1. Hello!

    First of all, welcome to MapleLegends! You are currently viewing the forums as a guest, so you can only view the first post of every topic. We highly recommend registering so you can be part of our community.

    By registering to our forums you can introduce yourself and make your first friends, talk in the shoutbox, contribute, and much more!

    This process only takes a few minutes and you can always decide to lurk even after!

    - MapleLegends Administration-
  2. Experiencing disconnecting after inserting your login info? Make sure you are on the latest MapleLegends version. The current latest version is found by clicking here.
    Dismiss Notice

Odd Jobs Community Thread

Discussion in 'Community' started by deer, Jan 7, 2021.

  1. Slime
    Offline

    Slime Pixel Artist Retired Staff

    641
    1,185
    381
    Apr 8, 2015
    Male
    Israel
    5:08 PM
    Slime / OmokTeacher
    Beginner
    102
    Flow
    ^ Probably my favorite rangifer's diary entry so far! Keep em coming!
     
    • Friendly Friendly x 1
  2. OP
    OP
    deer
    Offline

    deer Pac Pinky

    195
    398
    191
    Oct 27, 2020
    Female
    Oddville
    2:08 PM
    cervid
    Bishop
    130
    Oddjobs
    (You can also read the original diary entry here, and/or read the diary entry on the Oddjobs forum.)

    rangifer’s diary: pt. lv

    Taxonomising odd jobs, pt. ii: Building up a modern perspective. §1

    In the previous part (i) of this series, we asked a seemingly simple question: What even is an “odd job”? And I tried to answer this question in a way that hopefully codifies the way that odd-jobbed characters are typically thought of now. But, at the time of writing, the year is 2021; odd-jobbed MapleStory characters have been with us since at least 2003, and it would be truly surprising if this “modern” perspective was an accurate reflection of attitudes towards odd-jobbed characters even from the very beginning. Indeed, we will find that our approach in part i of this series was (intentionally) anachronistic. We have in mind two tenets, in particular, that were expounded in part i:
    These two tenets help to produce a conception of “odd job” that is highly idealised. As such, we expect concrete manifestations of these ideas to be simultaneously less perfect, and more nuanced, than their corresponding “platonic ideal(s)”, so to speak. In much the same way, the concept of “an ordinary fighter/crusader/hero” appears simple: the character is STR-based, chooses to either use swords or axes, and follows the {fighter, crusader, hero} throughclass. In actual reality, things are a little more complex; each player has a different idea of what they want their character to be and to be capable of, each has a particular implementation of the game that they are working within (including various versions of KMS, of GMS, of MapleSEA, of EMS, etc., as well as various versions of private servers), each has their own in-game resources that they have to work with (including, among other things, how “funded” they are, what time-limited items they have access to, etc.), and each has their own knowledge and understanding of the game-world and its mechanics. As a result, each concrete realisation of such an “ordinary fighter/crusader/hero” may differ from another in many ways. Skill builds, AP builds, equipment choices, and so on, can all vary in trivial ways — e.g. minute and usually unimportant details of SP allocation, or extra points in INT and/or LUK due to the initial roll of the dice at character creation — as well as nontrivial ways, e.g. hybrid sword/axe builds, specialisation in a particular handedness (one-handed vs. two-handed), and even “mixed”/“confused”/“bloodwashed”/“bluewashed” AP builds that intentionally allocate a nonzero minority of acquired AP outside of STR/DEX.

    Concerns about “knowledge and understanding of the game-world and its mechanics” are particularly relevant for what we want to do here: building up a modern perspective. This process of “building up” means taking a serious look at the history of odd (and more generally, experimental) MapleStory character builds, and in doing so, we will need to reconcile — and, more importantly, incorporate — antiquated views of MapleStory character builds within our so-called “modern perspective”, to create a synthesis onto which we can impose a taxonomy. In the previous paragraph, I hyperlinked to the English Wikipedia article on “diegesis” when writing “the game-world”; but this is actually misleading. MapleStory, as a video game, does include some truly diegetic (read: narrative) elements; but, these elements are largely fragmentary (and often lacking in consistency), and furthermore, are not generally considered to be the main focus of the game. MapleStory, as the MMORPG that it is, is characterised neither by diegesis nor mimesis; rather, it is a social game that is most largely characterised by the social interactions of the people who play it. As a result, the culture of the game informs what character builds are conceivable, and it is our highly advanced and optimised pre-Big-Bang (pre-BB) culture — which permeates the existing pre-BB MapleStory private servers (e.g. MapleLegends) — that allows for the highly-idealised concept of “odd job” that was put forth in pt. i of this series.

    As part of this pt. ii (“Building up a modern perspective.”), I want to take a walk through some historical (or perhaps, archaeological…) artefacts of MapleStory odd-jobbed gameplay from the past. Doing so will allow us to build up a modern synthetic perspective by not only comparing & contrasting past MapleStory cultures with our present one(s) — although this is extremely important — but also by giving us an idea of what odd jobs and odd concepts could be considered historically… primordial. It is easy enough (or rather, difficult enough…) to impose a taxonomy from scratch, relying on our modern understanding of individual odd jobs, but it might (or might not, who knows!) be useful to us to come from a historical perspective when we start arbitrarily drawing the lines that separate our taxa.

    capreolina’s 120 party!!!

    My woodswoman capreolina has, at very long last, achieved the most prestigious rank of bow-specialised woods(wo)men: woodsmaster. And like my pure STR bishop cervid and my darksterity knight rusa before her, I held a special level 120 party for her. You can watch a video of the party on the Oddjobs YouTube channel!!

    And along with the video comes capreolina’s Theme, which you can obtain a score for (compatible with MuseScore 3+) here: https://codeberg.org/deer/gists/raw/branch/master/capreolina-s_Theme.mscz :)

    capre getting in on some of that 4th-job goodness

    By the end of the party, capreolina was a woodsmaster, but only had a single fourth-job skill: level 1 Bow Expert. This is a nice little passive damage boost for just one SP, but not exactly exciting yet. In order to get to the juicy stuff, I needed to embark on some 4th-job skill quests:

    First was the Hurricane quest, which asked of me that I hunt Kentauruses in search of Helena’s Old Gloves. These, according to my fourth job instructor Legor, would be the gloves that the archery instructor in Henesys bestowed upon Legor when he was a fledgeling archer seeking to learn the ways of archery despite being blind from a very young age. Players of MapleSEA, GMS, and EMS will now be asking the question: who is this “Helena”? And what do her gloves have to do with anything?

    The answer is that the name of the first-job archery instructor in MapleStory is inconsistently localised into English. In KMS, JMS, TMS, and CMS, this NPC is known as Helena (e.g. ヘレナ in JMS; for KMS in particular, see transcription into Korean), but this name was ported into English as Athena (Pierce). This is a little odd, although in classical Greek mythology, Helena (of Troy) and Athena are technically siblings, as they both share Zeus as their father. That being said, Zeus was the father of, uhm, a lot of various mythological figures, so perhaps the relation is not as close as it sounds. More to the point, Helena and Athena are two very different figures: Helena is best known for starting a war that culminated in the fall of Troy due to simply being the most beautiful woman in the world. Athena, on the other hand, is the goddess of wisdom, craftsmanship, and warfare; indeed, the wisest figure in classical Greek canon. Using Athena’s name for the first-job archery instructor makes a good deal more sense, considering that Athena is associated with the more deliberate and strategic elements of warfare — as opposed to her violent and untamed brother Ares, better associated with warrior skills like Rage and Berserk. The connection with Athena even suggests a skew towards the woods(wo)man; Athena is usually depicted with a spear, and is also symbolised with olive trees. (Side note: Athena was also associated with owls, and romanised as Minerva, hence the “owl of Minervain MapleStory.) It is possible to suggest that the KMS/JMS/TMS/CMS name of Helena is from Hellen, the progenitor of the Greeks, but this leaves the final vowel -a unexplained (it would be phonotactically valid in Korean to leave it off, as far as I know, as the syllable structure in Korean is (C)(G)V(C)), and Helena & Hellen are actually understood to be etymologically unrelated. And naming the job instructor after Hellen makes, still, less sense than naming her after Athena. It should be noted that, in any case, the basis in Greek antiquity is clear & intentional; two of the first-job archer skills (The Blessing of Amazon and The Eye of Amazon) are named for the Amazons, female-only bow-wielding warrior-hunters renowned in ancient Greek and Latin literature for their combat expertise (particularly in archery) and strictly female society (interacting with men only briefly, strictly for the purpose of reproduction). It seems likely to me that the KMS/JMS/TMS/CMS name was chosen simply to make her “sound Greek”, rather than in reference to any particular figure.

    In any case, I went off to kill some random Kentauruses in hopes of finding this stinky old pair of gloves:

    [​IMG]

    [​IMG]

    And lo, there they were!:

    [​IMG]

    I took these old gloves back to Athena, much to her surprise, and realising that this meant I had Legor’s blessing, she decided to give me a new skill: Hurricane.

    [​IMG]

    Now, at level 1, Hurricane isn’t all that useful compared to my level 30 Strafe; the DPS just doesn’t match up. However, it is really fucken cool, and there’s some neato things you can do with it, like feathering it while you walk to get attacks that don’t actually interrupt your walking at all!

    And then I went to get my birdy, which was honestly the skill I was looking forward to the most. For this, I had to travel into the depths of El Nath. First, up the mountain:

    [​IMG]

    …got some Riche cards along the way…

    Into the Caves of Trial:

    [​IMG]

    And in there, a special map where the Phoenix keeps its nest. I stole(?!) an item from there and brought it back to Leafre, where I used it to summon the Phoenix itself:

    [​IMG]

    [​IMG]

    I had to subdue the Phoenix, which is no trivial task; the Phoenix is a proper boss, with 1.35M HP and all that. But after a little while, I claimed victory, and in turn, the Phoenix swore to follow me wherever I go:

    [​IMG]

    That is “its rule”, apparently. Kind of a weird rule, but sure. I appreciate my firey birdy taking out the trash for me:

    [​IMG]

    General, undifferentiated, all-inclusive, generic bossing section

    I did a few quick boss runs with Harlez, xBowtjuhNL, and Eoka before capre’s 120 party:

    We were joined by a nightlord by the name of Manneqquin when we did Papu, which made that run quite a bit more swift:

    [​IMG]

    And I did a trio Rav run with Harl and Bow:

    [​IMG]

    Later, after capre ascended to woodsmaster, I hunted down a BF with Bow, and we were considering duoing it (not sure that would have went too well…), but ran into Matchy along the way, and partied with them to take down a BF:

    [​IMG]

    I had actually met Matchy back when I was playing capre at LPQ, when they were playing their then-page (now pally) Nobless! So it was nice to bump into them in a random bit of the Crimsonwood forest~

    I actually finished the Capt. Latanica card set on my darksterity knight rusa:

    [​IMG]

    And, in my quest for points, I actually started soloing Capt. Lat on rusa while taking along my STR bishop cervid and my I/L magelet for the ride, just for the points…

    [​IMG]

    Thanks for the free points, pet autopot!

    And I also have been soloing Capt. Lat on capreolina, for the points:

    [​IMG]

    The final Capt. Lat of the (unextended***) event, I did with ducklings, STRginner of DuckNation:

    [​IMG]

    The first go-around, ducklings unfortunately didn’t make it (it’s imperative that she stay inside the boss, otherwise she risks getting one-shot by any of the magical attacks), but during this last one, she successfully made it through and got those sweet, sweet bossing points :)

    cervine does some more card-hunting!

    I decided to do a little card-hunting on my I/L magelet cervine, and get some event coins along the way~
    Being only T1 so far, I am still grinding for cards on cervine exclusively in Victoria Island. Going through my monster book partly in order, and partly by location, I went first for Axe Stumps:

    [​IMG]

    And then I headed to Golem’s Temple I to hunt Blue Mushrooms and Stone Golems at the same time:

    [​IMG]

    [​IMG]

    And then to Line 1 <Area 2> to hunt Jr. Wraiths and Stirges at the same time:

    [​IMG]

    By the time that I had gotten to 5/5 Jr. Wraiths, though, I was still 0/5 Stirges… so I headed to the Transfer Area to finish off the Stirge set:

    [​IMG]

    And with that many more event coins in hand, I finished cervine’s first event ring:

    [​IMG]

    Ouch. The Colorful Cube scrolls have a 90% success rate, but I managed to fail quite a large proportion of mine, including two of them on this one ring. This is particularly bad for cervine, as she is arguably my most stat-starved character of all, so every single point of INT (more generally, TMA) counts for a lot. Thankfully, I was later able to perfect a ring, so I prefer to think of it as equivalent to two slightly-scuffed (“(+2)” and 0 slots) rings :)

    70%

    70% is the success rate for [Mastery Book 20] Berserk, the first mastery book that I have ever actually used in my Maple career… and I passed it!!:

    [​IMG]

    As punishment for my crime of actually passing something that involves RNG, I failed two 70%s on the first two slots of the Aluminum Bat (not to be confused with the Aluminum Baseball Bat, which is also a two-handed sword that lacks job requirements and stat requirements) that I was gifted during capreolina’s 120 party:

    [​IMG]

    Harlez kindly gave me the 70%s that I needed to keep going, at which point I proceeded to fail the next slot as well, and then boom the thing on the slot after that (the fourth slot, in total). Wow~

    EPQing with Numidium and Medulla

    I did some nifty EPQs with fellow odd-jobbed friends, pure STR cleric Numidium (Taima, Tacgnol, Boymoder, Hanyou, Nyanners, Gambolpuddy, Inugami, Yotsubachan) and dagger fighter Medulla (Cortical, GishGallop, BowerStrike, WizetWizard, Amygdala, Subcortical, MageFP), on my DEX brawler LPQ mule sorts!:

    [​IMG]

    As it turns out, sorts makes a pretty good EPQ mule as well; besides being quite capable of DPSing down the boss (Big Bad Rock Man, or BBRM), I can use CSB during stage 2 to corral the stone buggos towards the centre of the map:

    [​IMG]

    All in all, we did a full 17 EPQs, enough to get sorts form 0/250 party points to 250/250 party points in that one session! Whew~

    A few more MPQs to wrap up alces’s points~

    And, finally, I needed to top off some of the points that I was saving up for temporary event gear. For my daggermit alces, the best way to do this was MPQing my way into another completed (read: zero slot) event ring. Here is ome screenshot of an Angy Fanky fight that I had in one of my MPQ groups:

    [​IMG]

    alces is, actually, already “done” with MPQ; I have long since finished a quite nice Horus’ Eye, and completed the Franken Lloyd card set. And by now, I have far too many excess pink marbles… But now that alces is level 85, it will be just one more levelup before she graduates MPQ forever~
     
    • Like Like x 2
  3. SiriusPlaque
    Offline

    SiriusPlaque Slime

    23
    0
    29
    May 15, 2020
    10:08 AM
    Lmaoo your bad luck with rng is almost comical at this point :’)
     
  4. OP
    OP
    deer
    Offline

    deer Pac Pinky

    195
    398
    191
    Oct 27, 2020
    Female
    Oddville
    2:08 PM
    cervid
    Bishop
    130
    Oddjobs
    (You can also read the original diary entry here, and/or read the diary entry on the Oddjobs forum.)

    rangifer’s diary: pt. lvi

    Anniversary event extension~ Wow~

    Taxonomising odd jobs, pt. ii: Building up a modern perspective. §2

    In the previous section of this part (§1), I said that “I want to take a walk through some historical (or perhaps, archaeological…) artefacts of MapleStory odd-jobbed gameplay from the past”. So I want to take a look at the archive which I maintain on the Oddjobs website. This archive is, as of yet, exclusively English material, but there are still some fairly old entries that we might find useful in our quest to build up a modern perspective of odd jobs.

    A Guide for an HP Warrior!

    The oldest (as far as I know; many of the dates are somewhat tentative, as many entries are likely to be reposts or to lack proper dating) entry on the archive so far is “A Guide for an HP Warrior!”, by ShurikenRonan, which is dated at 2006-05-07 (although as you can see on the post, it was updated continuously until 2007-08-14). This entry is what it says on the tin: a guide to HP warriors.

    The guide starts off (after some images) with the following:
    You can find similar themes in the context of other odd jobs (permabeginners, STR mages, etc.), but it might be somewhat surprising (or, maybe not…) to modern readers that odd and experimental jobs often got a lot of hate in the early days. It seems that, somewhat ironically, as the game got older and highly-optimised characters became the norm, odd-jobbed characters became more clearly defined and clearly intentional. Instead of an HP warrior (or any other odd job) looking like someone screwing up the game, the playerbase has matured to see them as legitimate ways for players to challenge themselves and play the game in a new way. I remember (albeit vaguely) my experience playing a STRginner in ye olden days, when I decided to accumulate defames (e.g. by asking the other person to defame me when trading fame) partly because I expected to be defamed simply for not taking first job advancement.

    Indeed, the guide goes on to explain the existence of HP warriors:
    This more or less outlines our modern concept of HP warrior; they are capable of doing PG training on monsters far above their level for fairly efficient experience (particularly during the first half of second job), and they are capable of going places that other jobs cannot, simply due to their immense HP. Interestingly, ShurikenRonan emphasises the symbiotic relationship between an HP warrior and a priest; the HP warrior gets lots of EXP from monster deaths, while the priest gets extra EXP for healing the HP warrior. EXP from Healing is usually not (at least, not in MapleLegends particularly) considered a large part of EXP intake for clerics/priests. And furthermore, as you can see in the quotation above, even HP spear(wo)men are considered. Normally, nowadays we think of HP warriors as being either {page, white knight, paladin} or {fighter, crusader, hero}, but this guide suggests HP spear(wo)man as a way to make what is effectively an HB/IW mule. This would have made more sense back then, as characters were generally lower level, HP washing was not a thing, and bossing strategies were not as highly-developed (especially with the complete lack of grade 4 skills) — so being able to reliably and efficiently take any hit (even from, say, Zakum touch damage) without dying would have made for a great HB mule.

    Other than that, this ancient guide pretty much upholds our modern concept of HP warrior. They really are pure HP, other than the 35 base STR needed for first grade advancement.

    LUK Warrior

    The second-to-oldest entry in the archive is a little funkier: “LUK Warrior”, by Dareka, dated at 2007-02-06. The thread starter suggests making a warrior that swaps out DEX for LUK, with the advantage being higher AVOID (in exchange for lower WACC). This thread gets slightly derailed here and there by users who are unfamiliar with the game mechanics, but are still confident that they understand. If you thought that this was too long ago for people to have a good understanding, though, you would be wrong. Already on the first page, kholdstare62 and zedster correctly point out that each point of DEX gives 0.8 WACC, whereas each point of LUK gives 0.5 WACC. And zedster and MagicalBow are able to correctly surmise that each point of DEX gives 0.25 AVOID, whereas each point of LUK gives 0.5 AVOID. On the second page, Sir Sigmund Freud points out a pure DEX warrior with the IGN MitGas, and says that they are personally planning on making a kind of dagger warrior/dex warrior hybrid, whose STR is capped at a maximum of 65 simply to be capable of equipping warrior-rogue daggers (like e.g. the Cursayer).

    However, it’s pretty clear from the start of the thread, and reading onwards, that the suggestion of making a “LUK warrior” is not exactly the LUK warrior that we know now; Dareka intends on adding STR, and just adding LUK when needed for WACC and/or AVOID. On page 3 (the final page) of the thread, SpiraeaKozak confuses the original poster’s intent (thinking that they really want a pure LUK warrior), and confuses the game mechanics as well, thinking that daggers use LUK as their primary stat (this is true only for rogues; other classes do not factor LUK into their damage with daggers). But Dareka corrects them:
    This lack of purity (still STR-based, not actually pure LUK) would disqualify this version of “LUK warrior” from being “odd” in the sense defined in pt. i, although this would still be an experimental build. I would hazard a guess that part of what made this make sense back then, was warriors actually being seriously WACC-starved. In, say, MapleLegends, it is not difficult for ordinary warriors to go pure STR (read: DEXless & LUKless) and still have the accuracy that they need. Therefore, “LUK warrior” would only make sense in the context of an odd build, in which case we expect it to be truly LUK-based (rather than being STR-based in actuality). But the replies to this thread (particularly that from Sir Sigmund Freud) indicate that players were still taking these kinds of experimental builds seriously, trying to apply game-mechanical knowledge to them, and even experimented with builds that would be considered truly “odd” even today, like DEX warrior and dagger warrior.

    STR Mage v3

    Okay, one last one for now.

    Going one entry further down the chronological list, we have “STR Mage v3”, by Hyperhal, dated at 2007-03-04. The “v3” is due to this being the third time that Hyperhal posted a STR mage thread, as the older versions are unfortunately lost to the ages due to Hidden Street suffering some sort of database issues (although v1 is included at the end of v3). Again, like the HP warrior guide that we looked at above, this thread starts by pushing back against haters:
    Hyperhal explains why she made a STR mage in the first place, saying that she originally started out by simply trying to make a mage in the most wrong way possible, adding AP into all 4 stats, maxing the skills that were considered the worse grade 1 mage skills, etc.:
    And here we have it: the STR mage is like a permabeginner, but with more skills (this notion will become important later). HyperHal points out that STR mages lack access to the beginner-only weapons like Frozen Tuna, Sake Bottle, and Maroon Mop: “So by sacrificing your godly beginner weapons, you gain a class that is much more durable than your average uber-beginner.”.

    Hyperhal goes the STR F/P mage route, although their reasoning is somewhat unusual. Usually, when we think of STR F/Ps, we consider Poison Mist to be their main skill, at least when compared to the other STR mage routes. But Hyperhal has a different idea:
    It gets even more interesting when we get to the AP builds outlined by this guide. Hyperhal considers three kinds of AP builds: the “Warrior Build”, the “Hybrid Build”, and the “Everything and Anything” build. The warrior build is what we associate with STR mages: STR-based, with possibly some DEX to make up for missing WACC. But the other builds very much go into gish territory. This is a common theme that is seen with odd mages in general; there are not usually terms that allow people to distinguish different kinds of odd mage builds, so any build that adds any significant amount of STR goes under the “STR mage” umbrella, even if they are not actually STR mages in the usual sense.

    woooosa is my highest-level character now

    First off, I did a little more sadsadgrinding at CDs, with my STR bishop cervid and my darksterity knight rusa. This was enough to get rusa to level 123(!!):

    [​IMG]

    Wowie. Every level, I get a little more HP and 3 more SP to spend on zerk, and so now my zerk threshold is just above the 4.6k HP mark. Now, that’s not exactly impressive, but every little bit counts! Maybe soon, I would try zerking at a boss (like Capt. Latanica) for the first time! But I’m just not there yet.

    A quick MPQ with mae and Cort

    I did a quick MPQ with mae (xX17Xx, attackattack, drainer, strainer) and Cort (GishGallop, Cortical, Medulla, Amygdala, Subcortical, MageFP, WizetWizard, BowerStrike); we were joined by CharKoayTeow, who mae met while LPQing on her DEX page, attackattack.

    [​IMG]

    Here, you can see us; with me, as my daggermit alces, mae as her permarogue xX17Xx, and Cort as their I/L gish GishGallop. We have been trying to get xX17Xx a Horus’ Eye, as the massive buff to it has made it extremely appealing for a low-HP job like permarogue.

    SE???

    I was doing some solo grinding (again, at CDs) on my woodsmaster capreolina, and was joined by Flow member Level1Crook who was there to see me level up to 121 :O

    [​IMG]

    Which meant putting my first points into SE!!:

    [​IMG]

    SE is… pretty broken for classes who already have passive critical-granting skills (read: assassins, archers). Even with level 3 SE (which unfortunately only lasts 30 seconds, as the duration is 10 times the skill level), capre suddenly starts hitting like a truck with her bow as soon as I cast it. Here is a @dpm test that I did while fighting Capt. Latanica with self-casted buffs, Echo, and Cider:

    [​IMG]

    63.6M DPH. Y I K E S .

    Bossin’ wif da krew

    On rusa, I joined ducklings, Sockzz (also a DK), Harlez, and xBowtjuhNL for a pair of Capt. Lats, of which I think ducklings did actually survive one!:

    [​IMG]

    Later, the same crew tried Ravana:

    [​IMG]

    The runs were otherwise fine, but poor ducklings didn’t make it :x It’s definitely possible for her to make it, even without me bringing cervid to have the ability to Resurrect her, but it takes a lot of dilligence about when exactly I can and cannot Rush Ravana, as well as a healthy heaping of luck, for it to work out.

    I duoed Capt. Lat with Gruzz (who unfortunately did not make it through the first run, thanks to Cool Frog fumbling a potion or two):

    [​IMG]

    And we then went on to fight a BF with Harlez and xBowtjuhNL:

    [​IMG]

    Which went quite quickly; with my HB and his own MG, Harlez was able to survive BF’s touch attacks and contribute extra DPS with SB and ME!

    Meet d33r

    Oh no! Not another one of deer’s peepee poopoo garbo odd jobbers!!

    But actually, yes, and this time we’re talking even more peepee poopoo than normal.

    Meet d33r (I was originally going to call her “XXXdeerXXX”, but that proved to be too unreadable), my vicloc clericlet:

    [​IMG]

    What is “vicloc”?

    What is “vicloc”, you ask? Well, it was not my idea (the idea was brainstormed by Cort and Kels), but I did write up the rules here: https://codeberg.org/deer/gists/src/branch/master/vicloc/README.md. Basically, “vicloc” stands for “Victoria-locked”, and it’s kind of like islanding, except instead of Maple Island, it’s Victoria Island. Another inspiration is sleepylock. This recalls earlier versions of MapleStory, before Ossyria was added into the game. But it’s a little more complicated than that; in particular, because Victoria Island is not as isolated as Maple Island is, we needed a few extra restrictions besides “just don’t leave the island”. Vicloc characters are disallowed from operating gachapon machines, so the only thing that they can do with gachapon tickets is transfer them to non-islander, non-vicloc characters through the Cash Shop inventory. Vicloc characters are allowed to party with non-vicloc characters, but cannot receive items/mesos from non-vicloc characters for any reason. This makes vicloc characters (by our definition) kind of steelman-esque; they lack the ironman-esque requirement of isolating themselves from parties (besides vicloc-character-only parties), but retain the isolation from non-vicloc sources of mesos and items. Vicloc characters also, notably, cannot make use of Cash Shop items (Myo Myo the Traveling Salesman and weather effects that grant game-mechanical buffs) that affect the game mechanics and affect what items are available for purchase. This is actually a stricter requirement than normal islanding (on Maple Island), as the Cash Shop is always available, even to characters who have never left Maple Island.

    Some may be surprised to see that, besides areas that are obviously part of Victoria Island (Lith Harbour, Perion, Ellinia, Henesys, Kerning City, Nautilus Harbour, Sleepywood), we also include Amoria and Florina Beach in our definition of “Victoria Island”. We actually had a difficult time trying to figure out what exactly “Victoria Island” is, with three areas in particular being uncertain:
    • Amoria
    • Florina Beach
    • Coke Town
    The basic arguments in favour of, and against, each of these areas being part of Victoria Island are basically as follows:

    Note: “Transport” is defined as any travel that absolutely requires a cost (in terms of mesos and/or items, e.g. Magic Seed must be consumed to travel directly from Victoria Island to Leafre), and/or takes time to travel (e.g. the flight from Victoria Island to Orbis takes five minutes). So, non-transport methods of travel are effectively the same as just going through a normal teleporter, which suggests geographic contiguousness.

    Amoria

    In favour
    • Amoria is only accessible through Henesys (which is obviously part of Victoria Island), and does not require transport to go between Amoria and Henesys.
    • One map that is part of Amoria, namely Meet the Parents, is accessible from HHGIII via a normal teleporter at the top of the map.
    • Amoria is apparently thematically similar to Henesys, and features some of the same monsters as on Victoria Island.
    Against
    • NLC return scrolls are only usable from locations that are obviously part of Masteria, with the exception that they function normally in Amoria. This might suggest that Amoria is part of Masteria, not Victoria Island.
    • Amoria is not referenced in the Victoria Island world map.
    • https://bbb.hidden-street.net/ lists Amoria separately in their categorisation of “Maps” into nine categories (Amoria, Ellin, Ereve, Maple Island, Masteria, Ossyria, Rien, Victoria Island, World Tour).
    • Amoria has maps that spawn Sakura Cellions, but cellions are a genus of creature that is normally associated with Ossyria, being otherwise entirely absent from Victoria Island.
    • The Amorian mainland (i.e. everything except Meet the Parents) is reachable from Henesys with no transport, but still requires talking to a special NPC (Thomas Swift).
    Florina Beach

    In favour
    • NPC dialogue suggests that Florina Beach is geographically contiguous with, or at least very close to, mainland Victoria Island (“Have you heard of the beach with a spectacular view of the ocean called Florina Beach, located near Lith Harbor?”).
    • https://bbb.hidden-street.net/ lists Florina Beach as part of Victoria Island, and you can see here that they provide a map of Victoria Island from a more advanced version of MapleStory, which includes Florina Beach directly on the Victoria Island world map.
    Against
    • Going to Florina Beach requires transport.
    • While Florina Beach is accessible via transport from Lith Harbour (Victoria Island), it is also accessible via transport from Orbis (Ossyria) and Ludibrium (Ludus Lake).
    Coke Town

    In favour
    • Coke Town is only accessible from Victoria Island towns, and going to/from it is done without transport.
    • Coke Town is apparently thematically similar to Victoria Island, at least insofar as most of the monsters are simply Coke-themed or icy versions of Victoria Island monsters.
    Against
    • Coke Town was originally an event-only region, and is thus unlikely to have any intended geography at all.
    • Coke Town is not referenced in the Victoria Island world map.
    • Coke Town includes a monster known as Yeti and Coketump, which is modelled after an Ossyrian monster.
    • Coke Town was a paid promotion by Coca-Cola, and is thus unlikely to have any intended geography at all.
    • Going to Coke Town does not require transport, but still requires talking to a special NPC (Cokebear Administrator).
    Based on the above considerations, we decided that the former two are in fact part of Victoria Island, and that the latter is not.

    My vicloc adventures so far

    In any case, you can imagine that I’ve really set myself up for a tough time here. Vicloc is already a pretty severe challenge when we’re the first ones on the server to do it — there is no “economy”, so besides the three PQs that we technically have access to (HPQ, KPQ, APQ), it’s basically just ironmanning. But in addition, I have decided to play an odd job, and in particular, one that tends not to fare very well without good equipment: the magelet. So, when reading about my vicloc adventures, keep in mind that I’m not complaining: I did this to myself on purpose!

    xXcorticalXx (Cort’s vicloc STRginner) and I decided to do a few HPQs, to get ourselves some rice cake hats. We found out very quickly, however, that we were simply incapable of killing the monsters that spawn in HPQ. So, in order to defend our bunny friend, we were forced to lure the monsters away by aggroing them with our single-target attacks:

    [​IMG]

    This turned out to, thankfully, be very effective; we were able to farm up a rice cake hat for each of us. I refer to my rice cake hat as my “safety hat”, as its whopping 22 WDEF saves me a huge amount of potions. Oh yeah — about that… I literally cannot afford potions! I only use potions given to me by quests, and for a long while I couldn’t even buy myself gear like a Stolen Fence or set of Subis (my Sweet Fork Cake is quite useful against monsters with relatively high MDEF, as I am pure LUK after all).

    My first big break was not actually mesos or potions. Instead, as the event was about to end, I scraped up enough points for exactly one (1) maple equipment box:

    [​IMG]

    And I got a 4/2/2/2 dana!! This is actually pretty much the worst possible blue dana for d33r, but really all I was asking for was a dana with >0 INT, and that’s exactly what I got!

    And with that (and spending my leftover 5 points on some goofy glasses), I put together d33r’s look:

    [​IMG]

    I already have an odd-jobbed cleric/priest/bishop (cervid), and her look is quite dark; so for d33r, I went with the polar opposite. Normally, outfits of this kind are traditionally associated with tragedy, but that was not my inspiration here.

    And I went off to do some quests. Being vicloc means that I very nearly have an obligation to complete every (or almost every) quest that is completable by vicloc characters. Annoyingly, some quests that are classified as “Victoria Island” are actually not doable by viclockers. But I went off to do the early Stump and Slime quests like The Stump Horror Story, Beauty or Beast, my first job training quests from Grendel the Really Old, I’m Bored, and I Need Help On My Homework!:

    [​IMG]

    [​IMG]

    Along the way (not a lot of return scroll use, and no taxi use, thanks to my nonexistent mesos) I ran into a sad gachapon ticket:

    [​IMG]

    Well, I’ll just have to redeem that one on another character…

    Here I am, doing Fixing Blackbull’s House and collecting some Leaves for Maya and the Weird Medicine later on:

    [​IMG]

    And then I came to the realisation that doing Victoria Island quests would mean doing some JQs… I’m not exactly a JQer myself, but thankfully this first one (John’s Pink Flower Basket) was doable after some frustration:

    [​IMG]

    And then I went to kill a bunch of Green Mushrooms for Arwen and Ellinia and DANGER! <1-G. Mushroom>:

    [​IMG]

    Which brings me, along with some other quests, all the way to level 21~ Next for d33r is a lot of KPQing, and I’ve already (finally) bought some actual gear in preparation: a Pink Starry Shirt, some Ice Jeans, and a Stolen Fence!

    cervid working towards getting Genesis~

    On my pure STR bishop, cervid, I have a spare SP laying around that I’ve not spent on MW, as level 8 of the skill would only increase the duration. So I suppose it’s high time, then, that I get Genesis. As a party trick, I guess. Who doesn’t want to see 15 MISSes at the same time, from a single spell?

    Anyways, I had not started the quest yet, as I knew it meant farming up the ETC items necessary to enter the Cave of Life. And those monsters are tough. But I was determined to try it anyways, and Gruzz and xBowtjuhNL offered to help, so I went around to a bunch of NPCs to start the quest off:

    [​IMG]

    Oh, Wiz… if only you knew how STRong I really am.

    The crux of this quest is that Shammos is actually an incredibly powerful magician — so powerful in fact, that when he threatened to annihilate the Maple World as we know it, Grendel the Really Old and Robeira agreed to lock Shammos, once their friend, into a dungeon below the Chief’s Residence in El Nath. They claim to only have spared Shammos’s life because Shammos was such a dear friend to them; but Shammos has, through self-corruption by the incredible and uncontrollable power of his own magic, and through years of isolation in the dungeon, become a creature barely resembling his original form at all. Shammos’s memory is foggy and distorted, and not everything that he says is coherent, but he insists that Grendel and Robeira did this to him out of jealousy for his magical power. If I could bring back one last sacred artefact necessary to unlock a new magical power — that artefact being the Life Roots — Shammos might teach me to wield this power myself.

    Sounds just slightly evil, but OK — unfortunately for me, I would need to enter the cave of the Horned Tail to even recover this artefact.

    [​IMG]

    I found Doom to be quite effective on these tank-like beasts:

    [​IMG]

    And we were fortunate enough to come across a Manon not too long after beginning to look for one:

    [​IMG]

    And so, not too long after starting — thanks to the help of Bow and Gruzz — I had all of the ETCs that I needed to enter the Cave of Life! Now, just to find someone selling HTPQ services…

    A few PPQs

    I did a handful of PPQs with Boymoder (Taima, Tacgnol, Hanyou, Nyanners, Gambolpuddy, Numidium, Yotsubachan) the STRmit, on my swashbuckler hydropotina:

    [​IMG]

    Above, you can see us trioing the Lord Pirate with LinusSosa the hermit. Along the way, Boymoder leveled up to 86!! Goodbye MPQ~

    Oh, and this was enough PPQs for hydro to get her first PPQ hat:

    [​IMG]

    Nice! I’m finally a real pirate now!!

    Items!

    It’s time to talk about some equipment. Here’s a Daiwa Sword that I scrolled after getting it from gachapon (thankfully didn’t boom on the final slot due to a 70% this time):

    [​IMG]

    Cool!! A great weapon for DEX/LUK pages, white knights, fighters, and crusaders, like attackattack and Hanyou!

    And I crafted another NRC!!:

    [​IMG]

    Wow, really nice this time (only got 39 WATK clean last time). And I actually managed to scroll it very well, passing six 60%s on it!!! 55 WATK, woweeeeeeeeeeeee~

    Speaking of nice changes of pace from my usual abyssmal crafting/scrolling luck, I bought up four(!) 2 STR clean Navy Christmas Socks in anticipation of my precious Green Christmas Sock expiring. This sock is going to be shared between quite a few of my characters (cervid, capreolina, hydropotina, rusa, alces, hashishi, dama), so I really want a nice 2 STR one with high DEX. Having bought four clean ones, the very first one that I attempted to scroll came out quite well!!:

    [​IMG]

    Now I guess I just gotta farm up more 30%s to try for 9+ DEX :eek: Unfortunately, not doing well so far, as I failed a 30% (but did not boom) on the first slot of one of the other sox. But hopefully I can make an even cooler one, and if not, this is already very nice!

    And finally, OmokTeacher (Slime, Slimu) actually managed to scroll an even more amazing Crimson Arcglaive than the one that he made before — this time with an insane 128 WATK and 15 STR!!! And he extremely kindly gifted me his old one, a legendary weapon in its own right, for use by my pure STR bishop cervid!!!!:

    [​IMG]

    Thank you so much Noam, again, for this incredible weapon as a gift!!! I’ve never been gifted anything like this in my Maple career <33333…
     
    • Friendly Friendly x 4
    • Like Like x 3
  5. Cortical
    Offline

    Cortical Mano

    10
    23
    25
    Jun 19, 2021
    12:08 AM
    Cortical
    Beginner
    93
    Oddjobs/Flow
    Vic islanders represent! And congrats on the socks and glaive, well deserved :)
     
    • Friendly Friendly x 3
  6. Charles
    Offline

    Charles Brown Teddy

    61
    81
    75
    Feb 2, 2015
    Male
    Ga
    10:08 AM
    deerdeer

    May join the Vicloc this weekend too :D seems refreshing!
     
    • Like Like x 1
  7. Cortical
    Offline

    Cortical Mano

    10
    23
    25
    Jun 19, 2021
    12:08 AM
    Cortical
    Beginner
    93
    Oddjobs/Flow
    Of course! I'll send you the rules and stuff in the Oddjobs discord :)
     
    • Friendly Friendly x 1
  8. Slime
    Offline

    Slime Pixel Artist Retired Staff

    641
    1,185
    381
    Apr 8, 2015
    Male
    Israel
    5:08 PM
    Slime / OmokTeacher
    Beginner
    102
    Flow
    v62 map also includes Florina Beach :p
     
  9. OP
    OP
    deer
    Offline

    deer Pac Pinky

    195
    398
    191
    Oct 27, 2020
    Female
    Oddville
    2:08 PM
    cervid
    Bishop
    130
    Oddjobs
    This statement is in reference to the fact that Florina Beach was originally not adjacent to Nautilus Harbour in versions of the game that didn't have pirates. Because there was no such thing as Nautilus Harbour!
     
  10. OP
    OP
    deer
    Offline

    deer Pac Pinky

    195
    398
    191
    Oct 27, 2020
    Female
    Oddville
    2:08 PM
    cervid
    Bishop
    130
    Oddjobs
    (You can also read the original diary entry here, and/or read the diary entry on the Oddjobs forum.)

    rangifer’s diary: pt. lvii

    Taxonomising odd jobs, pt. ii: Building up a modern perspective. §3

    Woodsmen

    The next artefact in our archive, and the next one I want to look at, is “Woodsmen”, posted by Zambies on the Hidden Street forums on 2007-08-12. The original post, in full, is as follows:
    Already in the second sentence, we see the term “jobbed beginner” used. This term will be important to keep in mind when creating our taxonomy, as it accurately reflects how certain odd jobs have been traditionally conceptualised and grouped together. Zambies goes on to imply that STR mages are part of this group. Based on the use of the term “jobbed beginner”, it seems that what Zambies has in mind is very similar to what are still known as woods(wo)men today.

    The first two responses to the thread are somewhat confused, not understanding the original post, and Sevin asking what a “woodsman” even is. Majal goes on to define woodsman as “Perma begginer that is a Bowman. They dont shoot. They just take advantage of the acc.”. This conception of woodsmen (and indeed, most jobbed beginners in general) as being essentially permabeginners who take advantage of some minor, but significant, benefits of their chosen class — usually passive skills and buff skills — is reflective of traditional attitudes towards these jobbed beginners. The worry here is about WACC; STRginners have obvious WACC issues, as they want to put as much of their AP into STR as possible, but at the same time have nothing built into their class that helps them with WACC (unlike e.g. warriors, who can passively gain +20 WACC just from mastery skills). The bowman class would seem to help solve this problem: The Blessing of Amazon passively grants WACC, and Focus gives a hefty WACC (and AVOID, for that matter) buff on top of that. Unfortunately for the woodsman, archers gain less WACC per point of DEX, and per point of LUK, than a beginner does: 0.6 WACC instead of 0.8 per point of DEX, and 0.3 WACC instead of 0.5 per point of LUK. But this mechanic is not even mentioned in the thread, as its effect is quite minor at lower levels and with lower-powered gear (read: less DEX and LUK from gear), in comparison to the large constant effects of The Blessing of Amazon & Focus.

    My guide to woods(wo)men is based on a more hybrid view of this odd job: one that is proficient in both melee and archery. But, looking at a more primitive attestation of the woods(wo)man odd job suggests that this hybrid view is more modern — or maybe not. One of the replies that I referred to as “confused” earlier, by pso_kik, mentions what they call a “strength bowman”: “Other than a strength bowmen (which means much less damage, but more stable), there's not much you can do with the archer class in terms of novelty builds.” The last part of this statement proves to be mistaken, but their description of a “strength bowman” as having much less damage, but more stable damage, is basically accurate to the ranged fighting style of woods(wo)men. But, at least when looking at this thread, there is no evidence of really putting two & two together here, and thinking of woods(wo)men as a hybrid melee & ranged job.

    Argentin makes the last reply in the thread, explaining what a woods(wo)man is, and how to allocate AP for one:
    The sentence “you will only need 25 DEX so you can get the first job and the rest you put into STR” is particularly notable here, for reflecting our modern conception of woods(wo)men as being pure STR archers. But again, similar to the “LUK Warrior” entry that we considered in §2, the emphasis is not on purity; it is accepted that any archer will need at least some DEX in order to get the WACC that they need to hit monsters.

    Brandini's Islander Guide

    Finally, I want to look at the next entry in chronological order: “Brandini's Islander Guide”, by Cacophobia (IGN MrBrandini), first published on 2007-11-05. This guide is a fairly straightforward guide to islanding (unfortunately somewhat uprooted by the later changes to Maple Island), but I want to focus on the “Builds” section.

    The section starts by saying that “Most people assume there is only one build all Islanders follow which is going Pure Strength, they are wrong.”, and goes on to list five different builds: strength islander, dex islander, hybrid islander, magelander, and luck islander. The classification of usual types of islanders has been done many times, in various somewhat different ways, but this basically captures the kinds of distinctions that are usually made. This will become important later on, when trying to fit islanders into a taxonomy of odd jobs. These different “flavours” of islander, if you will, would almost certainly be considered subjobs by the definitions laid out in pt. i, but because islanders are such a primitive and core odd job, the classification of at least some of these subjobs proves to be an important task. And, furthermore, magelander in particular is so distinct from other islander flavours — and so old — that it is worth mentioning separately.

    d33r prepares to take on KPQ anew

    I tried doing a KPQ on my vicloc magelet d33r at level 21, and pretty much instantly regretted it. I felt so utterly useless that I set a task for myself: to level up and come back!

    So I went to a nearby map (Caution Falling Down) to grind:

    [​IMG]

    Having hit rock bottom, I was indeed forced to train with my anniversary event claw:

    [​IMG]

    After grinding out a level or two, I thought I might make some progress on quests (without actually turning them in, as I wanted to save that for after KPQ, i.e. level 31+). Hopefully this would give me a nice break from the Octopus grind. So I went to collect ETCs for a level 15 quest, Don Hwang’s Request. Unfortunately for d33r, this didn’t turn out so well:

    [​IMG]

    So I went back to the g-g-grind:

    [​IMG]

    And eventually, I hit a much healthier level of 24!:

    [​IMG]

    Maybe now, I’m good enough for KPQ…?

    sorts gets a zhelm!

    I wanted to get a zhelm for my LPQ mule and DEX brawler sorts — after all, what LPQ mule doesn’t have a zhelm? So I went to collect some gold teeth for the Zakum prequests, on my darksterity knight rusa. Along the way, I got quite lucky, and was able to quickly finish up my Riche card set!:

    [​IMG]

    Once I had the teeth that I needed, sorts was off to the Door to Zakum:

    [​IMG]

    Not pictured here is me soloing the first prequest (yawn), and also not pictured is me doing the second prequest, in order to save myself the embarrassment. But rest assured that I eventually did complete both, and got my five EoF (not to be confused with EOF):

    [​IMG]

    And I bought an AFK Zakum run for her, and ended up with this slightly underwhelming helm, with no chance to reloot:

    [​IMG]

    …But hey, it’s not so bad. At least the stats aren’t 13/13/x/y or something!

    Doin’ woodsmaster stuffs

    I did some woodsmastering with my woodsmaster capreolina!

    The first order of business was to get capre access to some of the skills that don’t come as default skills for bowmasters. First off was Dragon’s Breath — this skill is only attainable by using a skillbook, but that skillbook can only be acquired in the El Nath PQ (ENPQ), and is untradeable. Again, this is a little weird, considering that ENPQ is supposedly just a quest that 4th-job warriors need to do to get Stance, but thankfully I have a warrior of my own who lacks access to Stance: rusa, my darksterity knight. So I failed to save poor Tylus (who, apparently, can never catch a break) once again:

    [​IMG]

    And I got quite lucky this time, getting the skillbook on only the second ENPQ that I tried:

    [​IMG]

    With that done, there was only one bowmaster skill left for me at master level 0: Concentrate. Starting this quest requires a handful of ETC items, one of which is the Ancient Book, which is a somewhat rare drop from Cold Sharks. I really didn’t feel like swimming around and farming these absolute tanks until I got lucky enough for it to drop, but thankfully for me, there was one on the FM for reasonably cheap, so I went ahead and took the shortcut there.

    But I did have to farm up some Ink Bottles from Squids (thankfully, these are not rare drops at all):

    [​IMG]

    And I took these ETC items (also including 30 Magic Rox and a USE item, Pure Water) to Wiz for inspection. Wiz said that I could use the materials to reveal hidden ink in the Ancient Book, and upon reading the invisible ink, he directed me to Athena Pierce:

    [​IMG]

    Athena Pierce told me that I’m a bowman. Excellent. So I tried talking to my third job archery instructor, Rene:

    [​IMG]

    Rene also agrees that I am definitely a bowman. Very well then.

    So I went to the Insignificant Being, who told me that I am a bowman. Just kidding — he told me that there is an ancient archery training ground built by the most powerful archer of legend, housed near the Insignificant Being’s location in deep Sleepywood. This training ground would allow me to train to be strong enough for the skill taught by this Ancient Book. But I would need something special in order to enter the training ground.

    It was at this point that I got confused; I was misled by the short guide (for 4th job skills) that I found online, and also misled by the entry in my quest journal in-game. I thought that, at this point, I just needed to party with another 4th-job archer and go in. I smega’d, asking if any 4th-job archers wanted to do the quest with me, and a new bowmaster by the name of wires responded and kindly agreed to help. I found that it was still impossible to enter, and eventually realised that I was mistaken about what I needed to do next in the quest. The 50 dark marbles that I needed to collect were not dropped by the monsters inside of the training ground; they were dropped by ordinary monsters in deep Sleepywood.

    Unfortunately, these dark marbles had a fairly low drop rate, I needed 50 of them, and they only dropped from Taurospears, Tauromacis, and Jr. Balrog. So wires said we could try the quest the next day. I hunted the marbles:

    [​IMG]

    And later, the next day, wires came back to assist!:

    [​IMG]

    Thank you wires!! And with that — 300 Jr. Balrogs killed — I was awarded access to the Concentrate skill:

    [​IMG]

    After that, I did some sadgrinding at CDs to get to level 122!:

    [​IMG]

    Cool~ Now my SE lasts for a whole 60 seconds :3

    d33r does KPQ

    Now that d33r was level 24, it was time to head back to KPQ much stronger than before. Well, okay, maybe not much stronger…

    In the image below, you may notice an abundance of “1”s and “MISS”es. This is where the Victoria guild gets their motto: “Killing KPQ gators, 1 damage at a time”.

    [​IMG]

    By the end, I was having to kill myself many times in order to stubbornly retry for the last item that I wanted from this PQ. I had gotten 5/5 on my King Slime set fairly quickly, but besides that, I just wanted a pair of 2 INT Squishy Shoes. I wasn’t prepared to be any pickier than that, so I would even have accepted x/y/2/0 shoes. But my luck in winning RollRolls was just nonexistent for a long time. Eventually, after many deaths, and many miserable KPQs in which I was still nearly useless, I won a roll. And it was then that I just about lost my shit, because I actually got x/y/2/1 shoes!!! Free from KPQ at last…

    cervid gets Genesis

    In the previous diary entry, we saw cervid, my pure STR bishop, working towards getting Genesis, with help from Gruzz and xBowtjuhNL. The only thing that was left to do was the Horned Tail Party Quest (HTPQ), in order to get a special quest item (Life Roots) that drops from the monsters in the PQ.
    I saw someone smega selling service for just this: HTPQ for Life Roots. So I took cervid into the Cave of Life, and was guided through the HTPQ by PaiKia:

    [​IMG]

    It was actually very quick, and much easier than I expected! Thank you, PaiKia. And with the Life Roots, Shammos was pleased:

    [​IMG]

    I had a skillpoint at the ready, so I rocked back to Victoria Island to test my newfound power. I was joined by viclockers Naganohara the clawpuncher (Spanolia, Taima, Tacgnol, Boymoder, Hanyou, Numidium, Gambolpuddy, Inugami, Yotsubachan) and xXcorticalXx the STRginner (moonchild47, Cortical, GishGallop, Medulla, BowerStrike), who got to witness cervid’s first ever (and possibly the first ever for STR mages in general?) Genesis:

    [​IMG]

    Questing with d33r

    Now that d33r graduated KPQ, it was time for some questing~

    The first order of business, as far as quests go, is a very special quest: the quest to advance to the rank of cleric! This would be my last ever job advancement, as 3rd and 4th job do not exist in a vicloc context.

    [​IMG]

    Eventually, I accrued 30 of those precious Dark Marbles, and was able to take the advancement!:

    [​IMG]

    Woweee~! Time to be really cautious about Teleporting so that I don’t waste my mesos!!

    I went and did the first part of the Mystery of Niora Hospital quest:

    [​IMG]

    I did the first two of Shumi’s jump quests. The first one, I was already somewhat familiar with. The second one… well, I’d probably done it once before, but that was a long time ago…

    [​IMG]

    But I managed (after some pretty severe laser burns, ow) to recover Shumi’s fat stack of cash. The first time, I was sympathetic. But Shumi really needs to stop leaving her money littered around these dangerous industrial booby traps!!

    I headed to Sleepywood to do some of the 99-kill quests, particularly the Evil Eye one, so that I would have tails for other quests like Strange Dish or Alex the Runaway Kid:

    [​IMG]

    I did the quest for 99 Evil Eye kills:

    [​IMG]

    And it took me every single potion in my inventory (including the HP, MP, and Magic Potions) over the course of roughly 35–40 minutes to get these 99 kills. Yikes… but I was determined to do it myself, crappy Magic Claw damage and all.

    I went and did some of Winston’s quests, including Planting Trees and Shawn the Excavator’s Request:

    [​IMG]

    And Winston’s Fossil Dig-up:

    [​IMG]

    A quaint little quest by the name of A Way to Fight Off Sleep:

    [​IMG]

    Mike, I hate to break it to you, but…

    And I was able to give Maya her medicine after getting some Charms of the Undead while I was doing the 99-kill quests:

    [​IMG]

    While I was in Henesys, I hit up the Camila’s Gem quest:

    [​IMG]

    Thx Utah. Little known fact*: every page in the Pig Illustrated is just another image of the same exact pig. There are no words.

    And I farmed at the mini-dungon attached to the Rain-Forest East of Henesys for 200(!) Pig Ribbons that I needed for Mrs. Ming Ming’s First Worry and Alex the Runaway Kid. Along the way, I somehow got no Ribbon Pig cards and 4 Pig cards:

    [​IMG]

    It was around this time that I started to get seriously worried about d33r. I knew that clericlet life was gonna be rough as a viclocker, and so I was unfazed by my rough experiences with the KPQ levels. But as I began to play her more in the post-KPQ levels, things didn’t seem to be shaping up quickly enough. Of course, I accepted my fate of always being very weak against non-undeads, but my original vision for this character was that I would be able to effectively — not nearly as effectively as a normal cleric, but still effectively — farm/grind on undeads across Victoria Island, like Zombie Mushrooms, Zombie Lupins, Jr. Wraiths, Wraiths, Mummydogs, etc.

    So, to get an idea of whether or not this would really be possible once I got some good gear and maxed out my Heal skill, I did some calculations with my damage calculator. Without going into the details, let’s just say that things didn’t look that impressive when I imagined a hypothetical level 41 d33r with godly (theoretically possible for vicloc, but not reasonably obtainable in the near future) gear fighting Jr. Wraiths (a level 35 monster) with Heal. So, with realistic gear, things were going to look much worse than even that. When I initially did these calculations and thought about what they meant, I’m not gonna lie, I was pretty brokenhearted.

    You might say: “deer, you have a damage calculator; why didn’t you do these kinds of calculations before making the character?”. The answer is simple:
    1. I already knew for a fact (and I knew this from my experience making the damage calculator) that Heal damage scales with LUK. In particular, the damage (after controlling for variables that we don’t care as much about, like skill level, MDEF, and target count) scales as essentially Θ((LUK + INT) ⋅ TMA). Obviously, my TMA would be quite low, and looking at the Heal damage formula, we can tell that the way to do the most Heal damage is to be pure INT (or close to pure INT) — but I was okay with that. Obviously, I expected to do less damage than a normal INT-based cleric. The point is that it wasn’t completely absurd to expect my Heal damage, as a pure LUK (and therefore INTless) cleric, to be workable.
    2. d33r was inspired by another character of the same build (clericlet) on another server, JoblessMS. Holy was (and still is) the highest-ranked character on the entire server, and that server is based on the same version of GMS (v62) as MapleLegends, so the Heal skill functions identically. Furthermore, JoblessMS had a lower EXP rate at the time (1x) and lower quest exp rate (2x), as well as far fewer functioning quests than MapleLegends. So I figured that if Holy could get to #1 as a clericlet/priestlet under those conditions, then I could do an okay job as a vicloc clericlet on MapleLegends. Being vicloc obviously makes it harder, but the much higher EXP rate would balance that out, right?
    Wrong, unfortunately. It seems that Holy has so much access to TMA gear that he would simply blow any character like d33r out of the water. So, for a bit, I was somewhat confused about what to do, as I felt quite demotivated by this realisation. But it is still true that d33r can function well as a supportive character, between her Heal, her Bless (not actually there yet), survivability due to MG, and mobility due to Teleport. So I tried that out:

    [​IMG]

    Here I am, fighting Stumpy with Sabitrama (Celim, Sommer, Copo, Fino, Vicloc, Bipp) the vicloc brawler. Sabitrama was kind enough to give me quite a few Stumpy cards, as they were hunting Stumpies and had already gotten 5/5:

    [​IMG]

    I think I’m 4/5 now or something? And I did some farming of quest ETCs with Sabitrama, including many monsters that I was incapable of reasonably taking on by myself:

    [​IMG]

    [​IMG]

    And we both finished up the Sauna Robe quest at the same time, as Sabitrama donated a few extra Diamond Ores to me that I needed to finish crafting a Diamond. I had four overall INT 10% scrolls at the ready (the Sauna Robe is the only overall in vicloc that d33r is capable of equipping… well, besides the Amorian ones that have no slots and no stats!), and to my (and everyone else’s) extreme surprise, the first one landed!!!!!:

    [​IMG]

    Y O W Z A .

    I decided against throwing any more OA INT 10%s at it, and the other 3 leftover ones can go to other vicloc mages. For this very special robe, I plan on very slowly & gradually finishing it as OA INT 60%s get generated in the vicloc world (it is only available as a drop from King Clang). 10%s and 60%s are the only OA INT scrolls that exist in vicloc, so this is the best I can do.

    And with that, we went to the Ant Tunnel nearby to farm some more ETCs for quests like Dr. Faymus’ Request and Jane and the Wild Boar:

    [​IMG]

    *Not actually a fact.

    Capt. Lat w/ OmokTeacher and Cortical

    I also did a fun pair of Capt. Lats with STRginners OmokTeacher (Slime, Slimu, JumpQuest) and Cortical:

    [​IMG]

    In which, during the second run, I was reminded just how difficult it can be to make Resurrection work. It appears that recipients of the effect get no temporary invincibility period whatsoever, and are of course dispelled upon death, making the resurrection very likely to result in an immediate second death…

    In which I am actually able to scroll items for some reason

    So, usually, when I scroll/craft things, I get horrible results and often omit them from my diary entries. The ones that I do include have given me a reputation for being just awfully unlucky at scrolling, for no apparent reason… But today I have a slightly different story to tell.

    Besides the 5/9 INT robe that I showed above, here are some other nifty 10%s that I managed to land on d33r:

    [​IMG]

    [​IMG]

    Not quite as impressive, and not impressive at all by outlander standards, but still pretty sweet.

    And I scrolled this Maple Crow, which I had leftover from a box that I opened during the anniversary event, and didn’t want to upgrade because it was already perfect clean:

    [​IMG]

    I have absolutely no personal use for it, but Yotsubachan made quite good use of it in LPQ!

    I also forgot to mention in a previous diary entry that I made some Cursayers for my daggermit alces, and wound up making a 107 WATK one and a 108 WATK one! Bipp borrowed the 108 WATK one for a bit before he hit level 100 (GDK time!), and other than that, it has been waiting for alces to become unstuck from MPQ and hit level 90.

    And last, but not least: The Skanda. I had, through my efforts during the anniversary event, gotten my hands on half a dozen clean Skandas, with a certain one of my characters in mind (I haven’t revealed her yet for a special reason, and she is not revealed in this entry for another, less special, reason: I’m too lazy), as well as Boymoder in mind, as she is a STRmit.

    Now, CFA 30%s don’t come cheap, so I was prepared to pay a smol fortune to scroll these bad boys. As expected, the first two boomed on the first slot, and another one just failed the first slot without booming. Then I scrolled this one:

    [​IMG]

    Yayyyyy~!! Now that is quite the Skanda!!! So good, it makes up for booming yet another Skanda on the first slot afterwards.

    Meet d34r

    Alright, before you accuse me of making this character just because d33r didn’t turn out the way I envisioned, I will have you know that I was already planning and discussing this character before d33r even hit level 30…

    Meet d34r, the dagger warrior. Cheesy as it may be, her name is actually two puns in one (d34r as in “dear”, but also as in “d33r+1”), so I couldn’t resist. Here she is, HPQing with fellow viclockers phloem (Jule) and Hailong:

    [​IMG]

    Once I had hit level 21, I headed towards Kerning City for the ol’ KPQ. Along the way, I took a slight detour to see if Mano existed:

    [​IMG]

    Very nice. That will make the Mano quest a lot easier when I do it at level 31 or so.

    d34r’s look is based off of two of my non-vicloc characters: my darksterity knight rusa and my daggermit alces. As my only other warrior, d34r inherits from rusa the face accessory, eye accessory (albeit in a different colour), and gloves — and as my only other dagger-only character, d34r inherits from alces the wings and generally dark look.

    And I did do some KPQs, like these ones with fellow viclockers SeaThief and Daggington (cobbages; also a vicloc dagger warrior, so I wasn’t very original this time :p), as well as Lv1Crook (Level1Crook; non-odd, non-vicloc baby corsair):

    [​IMG]

    [​IMG]

    I’m still very weak, but as expected, d34r is already turning out much more promising damage-wise in comparison to d33r… Right now, I plan on advancing to spearwoman so that we may one day go toe-to-toe with the biggest bad guy in Victoria Island, Ergoth. But Daggington is also going that route, so who knows. I may do fighter instead? :eek:
     
    • Like Like x 3
    • Great Work Great Work x 1
    • Friendly Friendly x 1
  11. OP
    OP
    deer
    Offline

    deer Pac Pinky

    195
    398
    191
    Oct 27, 2020
    Female
    Oddville
    2:08 PM
    cervid
    Bishop
    130
    Oddjobs
    (You can also read the original diary entry here, and/or read the diary entry on the Oddjobs forum.)

    rangifer’s diary: pt. lviii

    Every entry in this diary, up until (but not including) this one, has handled images in the same way. I used the most widespread losslessly-compressed raster image format available: PNG. As such, the screenshots are perfectly accurate (other than some minor edits that I do by hand in GIMP, usually just cropping) to the frames that were actually rendered by the MapleLegends client itself. To save on bandwidth (and memory, but mostly bandwidth), I aggressively optimise all PNG files before committing, by using zopflipng with the -m flag. zopfli is basically just a DEFLATE implementation written by some folks at Google. Using zopflipng with the aggressive -m flag essentially just takes a very long time (the README claims something like 2 or 3 orders of magnitude more time than libpng; probably more with -m tbh) to re-DEFLATE in more or less the most optimal way possible (or something like that).

    One of the images in this diary entry is very large, and I considered actually just putting a URL to it into the diary entry (rather than directly embedding it), requiring readers to manually navigate to the image with their web browser of choice, rather than having it automatically loaded. This kind of workaround is unfortunately necessary (in the case that I really wanted the loading to be lazy) because I don’t just push commits of this diary to the Codeberg repo — I also post to the Oddjobs forums as well as the MapleLegends forums. With the Oddjobs forums, <details> elements are not supported, so all images are loaded upon loading the page; furthermore, Flarum does not paginate threads, which makes for a particularly unpleasant & wasteful experience with lengthy & image-heavy threads like the rangifer’s diary one. The MapleLegends forums do support the BBCode equivalent (or at least, the XenForo-flavoured BBCode equivalent) of <details>: [spoiler]. Unfortunately, a certain truly awful web browser does not support the <details> element, which means that if XenForo wants to support masochistic users of said browser, they cannot translate [spoiler] elements into <details> elements — instead, a handful of <div>s and some clever CSS has to emulate this functionality. The problem is that this only works visually; the content within the <div> is still loaded normally. Reasonable web browser implementations will be lazy about loading the content of true <details> elements, because the content is not used at all, until the user explicitly interacts by opening the element up.

    I also considered if my use of zopflipng was aggressive enough. Maybe, by using some less common image format, I could achieve even smaller image sizes without losing any quality (i.e. still lossless). I already had in mind FLIF, which is, as far as I know, basically the cutting edge for lossless raster image formats in general. The FLIF website promises me a roughly 33% decrease in file size when compared to my old method (brute-forcing PNGs with zopflipng), which is pretty damn impressive, considering how aggressive zopflipng is. Unfortunately, FLIF has almost no web browser support whatsoever (see e.g. this issue w.r.t. implementing support in Firefox, which raises concerns about decode performance), so that’s basically enough to just throw it out of the window. However, I noticed that the FLIF website also compares to WebP, another image format which can support lossless (as well as lossy) compression. Because WebP was developed by Google, I guess they were able to push support for it on the web (thanks to their ownership of Google Chrome), and caniuse claims that WebP is supported by basically every major browser… again, except for IE. So, with WebP also promising me smaller file sizes than my old approach (although by a smaller margin than FLIF; the zopfli README quotes a 26% reduction), I am now switching to exclusively using lossless WebP for the images in this diary.

    If you encounter issues with this change, like not being able to view the images in the diary, I recommend switching to a supported version of a modern web browser like Firefox.

    Taxonomising odd jobs, pt. ii: Building up a modern perspective. §4

    dorangpuppy

    The last entry I want to talk about here is actually a series of three screenshots by BasilMarket user dorangpuppy. dorangpuppy was a KMS player, and posted some screenshots of old pre-BB odd-jobbed KMS characters on BasilMarket, saying the following:
    And these “funny characters” are indeed odd-jobbers in our sense, although unfortunately dorangpuppy never posted any screenshots of the second item on this list (islander).

    Here is one such screenshot, of a level 66 permabeginner wielding a Blood Snowboard at Tree Dungeon, Forest Up North IX. As can be seen from the old-style GUI, this is from a relatively early version of KMS before the GUI was revamped some time in 2005.

    And then we have two images ([1], [2]) of what dorangpuppy refers to as a “STR mage”, although upon closer inspection it seems that this F/P archmage is actually a gish, or perhaps gishlet, of some kind. In the first screenshot, we can see this level 125 mage hitting a Skelegon with their Elquines — this would require at least a combined total of 160 INT+LUK, probably more. For a STR mage, who has 20 base INT and about 4 base LUK, this would be more or less impossible. dorangpuppy gives a rough date on this “STR mage” (2007) and claims that they are the first ever F/P mage of this kind (at least, in KMS).

    As mentioned in an earlier installment, I have been focussing on just English-language resources for this part. So it is fortunate that dorangpuppy, a Korean speaker of KMS, has posted these screenshots on an anglophone site (BasilMarket) for us to see. This is also a testament to just how old odd-jobbed characters are in the history of MapleStory.

    So what?

    Well, let’s take a look at some of the odd jobs that have been explored, or at least mentioned, in the archive entries that we’ve looked over in this part:
    This would seem to outline some of the most primitive odd jobs: we have our permabeginners (both on-island and off), our weird-statted mages, our “jobbed beginners”, our weird-statted warriors, and last but not least, our blood warriors.

    Along the way, we learned some interesting things that might be surprising when coming from a more rigid modern perspective:
    • Unlike our modern notion of HP warrior only including fighters/crusaders/heroes and pages/WKs/paladins, HP spear(wo)men actually used to be a thing.
    • The idea of a LUK-based warrior is actually older than you might think, although our evidence for that suggests a much looser definition, that still has a sizable amount of base STR. Similarly, pure DEX warriors are also older than you might think.
    • The term “STR mage” has historically been used to collectively refer to any mage job that uses STR, including very different jobs like gish and gishlet.
    • The distinction between “woods(wo)man” and a mostly-STR archer has historically been a bit wishy-washy.
    • The colourful classification of various varieties of islanders has been with us for a very long time.
    • STRginners are about as old as MapleStory itself.
    Does this wisdom help us in making a(n) taxonom(y/ies) for odd jobs? Well, I hoped it might, but I will leave it as an exercise to the reader to figure out how useful this will really turn out to be :)

    I’m a rogue

    In the previous diary entry, I mentioned in a parenthetical: “I haven’t revealed her yet for a special reason, and she is not revealed in this entry for another, less special, reason: I’m too lazy”. Well, this time, I am forcing myself to be at least un-lazy enough to finally reveal this character in a diary entry. Many who know me in-game already know this character, and at this point she is already level 68! I can reveal this character to you in the same way that I revealed her to mae (technopagan, xX17Xx, drainer, attackattack, partyrock, maebee, strainer), who is the very person from whom this character was kept a secret, for this entire time. But first, a little context.

    mae plays a permarogue by the name of xX17Xx. This character has been around for a while now, and mae has often lamented that no one else plays permarogue — even though it’s obviously the ideal job. A while back, I had floated the idea of trying to get her hands on an NRC, but later, Cort (Cortical, xXcorticalXx, GishGallop, Medulla, Subcortical, MageFP) and I decided to do this, but keep it a secret so that we could eventually give it to mae as a surprise present. Cort also had the idea to make matching permarogues named xX16Xx and xX18Xx who would get to level 35 to meet mae at LPQ (where she was PQing with her DEX page attackattack at the time) and surprise her there.

    After a lot of effort and money poured into getting an NRC (and some 1M(?) mesos or so put towards a minimum WATK NRC with 0 slots left, to use as a joke) that would be suitable as a gift — NRCs are very rare and tend to be egregiously expensive, and I actually personally crafted two of them myself — we eventually got to the point where we were ready for the reveal. Except we were a little too slow and lazy to get very far past level 31… but oh well, at least 31 is high enough to be a true permarogue. Rather than do the reveal at LPQ (mae had moved on to EPQ with attackattack by this point), we just did it in the FM. Oh, and also, I wanted to play a permarogue seriously as well, and I didn’t want to steal mae’s name instead of using my own personal nomenclature, so I had created not one, but two permarogues by this point.

    You can watch the reveal here:

    mae meets the permarogues [YouTube]

    So, you won’t be seeing any more of xX18Xx (other than in the above video), but you will be seeing more of panolia:

    [​IMG]

    You can see panolia throwing some MTKs in the above image; the MTKs are not actually mine, but were lent to me by Cort. In any case, readers of this diary already know my daggermit alces; daggerlords tend to play a little like dagger-only permarogues insofar as they are stuck with Double Stab for life. So with panolia, I have been focussing more on the other side of the first-job rogue arsenal: claw usage with L7. It is also time that I reveal these shitty memes I made one day for no real reason… maybe I just wanted an excuse to use GIMP or something…

    [​IMG]

    [​IMG]

    Here is a smol screenshot of a usual interaction with OPQ randoms:

    [​IMG]

    Usually, when people ask “why”, I inform them that it’s really quite simple: I steal shit, and I beat people up. I’m a rogue. What else would I need to be?

    The screenshot further above, of panolia taking on a Wild Kargo, was taken while panolia was hunting for her Ellin ring. Here I am, farming the Firebombs for their ETCs:

    [​IMG]

    And with the ETCs collected for the final bit of the questline, I headed back in time to do the questline proper:

    [​IMG]

    [​IMG]

    [​IMG]

    Easily done, for panolia, shredding these poor old creatures to a fine pulp with Lucky Seven… And there it is! The ring~!

    [​IMG]

    The ring is known for its +1 allstat and its +100 MAXHP, and that +100 MAXHP is quite juicy for panolia. When using a claw, the +1 allstat only really matters because of the +1 LUK (as panolia is DEXless, and L7 does not take STR/DEX into account when calculating damage) — but actually, getting the ring at this time was quite crucial for another reason: HP washing.

    “Le gasp!!”, I hear you say. “deer, I thought you hated washing! Plus, isn’t it disallowed in your guild anyways?” Well, the short answer is: yes, and sort of. Oddjobs has pretty much always allowed washing, but only for permabeginners, as they are extremely squishy even when compared to an unwashed, but otherwise completely normal, archer/rogue character. I have never once prevented, nor even discouraged, any permabeginners from HP washing. Personally, my permabeginner (hashishi) is completely unwashed, and I intend to keep her that way for various reasons. However, it wasn’t until a bit later that I instituted another exception to the “no washing” rule, specifically for permarchers, permarogues, and permapirates. The reason for this was that I found out that, again when compared to a normal but unwashed archer/rogue, these perma-first-jobbers in particular (not permawarrior nor permamagician, for reasons that are perhaps obvious) were at a defensive disadvantage. At the time, two of the perma-1sts in the guild (xX17Xx and gogigagagigo) expressed some interest in experimenting with HP washing, so I decided to simply allow them to do it (after all, they are mechanically at a disadvantage), and lend some INT gear (borrowed from my I/L magelet cervine) to help out as well.

    When Cort encouraged me to make a permarogue (at this time, just xX18Xx), we discussed the possibility of doing some light HP washing (keep in mind that the no-leech rule is, and has always been, a hard rule for Oddjobs, so HP washing to a high degree is difficult or impossible anyways). xX18Xx (and xX16Xx) actually still, to this day, have 40 base INT each. Partly, we just wanted our characters to be more similar to xX17Xx (who has done some amount of washing as well), and I had never actually done HP washing in my entire Maple career, so I figured I may as well try it at least once — after all, I play in a server that is dominated by this mechanic, and I already incidentally have a bunch of INT gear anyways.

    So I taught myself the finer details of how HP washing actually works. I quickly realised that, in addition to HP washing obviously scaling with character level (more levelups with ≥10 total INT → more excess MP & fresh AP → more washes are possible to perform), which I knew already, the HP disadvantage of the three perma-1st-jobs listed above does not scale with character level at all. The advantage is simply a constant (albeit somewhat random) boost that is awarded at the point of second job advancement. Further job advancements (3rd, 4th) do not award any such boost. So I knew that, in the long run, simply allowing these perma-1sts in particular to wash arbitrarily would be kinda broken. Running some of the actual concrete numbers myself, I realised that I would be able to fairly easily (and at a reasonably low level, i.e. not having to get a very high level like 100+) meet the number of washes necessary to recoup the MAXHP lost from never taking 2nd job advancement. So part of my experimenting with washing on panolia is to show that my own guild rules are horribly broken: panolia has gone for 40 base INT, and with some of my own already-existing INT gear (plus a random INT cape I picked up for 3M mesos), I intend to prove this brokenness. It seems that (perhaps unsurprisingly…), no matter what way you slice it — the way that we are slicing it here being that we use unwashed non-odd archers/rogues as a baseline — the “washing problem”, if you will, is also impossible to really solve even if you institute self-imposed restrictions. Our guild rules are, of course, followed merely on an honour-system basis.

    So, anyways. I needed this +1 INT from the Ellin ring, plus my INT gear and some temporary equips from this past anniversary event, to achieve exactly 90 INT from gear (and thus 130 INT total) when in full washing mode at level 65+.

    One of the running jokes about permarogues is that, much like permamagicians mysteriously lacking Teleport despite all 2nd-job mages possessing it, permarogues mysteriously lack Haste, despite all 2nd-job rogues possessing it. So I get a lot of “haste plz”, some genuine and some sarcastic… Luckily for panolia, she has an in-game macro called “Haste” that seems to do the trick — for me, at least:

    [​IMG]

    (It’s slightly tricky to spot in this image, but the macro is just a macro for Nimble Feet.)

    And in OPQ particularly, now that I have hit level 65 and can use alces’s Fan, I like to switch between claw and dagger, depending on which one is more suitable for that stage (or that particular part of the stage):

    [​IMG]

    And here I am, fighting Papa Pixie with fellow odd-jobbers Hanyou (Taima, Tacgnol, Boymoder, Gambolpuddy, Numidium, Inugami) the DEX WK and Medulla (Cortical) the dagger fighter:

    [​IMG]

    Now that panolia is level 65 — and thus able to use the Skanda shown in the previous diary entry, as well as my PGC and Cort’s MTKs — she does absolutely too much DPS to poor papa pixel. One of the folks I OPQed with was shocked to see me consistently knocking papa pixel back, even when I was the only one on papa’s left-hand side, and horrified to see me level up upon completion of the PQ… I only grow stronger!!! We shall see how panolia performs once she hits level 86 and I wash away 36 of her base INT…

    d33r

    Iiiiiit’s vicloc time~!

    I’ve been playing my vicloc clericlet d33r a bit, while focussing on my vicloc dagger spearwoman d34r~ Here are some of the things d33r has been up to:

    [​IMG]

    Here is d33r, duoing a stumpers with fellow odd-jobbed viclocker, the bullet brawler SeaThief!

    And I managed to run across a Mano somehow, who was generous enough to drop a card for me:

    [​IMG]

    Later, I ran across another stumpers and was forced to solo it:

    [​IMG]

    Not gonna lie, that fight was a real struggle, especially because stumpers can heal itself! Speaking of stuggles, here I am with the usual face that I make when I’m trying to complete a JQ for the rewards:

    [​IMG]

    This is the second Sleepywood JQ, I think, and I actually was unable to complete it on my first series of attempts. Later, though, xXCrookXx helped me out with it, and with that I was actually able to finish this bad boy:

    [​IMG]

    Wowie~ And I scrolled myself a very nice cape!!:

    [​IMG]

    Not pictured here is me failing the first OA INT 60% that I was able to finally find (there is only one source of this scroll in vicloc: King Clang, a.k.a. KC or Casey). :(

    And finally, the following image I leave without any context whatsoever:

    [​IMG]

    d34r

    By the end of the previous diary entry, my vicloc dagger spearwoman d34r was level 23 or so. Well, I’ve been a tad bit lazy to write another diary entry, and in the meantime, I’ve been playing d34r a little. Okay, maybe “a little” is an understatement — she’s level 48 now :) No small increase in levels, especially for an odd-jobbed viclocker!!

    Also at the end of the previous diary entry, I was debating whether to go the fighter route or the spearwoman route. I was leaning towards the latter, and Daggington (cobbages, snoot; also a dagger warrior) sealed the deal for me:

    [​IMG]

    Rather than walk through the journey verbally, I want to try something slightly different here: I’ve composed many of the screenshots that I took while playing d34r into a linear collage. This is not really the first composition of this kind that I’ve done in this diary, but this one is quite large, and traces d34r’s journey from level 23 to level 48 more or less chronologically:

    [​IMG]

    As you can see, d34r takes after my other odd spearwoman (rusa, my darksterity knight) — soloing King Slime, Stumpy, Faust, and even King Clang! I also participated in the first-ever vicloc ZMM kill! Admittedly, I did none of the actual DPS (at least not to ZMM herself; I was mostly killing her Zombie Mushroom and Jr. Booger minions), for lack of WACC, but I did help support with Hyper Body!!

    I’ve gradually grown into a partial-NX look for d34r. All of my NX gear is designed around the original look that I made for her, but I’m really liking the exposed helmet, gloves, shield, and armour! It makes me feel tuff!!! My only real gripe is that the armour only actually looks good when I’m standing still or walking normally. A closer inspection (e.g. when attacking or jumping) reveals that it’s, bizarrely, actually a one-piece swimsuit — my shoes only appear to meet directly with my armour because of my dagger obscuring the bare part where my upper thighs show… >.<

    As you can also see, d34r is very str0nk now. But I hope to make her even str0nker!! There are still some things that I want to do that lie out of my reach for my vicloc characters at the moment. My next big milestone will be level 50, not least because I can finally shed my old level 30 Cass in favour of the sleek Sai~ I actually have one already, that Thinks (OmokTeacher, Slime) kindly traded to me, but it’s unfortunately only 61 WATK… I have been killing more Copper Drakes and Skeledogs (both of which drop the Sai) than I ever thought possible, and still have not actually seen a single one drop with my own eyes. If a get to level 50 without another Sai drop, I will just have to 100% the 61 WATK one!

    alces is stuck

    Some readers (maybe, I dunno) might be wondering why my daggermit alces is still level 85. The answer is that alces, xX17Xx, and GishGallop are collectively stuck at level 85 in an attempt to get xX17Xx her Horus’ Eye from MPQ. Here we are, with a priest by the name of MartijnHS, in stage 3 of the PQ:

    [​IMG]

    …In which you can see xX17Xx also throwing the same MTKs from earlier.

    So, for now, alces is in MPQ+suicide hell, but in the future, I hope to put on some real rogue gear!!

    <3
     
    • Like Like x 3
    • Great Work Great Work x 1
    • Friendly Friendly x 1
  12. lv1crook
    Offline

    lv1crook Capt. Latanica

    354
    187
    278
    Jan 19, 2021
    Male
    7:08 AM
    Level1Crook
    Corsair
    1
    Flow
    I’m just glad the HPQ bonus stage EPM test made the cut OnionCrazy
     
    • Like Like x 1
  13. OP
    OP
    deer
    Offline

    deer Pac Pinky

    195
    398
    191
    Oct 27, 2020
    Female
    Oddville
    2:08 PM
    cervid
    Bishop
    130
    Oddjobs
    (You can also read the original diary entry here, and/or read the diary entry on the Oddjobs forum.)

    rangifer’s diary: pt. lix

    Taxonomising odd jobs, pt. iii: Exploring the space of possible taxonomies. §1
    • In pt. i of this series, “What even is an odd job, tho?”, we defined some terminology that is used throughout this entire series. And then we explored an answer (specifically, my answer) to the question that gives the part its name. This can be summarised in the following checklist of items that every “proper” odd job (i.e. an odd job that is not a subjob of any other odd job) must satisfy all items of, on our view:
      • Defining a particular odd job (e.g. permawarrior) is simple.
      • Insofar as an odd job is “odd”, it is pure in its “odd” aspect.
      • Each particular odd job satisfies some intuitive notion of natural.
      • The name that we are choosing to use here, “odd job”, is simultaneously fortunate and unfortunate. The “simplicity”, “purity”, and “naturalness” of odd jobs (as listed previously) justifies the use of the word “job” in the phrase “odd job”.
      • Odd jobs are, to the extent possible, atemporal.
    • In pt. ii of this series, “Building up a modern perspective.”, we considered that — depending upon a large number of various factors — one’s perspective/view/conceptualisation of what “odd jobs” are in particular, might differ significantly enough that we want to dedicate some time to other perspectives as well. In particular, we went through and analysed a number of entries in the archive, trying to reconstruct something resembling a historied perspective.
    In this part (pt. iii) of the “Taxonomising odd jobs” series, I want to (try to) get into the meat of it. After all, ultimately the point of this series is to do as the name says — to taxonomise odd jobs. I strongly suspect that the taxonom(y/ies) that we arrive at in this series will only be a few in a large space of possible reasonable taxonomies. So, with this in mind, it’s time to do a bit of exploration.

    Phenetics vs. cladistics

    I thought a little bit about what starting points one might have for a taxonomy of odd jobs. By “starting point”, I mean a guiding principle that outlines the basic structure of the taxonomy. Maybe I didn’t think about it hard enough, but in any case, what I came up with vaguely aligns with a dichotomy often touted in biological taxonomy: that of phenetics vs. cladistics. For readers who are not biologists, a brief overview of these terms is in order.

    Modern biological taxonomy is based on cladistics. Cladistics is deeply rooted in evolutionary biology. Cladistics organises species into “clades”, where each clade is rooted in a single common ancestor, and additionally contains every descendent of that common ancestor (which might be zero descendants, in which case the clade is an isolate, having just one member). It does not contain anything else; this means that clades are monophyletic by definition (indeed, “clade” and “monophyletic group” are usually used synonymously). As you can see, this makes cladistics fundamentally based on decendancy.

    On the other hand, phenetics reflects an older (and, in a biological context, largely outdated, save for some numerical methods that can sometimes be useful in cladistics) approach. Rather than having anything to do with evolutionary biology, phenetics classifies its subjects based on their observable properties. In the biological case, this means morphological/physiological/genotypical/ethological concerns. The phenetician’s job, then, is essentially to catalogue the properties of various species, and then determine a way of clustering them — based on some notion of similarity — to form groups. From an evolutionary perspective, phenetic methods fail to distinguish between — for example — a distinguishing trait shared between two species that is the result of synapomorphy, and one that is the result of homoplasy. More simply put, a shared trait that is shared as the result of both species inheriting the trait from a shared common ancestor, is indistinguishable from that shared trait being the result of separate innovations that just so happened to innovate the same thing (kinda similar to multiple discovery). From an evolutionary perspective, this is a fatal flaw of phenetics, as it makes phenetics largely blind to evolutionary relationships.

    The analogy between the biological phenetics-versus-cladistics dichotomy, and the subject matter that we have at hand (MapleStory odd jobs), is maybe not super clear. A more phenetics-oriented approach is the obvious way forward here. This isn’t biology; there is no evolutionary process per se, and thus no clear notion of decendancy. But, nevertheless, it appears possible to me, at first blush, to fabricate some kind of hierarchy that may hope to (vaguely) resemble a phylogenetic tree (or perhaps a forest of such trees). As we explored in pt. ii of this series, some odd jobs can be thought of as more “primitive” than others, and may have occurred (i.e. were actually invented and played) earlier. Furthermore, the subjob relation could come into play here as well — sometimes, jobs give way to subjobs, and these cases might look like decendancy as well.

    On the phenetics-y side of things, I want to investigate some kind of “bag of properties” approach(es). I’m borrowing the phrase “bag of properties” (sometimes also “property bag”) from obscure programming jargon, where the pattern in question is commonly used in languages like ECMAScript that combine a structural-typing-inspired approach (as in e.g. OCaml) with a complete lack of types*. This bag of properties view, where each object (viz. each job, in our view; or each (sub)species, in the biological view) is merely a bag of properties, and objects are compared simply by pairwise comparing their properties against one another, distills down the ontology of the phenetics(-inspired) approach. Of course, one issue is deciding what properties exist — as we will see, there are many possible properties that can potentially be assigned to a single job, so our bags will perhaps look a bit larger than you expect. And perhaps even more of an issue is deciding which properties are relevant, and how relevant.

    *Some of you are currently yelling at your screens: “What do you mean, ‘complete lack of types’? If it’s dynamically typed, then surely it has types, they are just dynamic!”. Unfortunately, you are all mistaken!! The form of an object’s in-memory representation, and its type, are never the same thing. To conflate them is a category error — the former is physically manifested, and relevant only to code generation and runtime behaviour, whereas the latter is merely an abstract mathematical construct, relevant only to the proofs that a translator generates when typechecking an otherwise well-formed program. That’s right, there is no such thing as so-called “dynamic typing”!! Fight me about it!!! (don’t actually. or do.)

    PQ2

    In pt. liii of this diary, I attended a virtual rave (as my daggermit alces) at FM 4-20, called “PartyQuest”. The rave was hosted by osmanthus (IGN zkgui) on datafruits.fm, and featured technopagan (mae; xX17Xx, drainer, attackattack, strainer, partyrock, maebee) DJing as one of the main acts! I attended the sequel this time — PartyQuest 2 — which was set up similarly, except that it was hosted in channel 1 of Maya’s House (actually called “Henesys Townstreet” in-game for some reason). I attended as my I/L magelet cervine, and later as my besinner hashishi. You can see me as cervine in the screenshot below, which features permabeginners drainer, Cowbelle, and Cortical, as well as Monologue (you can’t see the nametag, as he is the bottom-left-most character in the screenshot), who (as I found out during the rave) was a major active figure in the permabeginner scene of GMS:

    [​IMG]

    The above screenshot was not taken by me — credit for the screenshot goes to CowbelleCowbelle!

    The adventures of d34r

    That’s right, it’s time for more vicloc adventures!! My vicloc dagger spearwoman, d34r, hit level 50 :O

    [​IMG]

    To celebrate this milestone, an outlander by the name of Boymoder (Taima, Tacgnol, Hanyou, Numidium, Gambolpuddy, Nyanners, Spagonia, Naganohara) taunted me with forbidden juice:

    [​IMG]

    Where the outlanders get these horrible corrupting substances, I do not know. But fear not — I resisted all temptation.

    I’m keenly aware that, now, I have not one, but two, odd DKs. Okay, technically one of them is not (and can never be) a “DK” per se (as 3rd and 4th jobs do not exist in vicloc), but you know what I mean. So I brought my darksterity knight rusa to meet with d34r:

    [​IMG]

    F6 indeed.

    I needed a Mushmom ETC in order to complete a level 40 questline by the name of Jane the Alchemist. I wanted to complete the quest earlier, as the main reward is access to Jane’s shop, which offers discounted(!) Unagis and Pure Waters, but I needed to be at least level 50 to be able to hold mapowner for the Mushmom map. Now that I was level 50, I picked a channel and camped out waiting for Mushmom:

    [​IMG]

    Eventually, a mushroom mother did appear!:

    [​IMG]

    At this point, I would still have some serious issues hitting Mushmom — she was 10 levels higher than me, has 27 AVOID, and I am merely a masteryless vicloc warrior. But I was able to defeat her, after not too much trouble:

    [​IMG]

    Yay! The ETC dropped!! With this, I had all that I needed for my quest, but I figured I would check some other mushroom mother channels in case I could get a second ETC for other questers (including my clericlet d33r). Along the way, I was unexpectedly slain:

    [​IMG]

    x.x

    deerhunter turned out to be the woodswoman of Spagonia/Naganohara, and the pun of her name is three-pronged:
    1. Because deerhunter is a woodswoman going the bow-using (hunter) route, she will actually permanently be a hunter, as 3rd and 4th job do not exist in vicloc, so there is no advancement to ranger nor bowmaster.
    2. She actually enjoys the music of the American band, Deerhunter.
    3. I am deer, the one being hunted. ;~;
    My search for other Mushmoms turned out to be shockingly successful. In just two generations of Mushmom, I was able to go from 0/5 Mushmom cards to 5/5!!:

    [​IMG]

    And with a third generation, I managed to find a rare and powerful artefact… the legendary Ilbi:

    [​IMG]

    Wowee~

    I later went on to do the Building a New House For Blackbull quest, which meant collecting some rubble from Stone Golems:

    [​IMG]

    I already had the other ETC items on hand, so I went to complete the quest, only to find that it is bugged:

    [​IMG]

    [​IMG]

    For some reason, I was offered a Bow WATK 10% as one of the possible rewards, and, figuring this would be more useful than the others (as I was not offered a Dagger WATK 10%), I selected it… only to end up receiving no scrolls at all. I reported this on the MapleLegends forums, replying to a bug report on this quest from over three years ago, 2018-05-13!

    In any case, I continued on with the Jane the Alchemist questline. I needed some Drake’s Blood, so I headed to deep Sleepywood and tried grinding on some Drake maps:

    [​IMG]

    [​IMG]

    And in the process, I finished the first part of the A Spell That Seals Up a Critical Danger questline:

    [​IMG]

    Also for Jane the Alchemist, I needed a Moon Rock. My only way of obtaining one was to ask Arwen to craft it for me, but Arwen needed a lot of ingredients for that:

    [​IMG]

    Fortunately, thanks to my hoarding ways, I was only shy of a Gold Ore or two, so after a bit of hunting Blue Mushrooms for that, I was able to complete that final part of the questline! Yay for sweet potion discounts~

    It was now time to farm some cards. In vicloc, there are a grand total of 69 possible card sets that can be obtained, meaning that the tier 2 Monster Book Ring is the highest attainable Monster Book Ring. Getting to T2 means getting at least 60 out of the 69 total possible sets, which is much more easily said than done — not every monster in Victoria Island is easily farmed, and some, like Jr. Balrog or Ergoth, are hard to defeat at all. But T1, requiring only 30 sets, shouldn’t be too hard. So I went to fill out some of the easiest sets in the book:

    [​IMG]

    [​IMG]

    [​IMG]

    [​IMG]

    [​IMG]

    [​IMG]

    [​IMG]

    [​IMG]

    [​IMG]

    [​IMG]

    And with that, the T1 ring!:

    [​IMG]

    I actually went and did the 2,300 dolls questline, Rowen the Fairy and the Cursed Dolls. For this, my farming location of choice was Monkey Swamp III:

    [​IMG]

    Throughout my zlupin grind, I came across a map that I wasn’t even aware existed:

    [​IMG]

    Very mysterious. It appears that this map was never made use of in pre-BB MapleStory, although later versions included some quests involving JM From tha Streetz and the Knocked Trash Can inside of this map.

    Eventually, after countless zlupins slain, I finished the infamous questline!!:

    [​IMG]

    Oh, and I hit 5k HP w/ HB too :))

    [​IMG]

    Now that I was getting into the 5X levels, it was time to try out FoG (or, if you prefer, “TfoG”…):

    [​IMG]

    Wowee. Some extremely impressive solo EPH… but at what cost? The answer is: absolutely way too many Unagis. The main issue here is that, at this point, my PDD has outpaced my ability to gear up. Even with the +60 WDEF buff from Iron Body, warrior PDD scales so swiftly with level, that I cannot hope to keep up at this point. So I take a lot more damage from monsters in general (even lower-level ones), now that I am getting higher in level.

    I did the Faust questline:

    [​IMG]

    Which, like the other area boss questlines, gives some quite good rewards!

    The gist of the questline is that, in his younger days, Grendel was more greedy and reckless with his learning and usage of magic. So reckless in fact, that he created a mysterious dark evil in the southern forests of Ellinia which was responsible for magically corrupting the mind of his disciple, Reef. Reef was known by Grendel to be a sweet fellow, but with the corrupting influence of Grendel’s magic-gone-wrong, Reef became obsessed with Arwen, and grew jealous and vengeful when Arwen didn’t reciprocate his obsession. Arwen is where the questline begins, and she attributes Reef’s rash and awful behaviour to him being a mortal (unlike Arwen, who is a fairy). Later, Grendel admits his wrongdoing and assures us that Reef had always been a perfectly good and upstanding person before the magic got to him. Somewhere in there, big monke gets involved, with the straw puppet controlling our big monke guy being in need of defeating. With big monke and its puppeteer defeated, Reef’s spirit is finally freed forever.

    [​IMG]

    (The quest shown in the screenshot above, which is catalogued separately from the rest of the questline in the Quest Journal, is entirely wordless, and simply awards the fame seen above.)

    I also finally got the NX I needed for a pet with some autoloot. I had never noticed it before, but there is a smol level 10 Victoria Island quest that only shows up in your Quest Journal if you possess a pet:

    [​IMG]

    I went and did The Alligators at the Swamp questline:

    [​IMG]

    And with that, I went to Florina Beach (Clang and Lorang in particular) to farm Clangs for that sweet, sweet Yellow Umbrella. The Yellow Umbrella is perhaps the main item that I really need to rehabilitate d33r… with 52 MATK on average, and no stat requirements to wear it, it is the best possible weapon for d33r, besides the level 54 Heart Wand/Staff.

    [​IMG]

    After many (many) Clangs killed, I was able to finally find one of these precious items (after, of course, finding just about every other equipment item available in Florina Beach, including the Colorful Tube)!! Somewhat underwhelmingly, it is 3 MATK below average (49 MATK), but hey, that’s a whole lot better than nothing!!

    I was able to do some duo FoG with Thinks (OmokTeacher, Slime, Slimu), to excellent results:

    [​IMG]

    Aaaand I did the King Clang questline, as well! The end of the questline has you begrudgingly roleplaying through a heroic damsel in distress sequence fabricated by Riel:

    [​IMG]

    Glad to be of service, I guess…

    The adventures of d33r

    In other vicloc news, more woe befalls my vicloc clericlet d33r. Although I did, thankfully, get a (rather a bit under average) Yellow Umbrella, I had less luck with scrolls. As readers of this diary will know, I miraculously passed an OA INT 10% on the first slot of d33r’s Sauna Robe. My plan was to 60% the rest of the way. The only source of OA INT 60% in vicloc is King Clang, so many King Clangs I have killed. Many King Clangs. Like, a lot. So many, in fact, that I managed to farm up not one, but two OA INT 60%s (along with a whole lot of other stuff, and a few dozen cards). And Daggington (Clinks, cobbages, snoot) very kindly donated another OA INT 60% to me! So with these three 60%s in hand, I proceeded to fail every single one of them on my precious robe.

    In other news, while questing, I incidentally ran across a Stumpy, and got a second solo Stumpy kill under d33r’s belt:

    [​IMG]

    Although still a tough fight, d33r has still improved considerably since the last one, so it was a good bit easier this time. Speaking of questing, I did the main Florina Beach questline with the help of d34r:

    [​IMG]

    This questline requires a grand total of not one, not two, but 65 Coconuts! That’s a lot of tree-whacking.

    Later, I took a trip to vom town with xXCrookXx (Level1Crook, Lv1Crook):

    [​IMG]

    …Who helped me with the Icarus and the Balloon quest!:

    [​IMG]

    And all hope is not lost; I managed to pass 3/5 Ear INT 60%s on my Leaf Earrings, and I scraped up a 4th OA INT 60%, which I actually passed!

    Hopefully we will see d33r confidently wielding a Yellow Umbrella in the future~

    Doing NLC quests w/ Boymoder

    STRmit Boymoder was looking to do the Urban Warrior questline, which involves completing a number of quests & questlines around NLC, in exchange for the final reward of a Warrior Throne. I decided to join her on my I/L magelet cervine, who had not done any of the quests in the questline except for the quiz (which is trivial anyways) and Cleaning Up the Streets (which is actually, believe it or not, repeatable, although the repeat reward is pitiful).

    We started off with Featherweights, an otherwise useless quest that rewards almost nothing for the effort put in. This meant slaying some Boomers:

    [​IMG]

    And also slaying some Gryphons (also necessary for Cleaning Up the Streets):

    [​IMG]

    The final segment of the Subani’s Legacy questline requires farming for some pesky MesoGear Map Pieces from I.AM.ROBOTs:

    [​IMG]

    And, perhaps most annoyingly, Maple Balrog’s quests require going to Ludus Lake and back each time. While in Omega Sector farming Barnard Grays for Lost!, we found a card or two:

    [​IMG]

    And for ReVersal, we headed to the Ludibrium Clocktower for some Tick-Tocks:

    [​IMG]

    [​IMG]

    And with that, we could sit back & relax on our rather impressive-looking and rather blue thrones!:

    [​IMG]

    pan oiler

    It is now time for our regularly scheduled OPQ time with my permarogue, panolia:

    [​IMG]

    With a number of honestly rather subpar (not gonna name any names…) OPQs under my belt, I ascended to the rank of 1st job, pt. iii!:

    [​IMG]

    Unfortunately, OPQing straight from level 51 to level 70 was only enough to get me 3/5 Papa Pixel cards, so I had to kermit sewer slide:

    [​IMG]

    Thankfully, I was able to buy two cards off of other players (for 1M mesos each) to finish my set after only one death, so graduate OPQ I did:

    [​IMG]

    Over the course of panolia’s OPQ career, I managed to rack up enough feathers to get myself 4 wristbands! In order to do better than my 12 WATK mittens, I would need to (assuming I don’t want to pay considerably more money) pass GFA60s on all 5 slots. Now, the probability of actually succeeding (assuming that I don’t resort to GFA70s, which are much more expensive) is 0.6⁵ ≈ 7.78%. But, considering that I had 4 tries, the odds of getting at least one success wasn’t altogether awful (although still strictly less than 50%). The probability of failing a given wristband is 1 − 0.6⁵ ≈ 92.22%; so the probability of failing all 4 is 92.22%⁴ ≈ 72.34%; so the probability of succeeding at least one is 1 − 72.34% ≈ 27.66%. Not terrible, but not good enough for me to succeed any of them this time. R.I.P.

    And I got to do my final OPQs with STRginner Furbs!:

    [​IMG]

    And now, panolia is off to MPQ~ Actually, that means that a certain someone is finally freed from MPQ…

    alces is free at last~

    That’s right, alces (my daggermit) is bacc~

    Rather than a proper graduation from MPQ (which I had already done quite enough of on this character), I joined STRmit Boymoder and F/P mage 2sus4u in doing some PPQs!

    [​IMG]

    Later, we encountered some confusion over my use of MU:

    [​IMG]

    And I continued PPQing even after Boymoder left, and 2sus4u took a break to advance to 3rd job F/P mage (grats again!), at which point I met lesweetpea, as well as DevilDog96 and Ellie01, who appeared to be hardcore questers by the look of it (175-ish fame each, guildless, unoptimised assorted gear)!:

    [​IMG]

    I did finally take a break after some 18(!) or so PPQs, when I finally hit level 86~! Some readers may remember that I started questing on alces at a lower level than I usually do (I tend to be a bit greedy and save my quests for level ≥86, when KPQ/LPQ/OPQ/MPQ are no longer options), but I put it on hold at around level 71 or so — knowing that alces would have a hard time grinding at level ≥86, being stuck with masteryless Double Stab forever. Now that alces is finally at that point, I started where I left off — killing 400 Jars:

    [​IMG]

    And I started working on some other quests around the area as well!:

    [​IMG]

    [​IMG]

    I’m excited to play alces more in the future c:

    These bosses are not on Victoria Island! Bah, humbug!!

    I know, I know — but I can only kill so many Big Monkes and Caseys… Here are some of the outland bosses that I killed alongside xBowtjuhNL, Gruzz, and Harlez!:

    I was summoned to help kill Rav on my darksterity knight rusa… Unfortunately, all of my NX equips had expired, and I had nearly no NX on that account as a result of voting on panolia’s and on d34r’s accounts instead >.<

    [​IMG]

    Later, I duoed BF with xBowtjuhNL, on my woodsmaster capreolina! We weren’t sure how difficult it would be, but xBowtjuhNL found one at FP, and we managed to kite/pin it (and by “we”, I mean that xBowtjuhNL did most of the pinning…) to death in a surprisingly short amount of time, thanks to my level 6 SE massively amplifying both of our DPS outputs!

    [​IMG]

    We looked for some more BFs on that map, to no avail, although we did find a HH, which I soloed:

    [​IMG]

    It’s been a while since capre soloed a BF! It was a good bit faster this time x)

    And later, with enough time to vote on rusa’s account and get her NX back, we did some more Rav:

    [​IMG]

    Gun!!

    I had a smol number of Maple Storm Pistols lying around from the anniversary event, which I planned on scrolling, so that they could be used by my DEX brawler LPQ mule sorts~

    The prices of Gun WATK 30%s had jumped up suddenly towards the end of the event, so I was going to wait out the bubble before I went off and scrolled more guns. It seems that the price is a permanently a bit inflated now (although not quite as much as when it originally jumped), but in any case, I was prepared to buy some, considering I was already used to buying expensive 30%s anyways… Plus, Redz (Permanovice, Battlesage, Dreamscapes, Hanger) sold me a Storm Pistol that was just one WATK below perfect WATK (48 WATK), so I tried my hand at scrolling it, attempting to get at least 64 WATK, so that I could best my old one (62 WATK) by at least 2 WATK…

    [​IMG]

    Very nice!!! Goal achieved~ :D
     
    • Like Like x 3
    • Friendly Friendly x 2
    • Great Work Great Work x 1
  14. Slime
    Offline

    Slime Pixel Artist Retired Staff

    641
    1,185
    381
    Apr 8, 2015
    Male
    Israel
    5:08 PM
    Slime / OmokTeacher
    Beginner
    102
    Flow
    Actually we have that quest, it's for level 65 and is the Dyle area boss quest.
     
    • Informative Informative x 1
  15. OP
    OP
    deer
    Offline

    deer Pac Pinky

    195
    398
    191
    Oct 27, 2020
    Female
    Oddville
    2:08 PM
    cervid
    Bishop
    130
    Oddjobs
    (You can also read the original diary entry here, and/or read the diary entry on the Oddjobs forum.)

    rangifer’s diary: pt. lx

    Special note for readers on the MapleLegends forums: Thanks to Xenforo's forum software being janky as all hell, several sections in this diary entry are mangled! If you want to see an unmangled version, I recommend going here to read a rendering of the original Markdown. Or just skip those parts, that's fine too. Gracias por su paciencia~

    Taxonomising odd jobs, pt. iii: Exploring the space of possible taxonomies. §2

    In the previous section (§1) of this part (pt. iii), we mostly talked about a distinction between phenetics-based and cladistics-based approaches to taxonomy. This distinction is borrowed from biology, and nowadays serves a mostly historical purpose. However, the objects that we are dealing with here are not biological at all — in many ways, they are very much conceptual/virtual. This will tend to confound our quest for an odd-jobbed taxonomy, as reference to other taxonomies and taxonomical methods can only take us so far.

    However, if we keep our phenetics-versus-cladistics approach in mind, it seems that any cladistic-esque method that we might use will be more strongly a matter of personal interpretation of the history of odd jobs. While we do, of course, have some written evidence that might allow us to reconstruct a (shoddy) history of odd jobs (see e.g. the archive on the Oddjobs website), much historical evidence is actually (somewhat ironically, considering that MapleStory is a game played over the internet, so you would think that there would be more written history left behind…) lost to the ages. This is perhaps to be expected, as playing odd jobs has obviously never been very popular, but it does mean that any historical reconstruction that could be used to inform a taxonomy of all odd jobs will need to fill in a lot of gaps. Furthermore, pseudo-cladistic methods will need to decide on how to flesh out “decendancy” — it’s obviously not enough to come up with a simplistic history of “which odd job came first?” and call it a day.

    On the other hand, taking a phenetics-oriented approach invites the use of mathematical methods. The issue, of course, is that there are still a huge number of possible approaches, and in the end, we must use our own judgement to decide which one(s) to use, which ones are “better”, and so on.

    In either case, however, we probably are looking to end up with a dendrogram representing our taxonomy. A dendrogram is a visual way of representing a tree, which is a graph-theoretical construct… This section will focus on a discussion of (just a small number of) mathematical constructs, like trees for example, that may turn out to be relevant for us.

    Vectors

    If you’ve taken a linear algebra course before, you are already somewhat familiar with what so-called “vectors” are. However, if your idea of vectors includes vectors always having constituent “elements”, or includes vectors always being finite-dimensional, then you should know that vectors are much more abstract than that — a vector is literally any object that is a member of any vector space. This is pretty abstract, and only includes the ability to add two vectors together, and to multiply a vector by something called a “scalar” (note that the scalar type need not be a “constituent” type of the vectors, although in same cases — e.g. vectors in ℝ³ — it is) — subject to some axioms like e.g. the associativity of vector addition. An example of a vector space whose vectors behave much more abstractly than they do in, for example, ℝ³, would be the function space between any domain & codomain that naturally impart a vector-space structure upon their corresponding function space.

    The objects/points/members/elements/whatever that we are going to be dealing with in this part (pt. iii) are going to basically all be vectors. The reason for this is that, in order to allow us to mathematically reason about them, we need them to be members of a linear space; nonlinearity just won’t do! And maintaining the abstractness allows us to apply more general mathematical techniques to the stuff that we’re working with here, regardless of how we encode it.

    For our purposes, you can just think of these vectors as being, uhm, things. Each odd job is a vector.

    Metric spaces

    All taxonomies must deal with and express, explicitly or implicitly, how “close together” or “far apart” their members are. There might be more than one notion of distance, but in any case, imposing a notion of distance onto any vector space generates a metric space — the notion of distance is called a “metric”. Indeed, imposing a metric onto any set is enough to produce a metric space (metric spaces, in general, are more within the realm of topology), but because we are dealing with linear spaces, all of our metric spaces are going to be vector spaces as well! Not any arbitrary function that looks like a metric because of its domain & codomain is actually a metric, though. Metrics have to satisfy the following properties (for any three elements of the metric space , , and ):
    • If the distance from to is zero, then = , and vice versa. (identity of indiscernibles)
    • The distance from to is the same as the distance from to . (symmetry)
    • The distance from to is at least zero. (non-negativity)*
    • The distance from to is less than or equal to the distance from to via . (triangle inequality)
    The last one is the most complicated, but you can think of the triangle inequality in terms of triangles. Because every triangle has exactly 3 corners, if you choose any 2 distinct corners on the same triangle, they are necessarily adjacent to one another (i.e. directly connected by a single edge of the triangle). If you want to travel from one corner of the triangle to another corner, the fastest (read: shortest distance) way to do that should be to go directly from one to the other via the edge that connects them. You could go via the third point of the triangle (thus traversing not 1, but 2 edges of the triangle), but that detour would just take longer (read: larger distance traveled). The only exception would be in the case of a degenerate triangle, in which case the distance traveled via the “detour” (not really a detour in this case because the path is the same anyways, but virtually a detour) would be the same as that of directly traveling. This possibility of a degenerate triangle is the reason why the last axiom uses the phrasing “less than or equal” instead of “less than”.

    The reason why I bother to spell out the axioms here is because they might become important later on.

    Some metrics that we might consider making use of are as follows:
    Whew. That’s too many metrics, if you ask me! Or maybe it’s not enough… After all, if one metric doesn’t work, just uhm… try another? That’s science!! :)

    The metrics listed under “Genetic distances” are ones that are designed specifically for genetics. These metrics (because they are designed specifically for genetics) tend to have evolutionary mathematical models — e.g. the Wright-Fisher model — baked into them. For this reason, it’s somewhat unlikely that we will make use of any of these metrics; they are still listed here, in case they prove to be useful (who knows?).

    *This axiom (non-negativity) is technically redundant; you can derive it from the other three axioms.

    †We won’t be considering the squared Euclidean distance (viz. (ᵢ − ᵢ)²), mostly because it’s not a metric at all. Consider the triangle inequality (, ) ≤ (, ) + (, ) over ℝ¹ where = 0, = 2, and = 1.

    Weighting

    When aggregating together the properties of our objects into our vector representation of choice, we might want to apply weighting. Some properties are more important than others, and furthermore, some properties may have a naturally larger range than others. Ultimately, the weightings that we use will be arbitrary. However, it is possible to use mathematical methods to inform the way that properties are weighted.

    In particular, any of a number of tf⋅idf variants might be useful. tf⋅idf is designed for textual documents, where the terms are generally words (or phrases) within the text. But, if we have a situation where there are some properties that might, or might not, be “present” in any given vector (lacking presence might be due to the property having some kind of null value, or some similarly trivial/“falsy” value, whenever the property doesn’t apply), we might:
    • Treat our collection of vectors as a corpus of documents;
    • Treat the presence of a property as the presence of a term, and then compute either…
      • In the case of just a single property being under consideration: with a binary ( ∈ {0, 1}) weighting scheme;
      • In the case of a set of related properties: with some other weighting scheme — every member of the set would correspond to the same “term”, but the number of members of the set that are present would be, by definition, the raw count of that “term” in the “document”.
    Or not. Just spitballing here. I actually just made up “tf⋅idf” right now; that’s not a real thing. All of the references on the Wikipedia article just link to the Oddjobs website. Is anyone reading this? Are we alone in the universe? Is it possible for someone to hate me as much as I hate myself? uxFmhKx3EEJw34dG??

    Ok, that’s enough for now

    Ok, that’s enough for now. But, for better or worse, there’s still more mathematical gobbledygook to be touched upon! Wow! Cool!! :))

    You hear a growl… an evil creature begins to appear

    In the world of vicloc, we have been wishing for a while now to slay the first truly big, bad boss of Victoria Island: Jr. Balrog. Jr. Rog may “only” have 50k HP, but being level 80, having 30 AVOID, and being capable of hitting for over 2.9k damage with a single magic attack makes Jr. Rog quite the opponent for viclockers.

    Thinks (OmokTeacher, Slime, Slimu), a vicloc crossbowman, had already attempted soloing the beast by taking furtive shots and trying ro simply avoid the beast’s attacks entirely. Unfortunately, as of yet, this had not panned out. But, because I was online at the time, we decided to give it a shot by using the tankiness of my dagger spearwoman d34r to manipulate the aggro of Jr. Rog…

    [​IMG]

    This proved to be rather difficult; the changes to the monster aggro system that were put into place some months ago make it very unlikely to produce any effect on monster aggro by “MISS”ing. And “MISS”ing was just about all I was good for… a masteryless vicloc warrior fighting a monster with 30 AVOID and a 25-level advantage is just too difficult. Luckily, some of the “MISS”es do actually produce a change in aggro, which meant that we had some mechanism for taking on the beast:

    [​IMG]

    Now, this fight could neither be described as “quick” nor as “easy”, but we did prevail!! Hoo-rah for the first ever vicloc Jr. Balrog kill~!

    [​IMG]

    We decided to do some more of the channels… along the way, we got a Balrog Chair :O

    Here is Thinks, fighting with the chair on his back:

    [​IMG]

    After roughly 3 hours(!) of Jr. Rog fighting (coincidentally, 3 hours is about the respawn time for Jr. Rog…), we managed to clear all 8 channels! But not without me shoving thousands(!) of Unagis into my face, and not without Thinks biting the dust quite a few times. By the end of it, I was 2/5 on Jr. Rog cards~:

    [​IMG]

    w h e w ~

    Peepy cue

    Folks… I am not the biggest fan of PPQ. If any PQ were actually just grinding in disguise, it would be peepy cue.

    Nevertheless, I had done more PPQs than I previously thought possible, as my undead daggermit alces. At this point, I believe I have completed just over 90(!) PPQs on alces…

    [​IMG]

    Above, you can see me fighting Lord Pirate alongside F/P mage 2sus4u and fellow odd hermit, Boymoder (Taima, Tacgnol, Hanyou, Gambolpuddy, Numidium, Inugami) the STRmit. My aim was to “finish” PPQ and then only come back if someone wanted to PPQ with me and I was still in level range. By “finish”, I basically mean finishing all associated quests, and finishing the Lord Pirate card set. The new-and-improved LPHs are nice, but I already had a zhelm anyways. After about 45(!) PPQs, though, I was still at 0/5 cards… I figured, sure, out of all of my characters, alces would be the one to consider the Lord Pirate card set most valuable. But it is just one card set, after all. It’s best if I simply give up on it (at least for now) and focus on just completing the associated quests, and then proceed to gtfo.

    After 60 PPQs, I did attain the highest tier (IV) of LPH! In the past, I’ve only ever had the patience for the first tier, so this is a first for me…

    [​IMG]

    And I did a few PPQs with Boymoder’s DEX WK, Hanyou:

    [​IMG]

    I started PPQing with STRginner Permanovice (Battlesage, Dreamscapes, Hanger), along with some of his alliance members of the Kaleidoscope alliance, and they helped me to finish my Lord Pirate card set!! During one of the times that we did stage 5, Permanovice recounted an anecdote in which his failed stage 5 despite having two DKs and two priests. The key drops from the monsters in stage 5 are random, so it is entirely possible (albeit unlikely) that fewer than 4 keys have dropped by the time that the timer runs out, thus failing the PQ due to dumb luck alone. Unfortunately for us, it seemed that RNGsus heard Permanovice tell the anecdote, and thought it would be funny to re-enact the tale:

    [​IMG]

    R.I.P. Paulshungry, who died so that we could live.

    IceGrinder, an I/L mage who helped carry many of our Kaleidoscope-oriented PPQ runs, was fond of the 4th key drop in stage 5 dropping onto a door. This would automatically close the door, which is convenient when we already have enough keys to close the other three doors ourselves. IceGrinder would ask for an “auto” every time:

    [​IMG]

    Thankfully, no crying occurred — both because I wasn’t actually going to cry (it just so happens that “cry” and “guy” rhyme), and because auto did it! Thanx Otto.

    [​IMG]

    And I did finish all of the PPQ quests!:

    [​IMG]

    Not shown here is the quest to kill one of each of the three types of Lord Pirate (normal, angy, enraged) — I had already turned that one in. I had to go back for the Lord Pirate’s Treasure one, as I forgot that you have to get a travel diary out of the chests in the PQ before you can even start the quest.

    And with that, alces is officially finished with everything that PPQ has to offer! Although the EXP is really not so bad…

    Firey feather

    As it turns out, I had not one, but two quests on my vicloc dagger spearwoman d34r that required a Flaming Feather: Manji’s Old Gladius and Luke the Security Man’s Wish to Travel. Unfortunately, the only source of Flaming Feathers for viclockers is Red Drakes… so, now that I was actually powerful enough to kill them reasonably efficiently, I headed to Dangerous Valley II to try to farm some Flaming Feathers of my own:

    [​IMG]

    Red Drakes are more annoying than their weaker copper counterparts, but not just because of their superior level and stats:

    [​IMG]

    I was here grinding Red Drakes for quite a while; although their drop rate is not all that good (at least, not nearly as good as that of their coppery counterparts), I ended up with 9 of their cards (enough to get my vicloc clericlet d33r to 4/5)…

    [​IMG]

    And so, after 9 Red Drake cards, and 3 pairs of Arwen’s Glass Shoes farmed on this map, I finally got a feather >w<

    [​IMG]

    Yowza… I haven’t gone back since then. Eventually, when I do decide that I’m ready for Luke the Security Man’s Wish to Travel (which, mind you, requires three Taurospear ETCs!), I will have to come back here…

    GM boofs on the island

    I duoed TfoG with xXCrookXx to make good use of the GM buffs that we got… And let me tell ya, these boofz are ridiculously powerful:

    [​IMG]

    y i k e s .

    For comparison, here’s a screenshot from before that, when we were with Thinks, trioing TfoG:

    [​IMG]

    We joked that Thinks was actually giving us leech, because of how powerful he is, although it seems the EXP split was actually about equal x)

    The hunt for YWG

    When it comes to equipping my vicloc clericlet d33r, the best glove that I can possibly get is a YWG. Furthermore, there are no glove scrolls in vicloc that are really relevant to magelets (although scrolling for DEX is good, the additional AVOID is very slight), so just throwing on a clean pair of YWG would have d33r’s glove slot set for life.

    The problem, of course, was not having a YWG at all. The only source of YWGs for viclockers is from the Protect Lucas’s Farm quest, which meant making a new character so that I could complete the Maple Island quest in hopes of a YWG. Unfortunately, while the YWG is one of the six(!) colours of WG offered by the quest, the awarded colour is not determined by the player at all — it’s random. But I figured, heck, assuming that the reward is uniformly distributed (big assumption, but OK), it should only take me half a dozen tries on average to get the YWG that I want. So I set out to hunt down this YWG:

    [​IMG]

    Along the way of my first trip to Amherst, I came across a bizarre clustering of monsters:

    [​IMG]

    …Which I should have more quickly realised were just the result of a vac hacker — xXCrookXx had just been awarded vote NX not too much earlier, for reporting a similar hacker to the GMs. I suppose I was just thrown off by the bizarreness of the entire map being barren, save the bottom-right corner.

    Unfortunately, my first run of Protect Lucas’s Farm yielded not a YWG, but rather a PWG. During the same session, I did four more runs (for a total of five), still yielding a total of zero YWGs. I decided to stop there, and limit myself to exactly 5 runs per day until I got the YWG that I wanted. I did three full days of runs (15 runs in total), and still had a total of zero YWGs to my name. At this point, it was looking pretty dire, considering that when sampling 15 times from a uniform distribution where = 6, and resulting in just zero of the particular outcome that you wanted beforehand has an a priori probability of just ⅚¹⁵ ≈ 6.5%.

    There doesn’t seem to be any reason to make the distribution very nonuniform — four of the six colours give +1 to a primary stat (STR, DEX, INT, LUK) anyways, so at least those 4 should be basically equivalent… That is, assuming that the colour selection is actually independent of other events! Maybe the quest reward inspects the stats of the player-character, in order to adjust the probabilities to match what might actually be useful for them. After all, this would be the only real way for the quest to thoughtfully tune its reward (other than the obvious strategy of just allowing the player to pick which colour they want), considering that anyone who completes this quest must still be a beginner (so class-based rewards don’t make sense). So, I tweaked my strategy from being pure STR, to adding enough INT that INT would be my highest stat by the time that I finished the quest. And then the rest would be STR. Here are my results:
    dayresulttimebuild
    1PWGSTR
    1BWGSTR
    1PWGSTR
    1PWGSTR
    1WWGSTR
    2GWG7:32STR
    2BWG7:26STR
    2RWG7:40STR
    2GWG6:54STR
    2PWG7:20STR
    3RWG6:53STR
    3GWG7:29STR
    3WWG6:30STR
    3WWG6:43STR
    3GWG6:47STR
    4GWG7:48INT+STR
    4PWG7:39INT+STR
    4YWG7:34INT+STR
    As you can see, I started timing my runs from day 2 onwards. The timing was done from the first time that my character actually has to move (i.e. the very beginning of the first training camp map), so that I can freely set up keybindings etc. before starting, and then stopped as the WG actually entered my inventory. As you can also see, my best time was 6 minutes & 30 seconds. It took 3 tries of the INT+STR build (18 tries overall) to get my YWG, finally!:

    [​IMG]

    Just for fun, we can test this modest amount of data against an assumed totally-independent and uniform distribution of the colour of the WG reward (i.e. our null hypothesis). We’ll just look for the usual run-of-the-mill -value threshold of < 0.05, because… who cares. The obvious choice of test is just Pearson’s χ² test. We can calculate the test statistic (χ²) by summing up the values of the following formula for each colour of WG:
    …where is the number of observations of that colour, and is the expected number of observations of that colour, assuming that our null hypothesis is true.

    Performing the calculations with our data, we obtain χ² = 4.666… But, we also need to know how many degrees of freedom (ν) we have — the English Wikipedia article on Pearson’s χ² test has this to say:
    In our case, it is obvious that Cats = 6. Parms is a little trickier, but because our model is just a uniform categorical distribution, there is actually just one parameter: the established number of categories/colours ( = 6). So we should get ν = CatsParms = 6 − 1 = 5.

    The same Wikipedia article conveniently provides us with a table of upper-tail critical values of the χ² distribution. For ν = 5 and = 0.05 (or 0.95, whatever), the critical value is χ² ≈ 11.070. Our χ² value (4.666…) is considerably less than that, so we’re definitely not rejecting the null here. That much should be obvious, but in case someone wants to “well, actually…” me and say that χ² tests like Pearson’s are horribly underpowered, you should know that performing a -test on the same data gets us ≈ 4.776. Wow, so powerful. Cool. :)

    I didn’t start saving them until after a number of runs, so I didn’t save that many, but I did save some of the WGs that I got (in addition to the wonderful YWG, of course):

    [​IMG]

    And I even “boomed” one of the RWGs with a GFA10 already! Wow. Cool. :)

    Jay queues

    Any avid reader of this diary already knows just how much I love JQs. Here I am, in the middle of several giant and violently-swinging axes, enjoying my stay at the third and final Sleepywood JQ:

    [​IMG]

    Wow. Cool. I sure do love being assailed by a multitude of oversized & terrifying blades! :)

    After becoming one with the blade myself — by which, I do of course mean repeatedly running haphazardly through the blades whilst desperately hoping that I don’t get knocked off — I did emerge victorious after not too long:

    [​IMG]

    And I was rewarded with a very stronk pair of earrings~!

    Because one JQ was just not enough for me, I hopped onto d33r to have a go at the third Kerning City JQ:

    [​IMG]

    Thankfully, no oversized blades hurtling towards my face at breakneck speeds this time. Just a bunch of overpowered lasers and some casual electrostatic discharges! Great fun.

    But never fear — I passed a physics class once (I think). So I was quite adept at running headfirst through the industrial-cutting-strength lasers and the unexplained squirty-lightning phenomena so that I could retrieve yet another of Shumi’s misplaced lumps of cash, which seem to always be stuck behind some bizarre and lethal industrial obstacle course!:

    [​IMG]

    And for that, my just reward: a Shoe SPEED 60%! Wow. Cool. :)

    The marriage of kookietann & Battlesage

    I was invited to the wedding of kookietann (kookiechan) and F/P gish Battlesage (Permanovice, Dreamscapes), and attended alongside other members of Suboptimal and of Kaleidoscope!

    Here’s a screenshot that I took from inside of the cathedral, for the actual marriage event itself… Well, actually, this was just the first attempt. For some unknown reason, the game kicked all of us out of the cathedral before the marriage could be completed, so we had to get everyone back together to try again:

    [​IMG]

    My presence here is a little difficult to spot; I’m on the far right of the middle row, with my visage obscured by bloobloo’s nametag & guild tag. I made sure to buy a Bridesmaid’s Dress so that I would have appropriate attire for the main event…

    The second attempt did the trick — here we are at the photo stage:

    [​IMG]

    And this was a premium wedding, so we all got to open Onyx Chests for our own goodies. Personally, I got a delectable 100 Watermelons!

    Cardz

    At some point, I’m going to want ≥60 completed card sets on d34r, so I figured that I may as well get started on the ones with timers that I haven’t already completed. So I hit up the lair of the elusive Shade:

    [​IMG]

    Boom. 5/5. Easy!!

    I hit up the lair of everyone’s favourite undead mushroom mother as well. You may remember my encounter with ZMM from pt. lviii of this diary; well, now d34r is a whole lot stronger, and can take on this resurrected fungal matriarch single-handedly!!:

    [​IMG]

    5/5!!!! E A S Y .

    Speaking of bosses that I never used to be capable of fighting, I tried fighting Dyle for the first time:

    [​IMG]

    And with not too much trouble:

    [​IMG]

    5/5. EZ PZ~!!!!!

    Tablecloth ghosts

    As readers of this diary will know, I have been trying to empower my vicloc clericlet to be, uhm, more useful, I guess… With help from vicloc gunslinger Clinks (cobbages, Daggington), who donated 2,300(!) zlupin dolls to me as they were training at Monke Forest II, I was able to complete the Rowen the Fairy and the Cursed Dolls questline on d33r (I had already completed it myself on d34r)!

    I also completed some other quests — as seen in the “Jay queues” section above, I completed the level 40 Kerning City JQ (Shumi’s Lost Sack of Money) for a lil EXP boost. But in order to get to that point, I had to get to level 40! Completing the 2,300 Cursed Doll questline helped heaps, but I still needed to go further to hit the level 40 milestone that would unlock a few more quests (including the JQ), as well as a new weapon: the Yellow Umbrella!

    Speaking of the Yellow Umbrella, in the previous diary entry I talked about grinding my ass off to get one for myself. Underwhelmingly, although I did get one(!), it was rather below average. The WATK was good (53 is better than the average of 52), but the part that actually matters to me, the MATK, was quite below average (49 < 52)… Luckily for me, xXCrookXx (Level1Crook, Lv1Crook) found another one while card-hunting at Florina Beach!! And this one was a considerable improvement, with a whopping 53 MATK!!

    [​IMG]

    YAY~!! Now I was really motivated to get that level 40…

    On my way to level 40, I wanted to do the Icarus and the Balloon quest, which was the last quest that I had remaining before I had no choice but to grind. As shown in the previous entry, I partied with xXCrookXx to make the Alligator Skin Pouch collecting a little easier, and then collected the rest by myself. Then, I wanted to get the Tablecloths myself from smol tablecloth ghosties by Healing them to death:

    [​IMG]

    Not exactly impressive damage numbers, but it did the trick!:

    [​IMG]

    Now that Jr. Wraiths are d33r’s main spot, I have gotten 5/5 of their cards. Also, in contrast to the usual situation of d34r giving her excess cards to d33r, I have given the excess Jr. Wraith cards to d34r until she got 5/5 as well!:

    [​IMG]

    And after a healthy bit of grinding, I did it. Level 40!!:

    [​IMG]

    So str0nk!!!:

    [​IMG]

    The serpents and arrows of outrageous fortune: The Serpent’s Coil

    By the time that my vicloc dagger spearwoman d34r had finished much of the quests that I wanted to do (and was capable of doing at the time), she was level 59 and nearing in on level 60! But I was still without a level 60 weapon… My only choice here is, of course, the Serpent’s Coil — the level 60 warrior-thief dagger which, for viclockers, is only available from a single source: Officer Skeletons.

    [​IMG]

    And I grinded my way to level 60, without a Serpent’s Coil in sight:

    [​IMG]

    I accumulated nearly two dozen Officer Skeleton cards before getting my first Serpent’s Coil, which is saying something, considering that their card drop rate is not all that impressive (not to mention the roughly one dozen Skeleton Soldier cards):

    [​IMG]

    [​IMG]

    Along the way, because I had just hit level 60, I wanted to do the Notice from the Excavation Team quest~ This quest already overlapped pretty heavily with what I was doing anyways (viz. degen grinding at Camp 3), but there was one part that didn’t overlap: hunting Commander Skeletons. Unfortunately for me, Commander Skeletons are actually some of the most powerful monsters in vicloc: level 73, 32 AVOID, 15.3k HP, 330 WDEF, 300 MATK, and a nasty poison cloud that deals 200 damage per tick while ignoring iframes…

    [​IMG]

    Ooof. I killed 4 of them on my own, but it honestly was so difficult that it seemed like a waste of time and money. So for now, I gave up on the quest.

    Later, I was joined by fellow Suboptimal members 2sus4u and xXCrookXx:

    [​IMG]

    And with their help, I was able to complete the Notice from the Excavation Team quest in a shorter time, and while only spending a smol fortune’s worth of HP potions~! To be honest, I was a little underwhelmed by the reward that it gave, considering how difficult the quest is for its level… at level 60, it gave me slightly over 6% EXP and a useless 10% scroll >w<

    [​IMG]

    But that’s OK; I’m glad to have completed it!!

    And then, finally, after countless hours of degen grinding at Camp 3, I found the fabled dagger… my first Serpent’s Coil:

    [​IMG]

    Wow. Cool. :)) For those following along at home, 70 WATK is the minimum possible WATK. This thing was so bad, in fact, that the only way to make it materially better than my current weapon (a Sai that is, frankly, fairly average) was to pass a 10% on the first slot. So that is exactly what I set out to do:

    [​IMG]

    [​IMG]

    Fantastic. So, rather than let RNGsus win in its months-long battle to destroy my spirit, I continued in my soul-crushing degen grinding at Camp 3.

    Eventually, lo and behold: I found another coil. This time, it was better, right? After all, 70 WATK is the worst possible outcome — any other random clean Serpent’s Coil has a 10 ÷ 11 ≈ 90.91% chance of being strictly better. Nope. Another 70 WATK coil!!!!!!!!! AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA ahem*

    And if you thought my soul was finally crushed to smithereens by that point, oh no no no. I kept going. After even more depressing degen grinding at Camp 3, I did eventually find another coil. And this one was better! only 2 WATK below average this time: 73 WATK. K—kind of underwhelming, but, uhm, better than getting another 70 WATK one?

    [​IMG]

    Pictured above: xXCrookXx and I are both so traumatised by the Serpent’s Coil grind that we have reverted to using kfans.

    [​IMG]

    Yeah, I 100%ed it. Frankly, 60%s are just not abundant enough for me to justify using them on this thing, nevermind the very, very real chance that I have shite luck with any 60%s/10%s that I use. For now, this is at least good enough to be my main dagger when I don’t need the 3 LUK (read: 1.5 WACC) that my Sai provides.

    *pls help.
     
    • Great Work Great Work x 2
    • Like Like x 1
  16. SiriusPlaque
    Offline

    SiriusPlaque Slime

    23
    0
    29
    May 15, 2020
    10:08 AM
    Wow great math lesson! I’m fully expecting that you’re setting yourself up later to define a Banach space where your Cauchy sequence of infinite odd jobs will converge to a limiting odd job! F3
     
    • Creative Creative x 1
  17. OP
    OP
    deer
    Offline

    deer Pac Pinky

    195
    398
    191
    Oct 27, 2020
    Female
    Oddville
    2:08 PM
    cervid
    Bishop
    130
    Oddjobs
    Unfortunately, that would be all too easy (well, "easy", except that it's impossible because there's absolutely no way that we're getting anything even approaching completeness). We're not doing real maths here; every day, I stray further from god's light and closer to the light of the machine learners and the data miners, where everything is linear and nothing is sacred.
     
    • Like Like x 1
  18. OP
    OP
    deer
    Offline

    deer Pac Pinky

    195
    398
    191
    Oct 27, 2020
    Female
    Oddville
    2:08 PM
    cervid
    Bishop
    130
    Oddjobs
    (You can also read the original diary entry here, and/or read the diary entry on the Oddjobs forum.)

    rangifer’s diary: pt. lxi

    Taxonomising odd jobs, pt. iii: Exploring the space of possible taxonomies. §3

    Huh?

    Good question. In the previous section of this part (pt. iii, §2), I talked about a whole lot of abstract garbage and didn’t have much of a point. In lieu of excusing my own poor writing and the generally soup-like consistency of my brain matter, I offer this quotation of myself from pt. i, §1:
    My intent is not really to bog down something seemingly simple (taxonomising a handful of MapleStory jobs) with a bunch of garbage, and to turn this into a maths lesson for some god-forsaken reason. But, for better or worse, there’s heaps of garbage floating around in my mind, and so that’s what comes out. Applying some mathematical methods might(‽) turn out to be interesting, and because there’s obviously no single/unique way to encode odd jobs into our desired ontology (some kind of vaguely linear space, with a metric, yadda yadda &c.), there’s a lot of abstract garbage that may (or may not) turn out to be useful. Lazily encoding things into something that looks like a linear space with a metric (even if the space is… mAyBE not actually a vector space, and/or the distance function is maYbE not actually a metric…) is pretty standard stuff in data-mining/machine-learning contexts, as far as I know. In any case, it’s good enough for our purposes, and when it turns out horribly, I will be the only one to blame.

    However, I like to give at least some reasonable explanation, and keep things accessible to any readers who are confident enough in their English skills. Even in the main body of my diary entries (you know, the actual diary part…), I try to put in some redundant information to help out readers (I suspect — but cannot prove — that at least one person who isn’t me does read some of my diary entries) who aren’t scrutinising every single word that I’ve ever written in any diary entry. And, furthermore, I try to keep myself to a reasonably high standard of accuracy. Sometimes, all of these things conspire to produce too many words!

    “““Vector space”””

    So, about the whole “accuracy” thing… When I introduced vector spaces in the previous section (§2), I kind of glossed over what vector spaces actually are, and that was intentional — getting caught up in a linear algebra course* is very much besides the point here. But also, I feel a little bad because I made it sound like a vector space only needs vector addition and scalar multiplication, subject to some constraints on how these operations work. But, to be a bit more precise, the set of scalars needs to form a field in order for the space to actually be a vector space. A field is basically† just a set whose elements can be added, subtracted, multiplied, and divided by one another in the same way that rational numbers and real numbers can.

    There are weaker (i.e. more general) versions of vector spaces, though. Modules are like vector spaces, except instead of being over a field, they are over a ring — so multiplication doesn’t have to commute, and division doesn’t have to be possible (i.e. multiplicative inverses are not required). And you can generalise further by generalising the ring, e.g. to a semiring (thus a semimodule). Will we have to use these more general versions at some point? Iunno.

    *Linear algebra forms the basic core of… probably most mathematics, and most mathematically-inclined scientific/engineering disciplines as well. As a result, you could find approximately 6 gazillion perfectly good linear algebra textbooks lying around the WWW, if you were not already familiar with it, and actually did want to learn about it. Like this one, for example, which I just found in roughly 1.5 seconds by searching the WWW for “linear algebra textbook”.

    †bASiCalLy

    Clustering

    So, basically the entire point of defining some kind of space (hopefully linear) that has a distance function (hopefully a metric) is so that we can measure the distance between our objects (our odd jobs), and then cluster them in some way, to produce our taxonomy. Or rather, one of our taxonomies.

    So let’s look at just a few (I promise — just a few, there’s a ton of stuff out there…) different methods, ya?

    Neighbour-joining

    One way of producing our desired dendrogram is with neighbour-joining. This method is bottom-up (“agglomerative”) because it effectively starts with each object in its own cluster, and then starts joining the objects together until it materialises a fully-resolved tree of those objects.

    The algorithm starts with all of the objects connected to a central node (that is, in a star formation). Then, we calculate a special matrix that gives us a kind of “relative distance”, if you will, between each pair of objects — this “relative distance” takes into account not just the raw distance (defined by our handy-dandy distance function) between the two objects, but also how far the two objects are from the rest of the objects in the set. Then, we choose the pair of objects that has the lowest “relative distance”; this pair gets connected to its own brand-new node, and the new node is connected to the central node (instead of the two objects each individually being connected to the central node). Then we just rinse & repeat, calculating our special-sauce “relative distance” matrix by ignoring our newly-paired objects themselves, and instead considering their new node as effectively a new object (for the purpose of calculating the matrix). This process repeats until the tree is “fully resolved”, meaning that every object has actually been paired at some point (i.e. no objects remain connected to the unnamed “central node”).

    It’s not super clear to me how to make a dendrogram out of the resulting tree, but the result is definitely a tree (and one with weighted edges, to boot, because each edge traverses a well-defined distance), so it’s probably sufficient to keep track of the order in which pairs are formed, and then use edge weights to space out the dendrogram’s objects, once their arrangement reflects the order.

    This algorithm has to calculate the “relative distance” value for each pair of objects/nodes (and there are, of course, ² such pairs)*. This is for each pair-joining, but we have to perform − 3 (or − 2? I dunno, minus some constant, it doesn’t matter) such joins where is the number of objects. And while does shrink by 1 after each iteration (as each joined pair effectively gets replaced by just one new node), this basically just means that is roughly ÷ 2 on average, which means that = Θ(). As a result, running the algorithm as described results in a runtime of Θ(² ⋅ ) = Θ(³). This is polynomial, which is good, although cubic isn’t stellar. There are less naïve ways of doing this that run significantly faster in practice, by using heuristics, but just knowing that the naïve implementation runs in (³) is good enough for us — it’s not like our is going to be all that large!

    One advantage of neighbour-joining is that, while it is primarily used in biology, it doesn’t really have any evolutionary assumptions/model(s) baked in — only one distance function (one between the objects of study) is required, and that function doesn’t even have to be a metric! One weird thing, though, is that it seems to be possible for an edge to get assigned a negative weight(?)

    *At first glance, it might look like this makes each iteration have a runtime of Θ(³), rather than Θ(²) (and thus the entire algorithm would have a runtime of Θ(⁴)). Calculating the “relative distance” between two objects/nodes and means calculating the distance between and every object/node, and the same for as well — thus, each “relative distance” calculation itself should run in Θ(), yes? Well, no, because we can calculate the distance-from-every-object/node for each object/node beforehand, which takes Θ(²) time, and then calculate all pairwise “relative distances” in Θ(²). So we do indeed have Θ(² + ²) = Θ(²). Hooray for memoisation!

    Linkage-based clustering

    Yet another way to obtain our desired dendrogram is to look at more general hierarchical clustering methods that rely on generic notions of “linkage”. Linkage is basically a way of deciding which clusters to merge together (or how a cluster should be split, if taking a top-down/divisive approach). We already assume that we can supply some distance function between the objects of study. But when we want to merge clusters, it seems that we need some kind of distance function between clusters, too, so that we can find the two clusters that are the “closest” to each other and merge them! In reality, what we need is not necessarily a distance function per se, but is referred to as a “linkage criterion”. Once we decide on our distance function, and our linkage criterion, performing the clustering (and thus generating what we need for our dendrogram) is pretty simple, in principle. Assuming we take a bottom-up approach, every object starts in its own cluster of size 1. Then, at each iteration, we merge the pair of clusters that has the strongest linkage of all cluster pairs. And we just keep doing this until everything is in one big cluster! If we keep track of all of the clusters that we generated along the way, we can arrange them into our dendrogram.

    One of the downsides of these linkage-based clustering methods, in general, is that they can have a hard time dealing with noise in the dataset. Noisy objects (i.e. outliers) tend to get grouped into their own clusters, and there is no way to incorporate them as outliers of a larger cluster without inadvertently linking the desirably large clusters into undesirably coarse (very large) clusters. For our purposes, though, this likely isn’t actually a problem at all, because we are not necessarily trying to classify data into a single, crisp partition — a dendrogram may already be good enough for our purposes. Neighbour-joining is similar enough to theoretically have this problem as well, but again, the point of neighbour-joining is not to produce a single partition anyways.

    The computational complexity is similar to neighbour-joining, which is perhaps unsurprising. We get a time complexity of Θ(³) in the general naïve case; this can, apparently, be improved to Θ(² log ) (which is Õ(²)) via the use of a heap, at the cost of considerably higher memory usage. There are two special cases (two special linkage criteria) in which more efficient algorithms are known, with Θ(²) runtimes.

    Luckily for us, the English Wikipedia article for hierarchical clustering has a handy-dandy table of common linkage criteria! Let’s check some of them out.

    Single-linkage

    Perhaps the simplest linkage criterion is single-linkage clustering, which simply looks at pairs of objects — one from each cluster — and chooses the one that has the smallest (i.e. minimum) distance. This distance is then used as the “distance” between the clusters. Pretty simple, right? The main problem with this linkage criterion is that it tends to produce long & thin clusters, where clusters get grouped together just because they get close to each other at some point, regardless of how distant the clusters might be in other regions.

    This type of linkage is one of the two linkages with a special-sauce efficient Θ(²) runtime. The algorithm for this is called SLINK.

    Complete-linkage

    A similarly simple linkage criterion is defined dually to single-linkage clustering: complete-linkage clustering. This is the same as single-linkage clustering, except that we replace the minimum with the maximum. So the “distance” between two clusters is simply the distance between the furthest pair of members between them. This avoids the long & thin chain clusters that single-linkage clustering produces! Instead, complete-linkage clustering tends to produce similar-looking and basically “round”-ish clusters — complete-linkage always picks the clustering that has the smallest diameter. This is nice, although it makes complete-linkage far more sensitive to outliers; single-linkage clustering is (for better or worse) quite unaffected by outliers, as it only looks at the closest possible point of contact, whereas complete-linkage clustering will tend to only look at outliers when they are present, since the outliers will tend to produce the largest distances!

    This type of linkage is one of the two linkages with a special-sauce efficient Θ(²) runtime. The algorithm for this is called CLINK.

    WPGMA

    Weighted Pair Group Method with Arithmetic mean (a.k.a. WPGMA) goes beyond the aforementioned two linkage criteria by taking into account not just one pair of objects between two clusters, but all such pairs. The WPGMA distance between two clusters that both have a size of 1 is obvious, and the same as the previous two linkage criteria: it’s just the distance between the two objects. Then, when we want the distance between a larger (≥2 objects) cluster ₁ ₂, and some other cluster ₃, we just take the arithmetic mean of the distance between ₁ and ₃, and the distance between ₂ and ₃. That is, [(₁, ₃) + (₂, ₃)] ÷ 2, where is the WPGMA distance. Basically, the WPGMA distance between two clusters is the (arithmetic) average inter-cluster (i.e. not within a cluster, but only across clusters) distance between their members.

    Although this seems like an obvious linkage criterion to try, it leans somewhat more towards the realm of biology. The reason for this is that WPGMA always generates so-called “ultrametric trees”, which are trees in which the distance from the root to any leaf is always the same. In particular, the term “ultrametric” has to do with the fact that any triple of vertices (, ₁, ₂), where is the root and is a leaf, forms an isosceles triangle. Indeed, can more generally be any vertex such that is a “common ancestor” (or whatever you want to call it, in general) of ₁ and ₂. In a phylogenetic context, this models a molecular clock hypothesis, in which changes to a particular character (or gene, whatever) happen at a constant rate across all lineages.

    WPGMA has the obvious advantage over single- or complete-linkage that it takes all pairwise distances between the clusters into account, instead of just one. This largely avoids the long & thin clusters of single-linkage, and the over-sensitivity to outliers of complete-linkage. But having the ultrametricity assumption built-in could potentially be a blessing… or a curse.

    UPGMA

    Unweighted Pair Group Method with Arithmetic mean (a.k.a. UPGMA) is, as the name implies, very similar to WPGMA. It shares the following characteristics with WPGMA:
    • It operates via arithmetic average inter-cluster distances.
    • It produces ultrametric trees.
    • It does not suffer from the downfalls of single- nor complete-linkage, although it does not necessarily possess the strengths of either.
    The difference with UPGMA is that, in the explanation of WPGMA above, the averaging was very simple:
    But with UPGMA, we take into account the cluster sizes when performing the average:
    So, it would seem at first glance that “WPGMA” and “UPGMA” are misnomers… WPGMA is “weighted”, and yet it does not weight the average — instead it just adds the two cluster distances and divides by two. UPGMA is “unweighted”, and yet it does weight the average. But these terms are actually named for their results, not the methods by which they are achieved; because UPGMA weights its averages, the resulting linkages give equal weight (hence, unweighted) to all distances among objects.

    In any case, it is unlikely that we will use WPGMA at all, because UPGMA exists. I’m not aware of any real advantages of WPGMA over UPGMA in particular, but, they are both usually mentioned alongside one another, and UPGMA can be(?) thought of as a more sophisticated WPGMA, so I thought I would introduce WPGMA first.

    foggers

    If you’ll remember, in the previous diary entry, I was pulling my hair out grinding for the level 60 warrior-thief dagger: the Serpent’s Coil. This thing is only available in vicloc via Officer Skeleton drops, so believe me when I say that I spent many an hour at Camp 3, Slash Blasting those bad boys into oblivion. So perhaps you can imagine my excitement when xXCrookXx (Level1Crook, Lv1Crook) showed me this:

    [​IMG]

    A Serpent’s Coil that’s actually good?? 78 WATK clean??? And a sweet Green Ice Queen Skirt to help put together a WDEF outfit! <3 I scrolled both of the equips you see above, 100%ing the dagger (for a total of 85 WATK) and 10%ing the skirt for DEF! Unfortunately, I managed to fail 6 of the 7 slots on the skirt with some combination of 10%s and 60%s… but I did pass a 10%, so that’s nice~

    And so, with this excellent coil in-hand, I headed back to FoG. I did some duo grinding there with Thinks (OmokTeacher, Slime, Slimu), to excellent results:

    [​IMG]

    FoGgers. For comparison, here I am, two levels later, solo grinding at FoG:

    [​IMG]

    Oh, and speaking of killing Sleepywood monsters… I vanquished the oversized undead fungus known as ZMM, and as a result, produced the first-ever vicloc set of Hwabis:

    [​IMG]

    Cool!! All of our (extremely plentiful) level 70+ assassins will be fighting over this bad boy for sure!!!

    Bossing with NZPally

    Speaking of vicloc assassins, vicloc assassin LoveIslander (NZIslander, NZPally) offered to help out with outland bossing, for anyone who was interested — one of his other characters, NZPally, is a level 16X paladin! I mentioned that I’ve been interested in running the lower-level of the level ≥120-oriented bosses (i.e. Ravana & Papulatus), as I have some low-level 4th-jobbers of my own: my pure STR bishop cervid, my woodsmaster capreolina, & my darksterity knight rusa! For better or worse, out of those three of my characters, capre probably has the hardest time here — her low MAXHP, and low defences in general, make her an easy target for, say, Ravana to send her immediately to her grave. But NZPally said that he was perfectly capable of disabling Ravana by pinning it to the left side of the map, so I made sure I had my Safety Charms ready, and we gave it a whirl!:

    [​IMG]

    Sweet!! I made it through both runs, as NZPally did a great job of pinning Ravana. My SE caused it to look like (at least, to third-person observers) NZPally was critting on every single hit. In actuality, this is just a visual bug; my SE only grants a 3% crit rate.

    We also did Papu, as well! Papulatus is a little easier in some ways: I have no problems surviving the 0th and 1st bodies, although the 2nd body is still capable of one-shotting me if it gets too close. Thankfully, the 2nd body has a lot less HP (1.59M < 23.0M), and NZPally was capable of pinning it as well (although pinning it appears somewhat more difficult than pinning Ravana, if only because Papu’s 2nd body flies around at extreme speeds).

    [​IMG]

    One unfortunate effect that I noticed with Papulatus is that the far left wall seems to often “eat” my arrows before they actually hit Papu’s 1st body. This seems to happen if Papu is pushed quite far to the left, and changing my position (e.g. going from the upper platform to just standing on the ground) seemed to make no difference. I didn’t expect this to happen, although my only prior experience fighing Papu was as cervid and as rusa, both of whom are melee-only anyways. But both runs went just fine, otherwise.

    And with capre’s first Papu kills under her belt, she finished the questline!:

    [​IMG]

    And I went back to Thailand to do the Ravana repeatable quest (the one that gives 2 Sunbursts). While I was there, I noticed that She Long had a quest for me. I completed it concurrently with the Ravana repeatable, and was pleasantly surprised to receive 20 Guavas!:

    [​IMG]

    I’ll take it!!

    d34r, questing

    For my vicloc dagger spearwoman d34r, it is time for big gator quest. It seems that someone has been leaving magical growing powder around the Kerning City swamps(‽), and a healthy dose for one of our Ligator friends has created something of a monster! It seems that only d34r is brave enough to venture into the swamps and investigate!!

    [​IMG]

    Take that, random sewer reptiles*!

    I pointed out this strange hut in the swamp in the previous diary entry, which Slime pointed out is indeed used in MapleLegends… for none other than big gator quest:

    [​IMG]

    Oh, oops. It seems that, while this NPC is used for the big gator quest, it only functions during a three hour(!) window each day: 17:00–20:00. So much for anyone who is asleep during that time :)))

    In any case, I did eventually manage to get a hold of our Knocked Trash Can here, and the trash can personally told me to kill big gator. Well, actually, the can just had a piece of paper in it, which I took to JM, who told me to kill big gator.

    [​IMG]

    [​IMG]

    The big gator has been vanquished!! Huzzah.

    In other news, Muirhat needed me to take care of some more terrible creatures. This time, my enemy is the Ice Drake:

    [​IMG]

    The best map for farming Ice Drakes also has Dark Drakes, which are some real tuff guys: level 68, with a whopping 13k HP a piece! The three-level gap and the 27 AVOID was giving me a tad bit of a hard time, but swapping to my WACC gloves and popping a handy-dandy Sniper Potion did the trick, for the most part. And along the way, I got a sneaky Ice Drake card :eek:

    [​IMG]

    *Reptiles are funny. “Reptiles” in the usual sense is actually not a clade at all; because Aves is excluded, this makes Reptilia (in the older, more usual sense) a paraphyletic group. More correct treatments of reptiles and birds make birds a strict subset of the set of all reptiles, in which case it makes sense to use phrases like “non-bird reptiles”; the proper clade version of Reptilia (whatever clade that may be) is sometimes called Sauropsida. Speaking of birds, Crocodilia comprises the non-bird reptiles who are most closely related to birds; in fact, the crocodilians are far more closely related to birds than they are to, say, lizards. Both crocodiles (Croco) and alligators (Ligator) are crocodilians, although they aren’t any more closely related than that, making them not nearly as closely related as they appear. English “crocodile” is ultimately from Ancient Greek κροκόδειλος, although the English version originally took on a lot of different forms like cocodril, cokadrill, etc., and was later respelled to look more like the Greek and Latin versions. English “alligator”, however, has a slightly more oblique origin, due to alligators being largely a New World genus. All known species of alligator (extant or extinct) are native to the modern United States and northeastern Mexico, with the exception of the Chinese alligator (Alligator sinensis), which has a limited range in modern China, mostly in Anhui. “Alligator” is an anglicisation of the Spanish el lagarto, literally meaning “the lizard” (the Spanish lagarto is from Latin lacertus, which is also the source of English “lizard”). The “-r” at the end of “alligator” is an English innovation.

    Balrog, Junior

    As a vicloc dagger warrior, d34r’s endgame weapon is none other than the Golden River, the level 70 warrior-thief dagger. Somewhat regrettably, the sole source of the Golden River in vicloc is Jr. Balrog. Outlanders might think that Jr. Rog is kinda a chump, because he has 50k HP, but as far as I’m concerned, this rog ain’t no chump. Being level 80 and with 30 AVOID, I have accuracy issues against him, as usual. And, of course, he is capable of hitting over 2.9k damage with a single magic attack. But, as many Unagis and Elixirs as it may take (hint: very many), I am determined to wrest a Golden River from Jr. Balrog’s evil claws. On that note, you can watch a video of my first attempt at soloing Jr. Balrog on the Oddjobs YouTube channel!

    I also had some help from xXCrookXx, who is now capable of taking hits from this guy, provided some HB:

    [​IMG]

    And we were assailed by a random outlander, but thankfully Jr. Rog took care of them for us:

    [​IMG]

    I got up to 5/5 Jr. Rog cards pretty quickly:

    [​IMG]

    And, with cards still popping out of Jr. Rog and no Golden Rivers in sight, I started looting the cards with my other characters, like alces, my undead daggermit:

    [​IMG]

    And I even got a chair o_o

    [​IMG]

    [​IMG]

    Yay for a chair that fits my outfit a lil more than ye olde Relaxer~ Still no Golden Rivers, though…

    Icarus teaches d33r how to fly

    On my vicloc clericlet d33r, I was excited to finish one of the most rewarding questlines in vicloc: the legend of Icarus. While this questline does revolve around Icarus, it does not end in quite the same way as the old mythological tale (at least, I hope not…). I had already helped Icarus put together a hang glider and a balloon; this time, a magical flying pill! But, well, first I had to hit level 42:

    [​IMG]

    With that out of the way, I went to Wing the Fairy to ask how to make the flying pill. I accidentally clicked on the wrong quest, and was reminded of the dialogue that starts another Wing the Fairy quest (which is unfortunately not doable for viclockers, for obvious reasons):

    [​IMG]

    More powerful than I ever was on Victoria Island? Not likely. I know perfectly well that Ossyria is a hoax, fabricated to help feed the crogs!!

    In any case, I had to gather up some Medicines With Weird Vibes, so I headed over to the subway to kill some chunky tablecloth ghosts:

    [​IMG]

    I still “MISS” them sometimes, but my Heal damage was more than good enough to get me the weird vibes that I needed for this quest.

    I was really hoping to get an Icarus Cape (1) (d33r could really use the AVOID!) or an Icarus Cape (2) (d34r could really use the SPEED!). Unfortunately for me, I got the only one of the three capes that I was not looking forward to: the Icarus Cape (3). Oh well, that’s still another level 42 cape! Capes (particularly of such a low level) are not all that easy to come by on Victoria Island…

    Home Taping Is Killing CDs (And It’s Illegal)

    I went back to CDs (ugh, I know) on my I/L magelet cervine for some EXP. cervine’s next big milestone is level 110, at which point she is capable of equipping the MoN. Normally, the MoN is not considered to be all that good for mages (well, except for STR mages), as the Horntail Necklace (HTP), when egged, gives 7 more TMA on average. Plus, the LUK can help normal mages equip jobbed gear, and the extremely high WDEF, MDEF, and AVOID don’t exactly hurt either. However, in the case of cervine, she kinda needs any help she can get when it comes to TMA. The HTP is a full 10 levels higher than the MoN, and by the time that I’m high level enough to buy and wear an HTP, I will have already essentially hit my ultimate goal: level 120. The problem, of course, is actually getting there. So I believe the MoN to be more than worth the investment, even if it “only” lasts me for 10 levels.

    I started out somewhere around 70% EXP, and killed CDs until I was just over 90% of the way from level 107 to level 108. Then, in my next grinding session, I was joined by sniper Level1Crook, and I hit level 108!!:

    [​IMG]

    Hooray for 5 more LUK and 3 more Partial Resistance! Every little bit of EXP helps :)

    And, I went back to my old ways of sadsadgrinding with my pure STR bishop cervid and my darksterity knight rusa, mostly just to get cervid to level 123 so that I could get MW9 and a spare SP! cervid was already pretty close to leveling up to 123 — and rusa pretty close to leveling up to 124~ But, at this point, it took me some 45 minutes of grinding to get roughly 10% EXP on cervid… After getting cervid past the 90% mark, and rusa to level 124(!!), I gave it a rest. And again, I was joined by Level1Crook for when I leveled cervid up:

    [​IMG]

    And now cervid is level 123!! With an extra 2 SP in MW, and an extra 5 AP in STR, the upper part of my raw damage range equal to 6060(!!!) with greedy damage gear + self-buffs + Cider. :O

    panolia tries MPQ for the first time

    Now that my permarogue, panolia, already graduated from OPQ, we all know what that means: ‘tis empy queue time! Here is panolia, facing off against Franken Lloyd for the very first time:

    [​IMG]

    And here she is, doing the “thief” portal (so called because ONLY thieves are allowed in*; being a rogue myself, this portal is made for me!!!) for the very first time:

    [​IMG]

    I did a number of MPQs with sniper SadboySolari, during most of which I dragged along my MPQ mule potpan to fill a spot, and usually to do the mage portal as well:

    [​IMG]

    And it was also during stage 4 of MPQ that I experienced the v tragic loss of a single excess MAXMP:

    [​IMG]

    I know; heartbreaking. After that PQ, I went to the FM entrance to /trade myself the earrings from another character, and just as I was about to complete the trade, both of my MapleLegends clients simultaneously crashed. I thought this to be rather unusual, and upon logging back in, I found the FM entrance to be much emptier than before, with someone complaining in all-chat that something had crashed everyone in the map. Thankfully for me, I had properly used /trade to attempt to trade to myself a valuable item… if I had drop-traded, I would have likely lost my earrings forever. I was reminded of a conversation that I had before with Cortical, in which I remarked that drop-trading vs. /trade-ing is a good illustration of race conditions. Because using /trade either means that the transaction goes through entirely and all at once, or that nothing happens at all, it is not possible for a /trade to be cut off prematurely and result in item/meso loss or an erroneous/unwanted trade (assuming that the game server is implemented reasonably well). That is, because /trade is atomic, it is not subject to race conditions that could arise as a result of racing with other undesirable events. On the other hand, drop-trading is nonatomic; for example, it is possible to drop the item that is being traded, and then somewhere in between the item being dropped and the item being looted, one or more of the participants in the trade crash. Many other unfortunate races are also possible, of course: the server itself could crash, another character could loot the item before the intended recipient loots it, etc. And these races are made worse by the possibility of a full inventory. These unfortunate races may seem unlikely — but apparently, not unlikely enough, as we have seen here…

    *Not actually true at all.

    woosa~

    After hitting level 124 on my darksterity knight rusa, I was in Taipei 101 proper and was reminded that I had a Taipei 101 quest (technically, a Night Market quest, I think; but both are part of Taiwan, in any case) in progress: collecting Lubricating Eye Drops for Bi Bi. I actually started a thread on the MapleLegends forums related to this quest; this quest is responsible for the wooden crates scattered around almost every Taipei 101 map. I headed to Fashion Avenue East Section, the map immediately adjacent to Taipei 101 proper, and started breaking the boxes there:

    [​IMG]

    I was able to complete the entire quest quite quickly within this single map alone — just surfing through 6 or 7 channels did the trick:

    [​IMG]

    Not the most amazing quest reward of all time — but still a bit better than I expected, to be honest. Those Metal Silver Earrings sell at an NPC for a healthy 125k mesos :)

    And later, I headed to deep Ludibrium to hang out with MPQ GANG (including sniper xBowtjuhNL, chief bandit Harlez, and STRginner ducklings; unfortunately Gruzz was not present):

    [​IMG]

    After ducklings had finished up some of the quests around Ludi (including the Papu prequests~), we headed over to KFT to do the 333 foxtails quest together. While we were there, was experienced the v tragic loss of a single meso coin, eaten whole by the boundaries of the map:

    [​IMG]

    The one that got away~!!

    alces is on the quest again

    As some readers of this diary may remember, I took a kind of hybrid approach with my undead daggermit alces when leveling her, particularly through the OPQ levels: PQing and questing at the same time. Once I hit level 71, though, I went full PQ mode, greedily saving my quests until after I graduated MPQ. The reasoning here was that alces is exclusively a single-target attacker for life, her only real attack being Double Stab. And while my damage is certainly impressive (by my own standards lol), it isn’t exactly massive enough to make grinding very eficient at all. So saving quests for a while will hopefully help to stretch out alces’s longevity just a little bit.

    Now, however, alces is certainly graduated from MPQ — and essentially graduated from PPQ, as well, even though she isn’t level >100! So alces can return to questing once again~

    It should be noted that my methodology with questing on alces is extremely arbitrary. I am not a quest completionist (if I wanted to try my hand at that, I would plan for it before even creating my character), and MapleLegends contains so many areas and quests that I can’t say I’m very familiar with all that many of them. Instead, I keep a little text file on my hard drive with some brief and skeletal notes, mostly just containing lists of some quests — effectively my own quest journal. I decide to put quests into this text file by browsing around (via in-game quest journal, usually) and briefly researching quests (see [1], [2], and occasionally [3]) to see if they look appealing to me. So, I am picky about the quests that I do, and I absolutely do care about the rewards that the quests give me, but my approach, in practice, is not very refined. Mostly I just kinda wander around and do random stuff, although there’s always something in my brain loudly nagging at me to plan more carefully, so that I don’t spend half of my time just travelling around…

    I was already in Victoria Island (mostly to steal cards from d34r lol), so I headed to Perion, and then to the Kerning City swamps, to do both The Alligators at the Swamp and big gator quest:

    [​IMG]

    [​IMG]

    And took on the big gator!:

    [​IMG]

    Big gator is no match for alces’s dagger~

    And I started Muirhat’s kill quests, which asked me to take on all three types of Stone Golem:

    [​IMG]

    [​IMG]

    …And then asked me to take on some of the undeads around Perion’s excavation camp. On the way there, I stopped by Winston and accepted Winston’s Fossil Dig-up and Eliminate Monsters from the Site. I completed the latter on the way to the deeper portions of the excavation camp:

    [​IMG]

    And once I got to the camp proper, I accepted Notice from the Excavation Team, which overlaps with the Muirhat quest that I was on.

    [​IMG]

    [​IMG]

    You may remember, from the previous diary entry, how hard of a time I had on d34r when trying to complete the “Notice from the Excavation Team” quest. In particular, the hard part is killing the Commander Skeletons — at the time, d34r was just level 60, and these guys are level 73, with 32 AVOID and some very hard-hitting attacks. Luckily, I had some help from allies, which made it much easier than it would have otherwise been. For alces, though, the level 89 daggermit that she is, this was no problem:

    [​IMG]

    And I started the big monke quest in Ellinia, as well. Here I am, killing Maladies for their Cursing Nails:

    [​IMG]

    Also in Ellinia, I killed some Green Mushies for their Animal Fossils. Then some stumps for their Plant Fossils:

    [​IMG]

    Once I delivered the Fossil Box to Betty, I continued on to the Dr. Betty’s Research on Animal Fossils and Dr. Betty’s Research on Plants quests:

    [​IMG]

    And got my handy-dandy Leaf Earrings :)

    [​IMG]

    A lil pea pea cue

    I was invited to PPQ by STRginner Permanovice (Battlesage, Dreamscapes, Hanger), and because my swashbuckler hydropotina was already in Mu Lung Gardens, I headed there as hydro:

    [​IMG]

    Not gonna lie, using Wings to hop directly from one large platform to the next during the Through the Deck stages is pretty fun.

    And here is hydro, with Permanovice, Yukinoumi, & LadyDemeter, fiercely battling the Lord Pirate:

    [​IMG]

    And later, I swapped over to my permarogue panolia so that panolia could try her hand at PPQ for the very first time~:

    [​IMG]

    And thanks to Permanovice and IceGrinder for kindly helping panolia do the PPQ quests~! <3
     
    • Like Like x 2
    • Great Work Great Work x 1
  19. OP
    OP
    deer
    Offline

    deer Pac Pinky

    195
    398
    191
    Oct 27, 2020
    Female
    Oddville
    2:08 PM
    cervid
    Bishop
    130
    Oddjobs
    (You can also read the original diary entry here, and/or read the diary entry on the Oddjobs forum.)

    rangifer’s diary: pt. lxii

    Taxonomising odd jobs, pt. iii: Exploring the space of possible taxonomies. §4

    In the previous section of this part (pt. iii, §3), I talked about some clustering methods, including neighbour-joining, as well as linkage-based hierarchical clustering methods. The latter methods need some way of determining linkage; we called these “linkage criteria”. There are more linkage criteria out there than I care to talk about, but there is just one more that I want to go over, in addition to the ones in §3:

    Energy distance

    The energy distance, as a linkage criterion, can be thought of (if you squint just a little bit) as a more sophisticated version of UPGMA (see §3). As its name implies, it is actually a full-blown metric (sOmETiMEs), which is nice. Defining the energy distance is not all that complicated, but because its origin is in defining (statistical) distance between probability distributions, I want to start with that kind of definition first.

    Between real-vector-valued probability distributions

    Let , ′ ℝⁿ be random vectors* each with distribution function , and let , ′ ∈ ℝⁿ be random vectors each with distribution function . Let {, ′, , ′} be independent. Then the square of the energy distance between and is defined as follows:
    It may be alarming that we have subtraction going on here, as this is the square of the energy distance, so to recover the energy distance itself, we have to take the positive square root. If the squared distance is less than zero, we will end up with an imaginary result, which we obviously cannot tolerate for a distance function. For the definition above, where we are talking about independent real random vectors, and comparing their respective distributions, it works out (for the same reason discussed below). However, it is common to generalise the energy distance to arbitrary metric spaces (which is basically what we want to do here). Before worrying about being a metric in the general case, here is the general case itself:

    Any metric space, not just real coordinate space

    Let (, ) be a metric space. Treat as a sample space†, and let , ℬ both be probability measures over it. Let , ′ ∈ be random variables each sampled from , and let , ′ ∈ be random variables each sampled from ℬ. Let {, ′, , ′} be independent. Then the square of the energy distance between and ℬ is defined as follows:
    It’s not really important to understand the details of this more general definition (or of the original definition, for that matter…), but I’ve taken Wikipedia’s definition here, and cleaned it up so that my head doesn’t hurt when I read it. As you can see, it looks basically like the first definition, except that the leading text is worded a little differently. This definition will lead us to the actual definition that we care about.

    In general, this is a metric iff is something called a “strict negative definite kernel” (Wikipedia calls it a “strongly negative definite kernel”). This term is defined in a 2005 paper, and it is also proven there that the Euclidean distance is such a metric, which is why our original definition (where the objects are random vectors over specifically ℝⁿ) works so well. Chances are, I’m not smart enough to prove (or disprove, for that matter) strict negative definiteness for some random-ass metrics. So, in the end, if energy distance proves to be useful in any way, we might have to just throw shit at it and hope that it doesn’t spit out any imaginary numbers. Realistically, we would just use the squared distance anyways (at which point, we can handle negative values), because this is a linkage criterion — we just need to know which pair of clusters minimises the “distance”, regardless of whether that “distance” is squared or not… and regardless of whether it’s a metric or not. Iunno.

    In any case, we can now get to the third and most useful formula of them all. By taking the generalisation above, and then replacing the expected value (E[·]) with its concrete (i.e. operating over individual concrete values/samples) analogue, the arithmetic mean, we get:

    Between finite sets of concrete values/samples

    Let (, ) be a metric space. Let , . Let || = , and || = . Let ∈ , ∈ . For the purpose of summation, select all , such that 1 ≤ , ≤ , and all , such that 1 ≤ , ≤ . Then the square of the energy distance between and is defined as follows:
    Wow, that is truly ugly. I guess plain text just isn’t the perfect medium for mathematical typesetting

    [​IMG]

    But why tho?

    Like UPGMA (and WPGMA, for that matter), this linkage criterion operates based on arithmetic means of individual object distances. Unlike either of those, however, energy distance takes into account intra-cluster distances, in addition to the inter-cluster distances that UPGMA/WPGMA already take into account. Energy distance apparently owes its name to the concept of potential energy in physics — particularly gravitational potential energy in the context of classical mechanics. In this analogy, individual observations are celestial bodies. These bodies are positioned relative to their hypothesised distribution, which acts as the “reference”/“origin”/“zero” point relative to which we measure the bodies’ potential energies. The ideal goal, then, is to minimise the potential energies of these bodies — at this point, the hypothesised distribution best fits the actual data (or, best fits some other abstract distribution, as the case may be).

    *A random vector is basically just a fancy name for a vector of zero or more random variables, which all have to be from the same probability space. For example: four flips of a fair coin (each flip having two possible outcomes, both equally likely: {Heads, Tails}) could be represented as a random vector of length four, where each element of the vector is a random variable representing a single coin flip.

    †We actually also need to define a σ-algebra over — call it — such that (, ) is a measurable space, so that we can define a probability measure over it. …Actually, it specifically should be a Borel σ-algebra, but who cares…

    Orderings

    I want to take a little break from the heavy-duty statistical/data-mining stuff, and look instead at another kind of mathematical tool in our toolbox. Whereas the statistical stuff is obviously oriented towards the “phenetics-inspired” approach (not that neighbour-joining, clustering, etc. aren’t used in cladistics — they are, but our objects have no genome or anything like that, so…), our “cladistics-inspired” approach requires a different kind of handiwork. Whereas the phenetics-inspired approach requires that we do all of the following (plus probably some more stuff):
    • Determine a suitable encoding(s) for the odd jobs.
    • Embed the above encoding into some kind of metric space (or, at least something that vaguely looks like it…?).
    • Determine a suitable algorithm(s) to taxonomise/cluster/whatever the data.
    • Implement the above algorithm(s).
    • Interpret the results somehow.
    On the other hand, our cladistics-inspired approach is much more hands-on with the classification of the data, because we will not be using statistical methods at all. Instead, we will want to put together our own “““phylogenetic tree””” that expresses the “““ancestral relationships””” between our objects (viz. our odd jobs).

    Although I’m pretty sure that I stressed this at least once before, I want to emphasise once again that the inspiration that I’m drawing from biological taxonomy is not supposed to draw a serious analogy; there is little or no actual similarity between the ontology that we have here, and the ontology of life on Earth. That being said, drawing this inspiration proved useful just for two reasons:
    • Biological taxonomy is highly developed as a discipline, so it lends many techniques and principles that we can make use of in our own way.
    • It gave me some kind of starting point for my thought process.
    With that out of the way, let’s take a look at some more tools that we can use to formalise some aspects of our handmade “phylogenetic tree(s)”.

    Weak ordering

    The first tool comes from order theory, which sounds scary, but never fear. It’s really simple: weak ordering. A weak ordering on a set is just a ranking of the set’s elements, with the stipulation that it’s possible for elements to be tied with one another. Two-way ties, three-way ties, four-way ties, etc. are all possible. For example, if we wanted to rank MapleLegends players by fame, we would expect there to be some ties. Say we have IGN famemesoicanwearmygear* (20 fame), IGN VeryNice (69 fame), IGN b14z317 (420 fame), IGN Honsterslut (0 fame), IGN Nice (69 fame), IGN fineStructure (137 fame), IGN BloodshotSclera (420 fame), and IGN sexnumber (69 fame). Then we can impose a weak ordering onto this set of 8 MapleLegends characters, by ordering them by their fame, which we can graphically represent like this:
    1. b14z317, BloodshotSclera
    2. fineStructure
    3. Nice, VeryNice, sexnumber
    4. famemesoicanwearmygear
    5. Honsterslut
    …Induced by a strict partial order
    One of the ways that weak orderings are usually defined is in terms of a partial order — particularly, a strict partial order. The difference between a “strict” ordering and a usual unqualified (not “strict”) ordering is the same as the difference between “<” and “≤”, respectively. “<” is strict because it reads (at least, in the context of real numbers and similar things) “strictly less than”, as opposed to “≤”, which reads as “less than or equal to”. Partial orders are called “partial” because they are not necessarily connected (sometimes “connected” is called “total”), which means that even if and are distinct ( ≠ ) elements of some poset† , it is not necessarily true that they are related. In order words, it is not necessarily true that < < . Elements that are unrelated in this way: ¬( < ∨ < ), are said to be “incomparable”.

    The point of mentioning strict partial orders is that, the incomparability itself is a kind of relation. We can think of our above example (with the 8 MapleLegends characters being weakly ordered by fame) in terms of a strict partial order as well: < iff has strictly less fame than , and furthermore, and are incomparable iff their fame values are equal. This strict partial order is special, because it has an additional property that not all strict partial orders have. Strict partial orders, by definition, already have:
    But now, we have another property that holds, because this strict partial order is not just any strict partial order — it is defined by taking a weak order and making it strict!:
    • Transitivity of incomparability: ( is incomparable to ) and ( is incomparable to ) → is incomparable to .
    This is easy to see, because the only way that can be incomparable to is if they are tied (equal fame values), and because fame values are just integers, equality between fame values is transitive. So the above definition of “transitivity of incomparability” can be rewritten, using our example, as: ( has the same fame as ) and ( has the same fame as ) → has the same fame as . Which is pretty obvious, right? This is kinda cool, because it means that our weak ordering is actually kinda two orderings at once — if you take one ordering and flip its “strict”ness, you get the other ordering! And they both satisfy nice (but mostly distinct) properties.

    Foobeeny

    You might reasonably ask exactly how many weak orderings are possible, given a set of || = elements. Well, you’re in luck, because these numbers are called the ordered Bell numbers (a.k.a. Fubini numbers): A000670 (OEIS). The OEIS lists all such numbers for ≤ 20, but looking at the Oddjobs website’s list of odd jobs, 45 odd jobs are listed, there alone. So I wonder, how many unique weak orderings are there for = 45? Well, thankfully we have a handy-dandy recursive formula for calculating the ordered Bell numbers. Let () be the number of unique weak orderings of a set of elements. For the purpose of summation, sum over all such that 1 ≤ ≤ .
    Wikipedia also offers a direct summation (both finite and infinite; we don’t care about the infinite ones) formula for ordered Bell numbers, which is defined in terms of Stirling numbers of the second kind. Thankfully, Stirling numbers of the second kind can be expressed using a fairly simple combination of factorials, binomial coefficients, and good ol’ fashioned integer exponentiation. So, to calculate the 45th ordered Bell number, I wrote up a little implementation in Python — actually, a few implementations. The mathematical version of the recursive formula is listed above, and for reference, here is the mathematical version of the non-recursive direct summation that I also used (for summation: 0 ≤ ≤ , 0 ≤ ≤ ):
    And for each of the two formulae above, two Python implementations. Thus, a total of four implementations:

    Code:
    def fubini(n):
        """
        ``fubini(n)`` is the ``n``th Fubini number (a.k.a. the ``n``th ordered Bell
        number).
        """
    
        a = 0
        for k in range(0, n + 1):
            for j in range(0, k + 1):
                a += (-1) ** (k - j) * choose(k, j) * j ** n
    
        return a
    
    
    def fubini_gen(n):
        """
        ``fubini`` defined by ``sum``ming generators, instead of using explicit
        loops.
        """
    
        return sum(
            sum(((-1) ** (k - j) * choose(k, j) * j ** n) for j in range(0, k + 1))
            for k in range(0, n + 1)
        )
    
    
    def fubini_rec_naive(n):
        """
        Do **not** use this function. It is only here for illustrative purposes,
        and is completely useless for any significantly large values of ``n``.
    
        Implements ``fubini`` using a naïve recursive method. I think the runtime
        is Ω(n!).
        """
    
        if n == 0:
            return 1
    
        return sum(
            (choose(n, i) * fubini_rec_naive(n - i)) for i in range(1, n + 1)
        )
    
    
    def fubini_rec(n):
        """
        Implements ``fubini`` using recursion, but memoises, in order to make the
        performance (read: asymptotic runtime behaviour) reasonable.
    
        <https://en.wikipedia.org/wiki/Dynamic_programming>
        """
    
        memo = {0: 1}
    
        def a(m):
            if m in memo:
                return memo[m]
    
            fubini_m = sum((choose(m, i) * a(m - i)) for i in range(1, m + 1))
            memo[m] = fubini_m
    
            return fubini_m
    
        return a(n)
    

    The first implementation (fubini) is the straightforward, imperative implementation of the second mathematical formula.

    The second implementation (fubini_gen) is very similar to the first, but uses generators (combined with the sum built-in function) instead of explicit loops.

    The third implementation (fubini_rec_naive) is a naïve implementation of the first mathematical formula (the recursive one). As the fubini_rec_naive implementation itself says: it is very bad. This naïve implementation has a runtime of, as far as I can tell, Ω(!). This is, to use a highly technical theoretical computer science term, “utterly useless garbage”. The reason that I wrote the naïve implementation, is for illustrative purposes: to show that implementing the recurrence relation as naïvely as possible does actually get you the right answers, but also the heat death of the universe may have already happened by the time that your computation completes.

    The fourth implementation (fubini_rec) is a less naïve implementation of the first mathematical formula (the recursive one). This implementation is indeed recursive, but it uses a little dynamic programming trick to avoid a poor asymptotic runtime: memoisation. I did a really shitty, informal performance comparison by running fubini.py with something at the end like: for _ in range(100): f(100). Substitute each one of the four implementations for f, and you get the idea. I ran the script like this five times per implementation, each time with the time command in front, like: time ./fubini.py. For each batch of five, I took the arithmetic mean by summing the times and dividing by five. For the fubini_rec_naive implementation, I took away the loop (no for _ in range(100):, so giving it a 100-fold speed advantage), but as expected, it did not complete. I let it run for 183 seconds, just for laughs I guess… I suspect that it would never actually finish (at least, not in a reasonable, human/deer timescale). I also did the same thing all over again, because I forgot to measure memory usage (/usr/bin/time -v ./fubini.py), and I wanted to see if that was interesting. I actually have a value here for fubini_rec_naive, as I just Ctrl+C’d after 60 seconds of execution time.
    implementationtime (s)peak memory (KiB)
    fubini1.1309144.8
    fubini_gen1.1239096.0
    fubini_rec_naive≈∞9180.0
    fubini_rec0.82710159.2
    As you can see, all three of the reasonable implementations are pretty comparable performance-wise. fubini_gen looks like it maybe performs slightly better (in space & time) than fubini, but probably not — I’m sure that’s just pure noise. fubini_rec is faster enough that I suspect it really is faster, which is kinda cool! Naïvely, I would expect the straightforward, imperative, direct summation to be the fastest — which would make fubini the fastest (or tied for the fastest). However, it seems that that is not so. However, as expected, fubini_rec does use noticeably more space. Here, the increase looks pretty modest, although I suspect that the actual overhead is larger than it looks here, as most of that memory usage is probably just the Python interpreter/runtime.

    Oh, right. We were asking what the 45th ordered Bell number is. Well, it’s this:

    1 255 482 482 235 481 041 484 313 695 469 155 949 742 941 807 533 901 307 975 355 741

    That is roughly 1.26 × 10⁶³, or 1.26 vigintillion. For comparison, this is much larger than the estimated number of kilograms of ordinary matter in the observable universe, which is a measly 1.5 × 10⁵³ kg — roughly 10 billion (10¹⁰) times smaller than the 45th ordered Bell number. Yikes!!

    I literally just forgot what I was doing

    I was going to cover some more stuff, but I got distracted. You might ask me: why did we even consider weak orders in the first place? Who cares?? Why am I reading this drivel???

    Those are all good questions, although I’m afraid that I can only answer the first one. The reason for considering weak orderings is that when we hand-craft our “phylogenetic tree(s)”, we may want to start with the most skeletal structure possible: ordering our objects (our odd jobs) in roughly “chronological order” (or rather, some notion of “primitiveness”). This will certainly end up being a weak ordering, as there should be many pairs of objects where we just don’t know if one “came first”, the other one “came first”, or they emerged at “the same time”. If is our set of odd jobs, and we had such a pair of objects (, ) ∈ × , then under our weak ordering (, ≤), it would be true that ≤ ∧ ≤ . We can then impose this ordering, later, onto our tree… or whatever it is. We shall investigate this next time, I guess.

    *I’m aware that some of these IGNs are too long to be valid. I think (IIRC) that the maximum length is 12 characters long?

    Partially ordered set.

    Grind sesh @ 3-4 F

    I was invited by xBowtjuhNL and Gruzz to have a little fun grinding at an unusual location (or, at the very least, unusual for such low-level characters): Sutra Depository 3-4 F. As some readers will know, the Sutra Depository (which is located within the Shaolin Temple part of the Mount Song / Shaolin Temple region of China) is home to several varieties of bald-headed monastic kung fu practitioners. True to their reputation, these humble monks are more than capable of beating the living shit out of you, if you get too close. 3-4 F of the depository, in particular, is home to Bronze Staffmen and, even more concerningly, Silver Spearmen. These Buddhist martial experts have gobs and gobs of HP — a single Silver Spearman, for example, has 4.5 times as much HP as a Jr. Balrog!! But they pay fairly well, if you can best them in combat — that single 225k HP Silver Spearman awards a base EXP of 11 935 (23 870 after a 2× multiplier)!

    So we gave it a whirl, with me playing my pure STR bishop cervid:

    [​IMG]

    We weren’t clearing these monks as quickly as I had hoped, so I decided to take charge:

    [​IMG]

    That’s right; with Genesis, cervid is capable of a whopping 0(!) aggregate DPM versus these guys. Incroyable.

    And, I found that Dooming the monks to turn them into wee blue snails was quite effective:

    [​IMG]

    Basic-attacking Blue Snails is just so much simpler…

    A lil bossin’ wif MPQ GANG

    I attended some of the usual Papu/Rav, also with Gruzz and xBowtjuhNL, as well as Harlez. Here, I was playing as my darksterity knight rusa. I managed to capture a rather chaotic moment (of which there are plenty, I suppose) during the fight with Papulatus’s 2nd body:

    [​IMG]

    You can’t actually see me in this image, per se… You can see most of the Spear Crusher animation, as well as my pet Rudolph, and even just a few pixels of my weapon (or rather, my NX pet hawk covering up the Sky Ski underneath).

    After the Papu fights, we had some fun with Chill Frog trying on various eye accessories. Here is Chill Frog, donning a very stylish snorkel:

    [​IMG]

    And we did Rav runs as well, of course. In which runs, our pets collectively reminded us of just how much they love the Pickpocket skill:

    [​IMG]

    So many shiny cylinders, so little time…

    d34r hits level 70~!!

    Now that the summer event maintenance was fast approaching, and my vicloc dagger spearwoman d34r was fast approaching the big level 70, I wanted to make the latter happen first! But first (even more first), of course, the obligatory “even more Jr. Rog hunting” — in futile hopes of a Golden River:

    [​IMG]

    As you can see above, I was joined by vicloc bandit xXCrookXx (Level1Crook, Lv1Crook) — now in full level 60 swag. Unfortunately, as usual, no luck with Golden River drops. Oh well.

    After even more FoG grinding — some of which was solo, some of which was also joined by xXCrookXx — I was joined by crusader Update, who I grinded with for not too long before hitting the fabled level 70:

    [​IMG]

    Yayyy!!! First vicloc level 70! o_o;

    I kept grinding with Update for a while afterwards, and also tried my hand at scrolling my only level 70 equip so far — a clean Silver Ancient Shield with very high stats — with my only shield STR scroll so far…

    [​IMG]

    u_u
     
    • Like Like x 2
    • Great Work Great Work x 1
  20. SiriusPlaque
    Offline

    SiriusPlaque Slime

    23
    0
    29
    May 15, 2020
    10:08 AM
    Honestly never thought I'd think about Borel sigma algebras in a mushroom game, kudos! I actually cared a lot....
    Definitely raised an eyebrow when you tried to define a probability measure on a metric space without specifying that it's measurable first, but I should have had faith that it was covered in a footnote.
    Althoughhhh I imagine the whole point of this exercise is to classify odd jobs based on a vector of features (but introducing randomness suggests each job is a random realization from a distribution of features????), and I bet theres a finite number of realizations. So your probability space (S, F, p) will have p be a discrete measure. Say S is still Rn, then I challenge you that you don't have to restrict F to a Borel sigma algebra, and the Lebesgue sigma algebra is OK.

    Tangent over, now back to your regular scheduled programming :yay:
     
    • Friendly Friendly x 1

Share This Page