10/1/2022 12 Comments 398 to 400: Twinz to ???ID 398: Twinz (Tsuinzu, Kageboozu, Wobbuffet)Oh boy. Tsuinzu—which is pretty easily romanized to “Twinz”—is quite the mess of a Pokemon. Originally a two headed ghost Pokemon, Twinz went through multiple reworks, changing into at least four distinct designs before ending up as Wobbuffet, a concept that was completely new and had only minimal connection to the original designs. Next to Sneasel, Twinz probably has the most complicated design history of all the Pokemon in Gold and Silver; and like Sneasel, despite being an early design that appeared in Spaceworld 1997, Twinz (Wobbuffet) was still being worked on just a couple weeks from the final release. So lets strap in, because there’s a lot to cover. Twinz, like I said, isn’t a horribly complicated concept: in its initial draft, it’s two ghost-looking creatures, connected by their tails, fighting with each other. It’s unclear why it's in Period 2c: either there was some reason the team wanted to include it in the initial draft roster for Pokemon 2, and so they hastily created a sprite slot for it before they hit #400, or this is actually the beginning of a bunch of throwaway sprites that I’m grouping together in 2d, and it happened to be good enough to be used by the time Spaceworld ’97 was created. There might also be some other reason Twinz goes together with the three baby Pokemon above it and the mystery frog, one that escapes us with the information we have. (Art by @RacieBeep) We don't have anything telling us what Twinz was like before Spaceworld '97, but by the time Twinz shows up in that build, it’s a Dark/Normal Pokemon that evolves into Girafarig at Level 29. It’s unclear if Twinz was originally even designed to be related to Girafarig or if it was grafted onto Girafarig, much like Gurotesu, Ikari, and Manbo1 were severed from their original evolutions and retconned into a family. There are arguments on either side. On the one hand, Twinz doesn’t look anything like Girafarig, and its typing (Dark/Normal) doesn’t seem to fit its obviously Ghost-type design. Furthermore, its placement in the Index doesn’t really make it seem like it was meant to be an evolution: Periods 2a and 3b are filled with evolutions of other ‘mons, but Period 2c doesn't seem like a part of the Index which was focused on creating evolutionary relatives that weren't babies. Potentially, maybe Twinz is found here in the Index because it’s a pre-evolution, much like Mikon et al? Arguing for the prosecution, it's easy to imagine that Twinz was created specifically to be connected to Girafarig. It’s a two headed creature conjoined at the body, just like Girafarig. One could even argue that Girafarig gets the spikes on its back from Twinz. Either they designed it that way on purpose or it was a big coincidence the team had two designs fitting the same gimmick, and they just had to pin them together. As I’ve mentioned previously, whether or not they were designed together, the Pokemon team clearly had difficulties making Twinz and Girafarig work as an evolutionary family. Given how different their body types are, the team must have struggled to find a moveset that made sense for both; by Spaceworld ’97, they're the only Pokemon family with only a placeholder moveset. They only have three level-up moves: Double Kick (which is perfectly in flavor for Girafarig, less so for Twinz), Growl, and Sand Attack. Not the most inspiring move pool for such an interesting design concept. In between Spaceworld ’97 and Spaceworld ’99, Game Freak tried a number of different solutions to make this family work. At some point, around April 1999, there was even work done on a Girafarig evolution! Sadly, we don’t have any sprites for it, nor any interesting information beside stats and evo data, evolving from Girafarig at level 34. It never ended up getting a name, and was ultimately overwritten by Pineco. It shared its typing with Girafarig, which in early 1999 was a Ghost-Normal type, presumably because Twinz made so much more sense as a Ghost type than as Dark/Normal. Of course, as I’ve mentioned, this introduced a new problem: Girafarig doesn’t look anything like a Ghost! It’s unclear what the team was doing at this point: they could have been creating a more ghostly and evil looking extension of Girafarig, in order to better sell the evolutionary line as a Ghost-type family; maybe that third form also merged some more elements from Twinz’s design into Girafarig, to make the family look more cohesive. It also could be the case that they were planning to make Girafarig itself the 3rd evolution of the family, and they were introducing a new second-stage Pokemon to better bridge the gap between Twinz and Girafarig. In fact, around mid 1999, a more conventional baby Girafarig pre-evolution was being conceptualized. By the time the Spaceworld ’99 demo came around, it had mostly been replaced with Kageboozu (early Wobbuffet), though its back sprite remained, taking the form of a baby Girafarig. This Pokémon has no back spikes, seemingly no snout, rounder ears, only one horn, and the "Chain-Chomp" which serves as its tail has no mouth. (Of note: the palette for "baby Girafarig" is reused from Twinz.) On the other hand, this baby Girafarig could have nothing to do with that second-stage they were conceiving; it’s more likely it was an attempt to redesign Twinz into something more like Girafarig, to make the transition between them make sense. After all, the backsprite is found in Twinz’ slot in Spaceworld ’99 (it also had a bizarre and unrelated front sprite!), which implies that Twinz was overwritten by this new design for a baby-Girafarig. As well, by Spaceworld ’99, Twinz had gained a version of Girafarig’s movelist and had become Normal/Psychic like Girafarig, so it’s likely it was also changed to look like Girafarig. The end result, either way, is the same: the Girafarig calf didn’t last more than a few months, and in either case, the team was trying to fix the family by reworking either the first, second, or third form. None of these ideas worked. Before Game Freak’s experimentation with baby Girafarig, Twinz had already gotten a redesign. By June of 1999, we have a completely new sprite for Twinz, which is far more than you can say for many of the other discarded Pokemon from Spaceworld ’97: about 24 of them disappeared altogether when development restarted, and very few got any sort of updated sprites. So the team clearly still had plans for Twinz by at least June. In that redesign, Twinz is no longer conjoined, and was now composed of two Haunter-like ghosts that rotated around each other. In fact, both of Twinz’s sprites—the Spaceworld ’97 ones and this new redesign—heavily resemble the contemporaneous sprites of Haunter. It’s definitely a coincidence, but it’s still interesting how they look more like a family than Twinz and Girafarig: Twinz was also now known as the “Spiral Pokemon,” which is kind of neat. Twinz never got a full moveset; by this point it had lost Double Kick and just had a placeholder moveset composed of Tackle and Growl. But it did have a Pokedex entry! Here it is: “Both heads are constantly swirling around. For unknown reasons, they always move clockwise.” Honestly, I kind of find this Pokedex entry a bit boring: figuring out why the ghosts swirl clockwise doesn’t feel like a mystery I, as a player, would be interested in delving into. The entry doesn’t really give a sense of Twinz’s personality, or how they live, just that they swirl. Twinz is reminiscent of an Ouroboros, or of a Yin-Yang, which seems fitting because of Girafarig's nature as a balance between good and evil. But swirling clockwise? Yawn. Obviously it’s in-progress, but between the boring nature of this entry and the lack of a moveset even this far into development, it does suggest that the team was still unsure of exactly what to do with Twinz. By Spaceworld 1999, as I mentioned, Twinz “exists.” But not really. While the name Twinz is being used in that slot, the backsprite is our friend the baby Girafarig, and the front sprite is a…a sketch of a potato? In retrospect, you can tell that the sketch vaguely resembles what Twinz would be replaced with—Wobbuffet—but on its own it’s pretty bizarre. (It’s a coincidence, but if you flip the sketch upside down, it almost perfectly matches the sillouette of Wobbuffet’s backsprite.) Twinz still has the typing and moveset of baby Girafarig at this point, and doesn’t yet resemble Wobbuffet either. Interestingly though, we have new Pokedex entries which hint at a very different background for this Pokemon, completely distinct from Girafarig, Twinz, or Wobbuffet. Here’s a Pokedex entry from Gold version on August 17th, a build a few days after Spaceworld ‘99 “They gradually disintegrate when exposed to light, so they live quietly in dark caves.” And here’s the Silver Pokedex entry, from the same time: “It lurks in dark places. It will not attack directly, but if attacked, it will return its foe's attack threefold.” Some of this does resemble what Wobbuffet would become, especially the information found in the Silver entry. The final version of Gold’s Pokedex entry mentions that Wobbuffet “dislikes light;” a far cry from literally melting when exposed to it, but the same general idea. The Silver entry also seems to imply this Pokemon counterattacks, which is basically Wobbuffet’s entire gimmick. Still, these entries suggest something that’s not quite Wobbuffet yet. The emphasis on how it lurks in dark places brings to mind some sort of monster, not the bright blue punching bag that Wobbuffet is modeled on. It’s also now the “Shadow” Pokemon in the data; not the “Swirl” Pokemon of Twinz but also not the “Endurance” Pokemon that Wobbuffet would become. The Pokemon we’re looking at in these entries is a shadowy ghost that fears the sunlight, protects itself from attacks, and is so incorporeal that it’ll melt when exposed to the clear light of day. Sounds like a Ghost or a spirit to me. Which is especially interesting, because there's evidence that between Baby Girafarig and Wobbuffet, there was even one more completely new Pokemon! Right after Spaceworld ’99, in the same build as these Pokedex entries, Twinz got renamed to “Kageboozu.” Kageboozu is an interesting, hard to translate name. “Kage” is easy; it’s the Japanese word for “Shadow” which makes sense for the “Shadow” Pokemon. “Boozu” could mean “Yokai” (Spirit) or bald head. But taking it literally is missing the point. “Kageboozu” is a play on the Japanese doll called a “Teru Teru Boozu.” They look like little ghosts, and are tied up in front of houses in order to wish for good weather. Kageboozu is, then, just a play on this name: Teru means “Shine,” and so “Kage,” shadow, is the opposite. Which sounds fitting for something that melts in the sun and tries to stay hidden in the dark (but, importantly, doesn't sound anything like Wobbuffet or Twinz). What’s more is that we actually have a sprite for this Kageboozu! Found on Wobbuffet's scratchpads is a strange sprite that depicts what appears to be a blobby ghost: exactly the sort of creature that is implied by those Pokedex entries. You can even see bits of the ghost melting off! Here’s its sprite, using Wobbuffet’s final palette, courtesy of @OrangeFrench: To sum up, it seems that Twinz did not change immediately into Wobbuffet: first, it became a baby Girafarig, and then it was briefly an entirely new Pokemon! Kageboozu probably did not last very long; the final game came out on a few months later and Wobbuffet must have been finalized nearly immediately. This was so immediate that Wobbuffet even still uses the menu sprite reserved for ghost Pokemon, even though it no longer resembled a ghost in the slightest. By the time the team made finalized Wobbuffet, its concept had strayed so far from the original idea for Twinz (or Girafarig, or Kageboozu) that it bared almost no resemblance to what came before it. Wobbuffet had three distinct inspirations now: It’s a punching bag, an Okiagari-koboshi (a Japanese paper-mache doll), and also an homage to the famous Japanese Comedian, Hayashiya Sanpei. The Okiagari Koboshi is a cute doll that traditionally has slits for eyes, much like Wobbuffet, and it is made so that it has a rounded bottom that will always wobble and reset itself facing forwards when pushed. Likewise, the punching bags Wobbuffet is based on were weighted on the bottom so that when a kid hit the bag, it would fall backwards, then swing forward and hit the kid back in the face: the harder it was pushed, the harder it would fall back upon its attacker. A perfect concept for a Pokemon that takes punishment and then deals the equivalent back to its opponent. Its third inspiration was a man named Hayashiya Sanpei. Sanpei was a well-known comic from the fifties, sixties, and seventies who participated in a style of comedy known as “Rakuga,” which was basically a genre in which the comic told long, rambling stories with a silly punchline at the end. He was known for touching his forehead in embarrassment when confronted by women, and his famous catchphrase, “So nansu, Okasan!” which translates to “That’s the way it is ma’am,” a catchphrase which inspired Wobbuffet’s Japanese name, “Sonan”. It’s also possible Wobbuffet took inspiration from another form of Japanese comedy known as Manzei. In Manzei, two performers, the "Tsukkomi" and the “Boke,” perform a routine in which the idiotic “Boke” constantly forgets things or humorously misinterprets what the other person is saying; he often gets hit over the head, with a slap or with a fan, when the Tsukkomi angrily corrects him. This sort of duo is common as a trope in a lot of Japanese media; I remember it most vividly in the characters “Solt” and “Peppor” from the game Chrono Cross. The Boke in Manzei is often round and overweight, and is characterized by his ability to take a lot of physical punishment as the Tsukkomi constantly hits him. Wobbuffet’s chubby looks make him look similar to the stereotypical Boke, and its experience of constantly taking damage from the opponent in a Pokemon battle seems reminiscent of a Boke constantly getting hit. It’s less of an obvious reference, but given that Japanese comedy was on the minds of Game Freak when they created Wobbuffet, I’m sure they at least considered the similarities between Wobbuffet and a Manzei routine. Gameplay-wise, Wobbuffet also evolved unrecognizably beyond it’s origins. Twinz, if Game Freak had found a way for it to work, would have had a typical moveset for a Ghost or Psychic Pokemon, probably mirroring the moveset that its evolution, Girafarig, had. But Wobbuffet was transformed into a gimmick Pokemon, based around the idea of a creature that only counterattacks back at its opponent. It only has four moves, all of which are designed to make either protect Wobbuffet or return any damage it takes to its opponent (assuming the player can guess correctly that incoming damage is either physical or special). My suspicion is that its incredibly tiny moveset (just Counter, Mirror Coat, Safeguard, and Destiny) and its counterattack gimmick was dictated in part by the incredibly tight time constraints the team had to finalize Wobbuffet. Kageboozu probably also used Mirror Coat, given its Pokedex entry, but it’s unclear if that was its only method of attack like Wobbuffet. Likely, Kageboozu never got a fully fleshed out moveset, and its use of Mirror Coat then inspired the team to further revise it to be based entirely around that move (and Counter). Rather than come up with an entirely new moveset that needed to be tested, they probably gave Wobbuffet a couple moves that went with its gimmick and called it a day. Time constraints may also have been the reason Wobbuffet is found in only one dark cave in Gold and Silver. It may have been the case that Wobbuffet’s encounter tables were entered into the game back when it was Kageboozu and thus more associated with darkness and caves. When it was turned into Wobbuffet, the team may have just left it to be found where Kageboozu had been placed, without really rethinking anything for the new concept. They left one reference to living in caves in Wobbuffet’s final Pokedex entries, for good measure. It’s very possible that the traits associated with Wobbuffet in the final were determined when it wasn’t even Wobbuffet yet! Despite Wobbuffet’s complete divorce from its origins, there’s one piece of Twinz left in its design. Ever wonder why Wobbuffet had that bizarre face on its tail? It’s the second head from Twinz! Someone, the conjoined twin ghosts survived, but one of them grew into a blue punching bag thing and now it drags around the other on the floor. I’ve always wondered why Wobbuffet had that bizarre tail; people have speculated before that its Wobbuffet’s real face, and that the rest of Wobbuffet is just a blow-up decoy. But now we know that the face is there as an Easter Egg, one tiny subtle hint to Wobbuffet’s humble origins. Twinz survived, in only a tiny way. Meanwhile, Kageboozu also has an interesting legacy. In Generation III, not only would a proper Teru-Teru Boozu Pokémon be introduced with Castform, but the name Kageboozu was also reused! Here, it is assigned to a new Ghost Pokemon, Shuppet. Unlike Gen II’s version of Kageboozu, Shuppet is supposedly more of a sock-puppet, though it still retains the flowing "skirt", the round defined head, and a horn evoking the string from which a Teru-Teru Boozu hangs: Could Shuppet be based on one of these earlier iterations of Wobbuffet? Maybe; none of the sprites we have look all that similar to Shuppet, but that doesn’t mean there weren’t more versions we’re unaware of. Shuppet’s Pokedex entries seem pretty different as well; though its Emerald entry mentions that it only comes out at night, that’s a pretty generic trait that many ghosts probably share, and its other entries just speak about how it feeds on negative emotions, which has no analogue in Twinz or the blobby ghost version of Wobbuffet. Most likely to me is that Shuppet was inspired by the name Kageboozu after they came up with it here; as it’s a pretty clever idea, the team probably had the leftover name and created a Pokemon that more literally embodied that name. There is one thing that’s a bit suspicious, however. Shuppet doesn’t just evolve into any random thing; it evolves into Banette, an evil ghost based upon a doll. Banette has more than a passing resemblance to Norowara, also designed for Generation II, and while it’s very unclear and unprovable, there’s at least a chance Banette was a revision of Norowara. If that were the case, was the Banette line cobbled together out of old Generation II concepts? It certainly seems like a possibility. It's amazing how memorable and iconic Wobbuffet has become, given that it was probably the very last Pokemon finished before the release of Generation II, give or take the revision of Sneasel. Wobbuffet was definitely iconic to me when I was growing up, and I associated Generation II with it as much as I thought of Marill of Togepi. On the other hand, I mourn Twinz, which was a great concept for a Pokemon that unfortunately never saw the light of day. Girafarig is a Pokemon defined by unused potential; if only Twinz existed, maybe both would have been more interesting and memorable as a result. ID 399: LedybaIn contrast to the complicated tangle that is Twinz/Kageboozu/Wobbuffet, Ledyba is notable for how stable its design remained throughout development. While its sprites were touched up in the transition from Super Gameboy to Gameboy Color, and its palette was changed, virtually everything else stayed the same. Its stats were identical throughout development; its typing and its name remained the same, and its name didn’t even get changed in localization. There’s some moderate differences in its movepool, which we’ll cover, but overall, this little guy was designed before Spaceworld ’97 and just left that way. Clearly the dude was a winner: the team seemed happy enough to leave him alone and focus their work in other places. In part, this relative stability could be because Ledyba was designed to be an early game Pokemon. As a result, it was probably a priority to get right early. Furthermore, Ledyba was one of the new Pokemon that could be caught in the demo shown at Spaceworld '97, which probably meant that the team was pretty confident in its design as it stood. Looking at the spritework, I also think there’s a chance that Ledyba was a Sugimori design. As we’ve seen in the past, the Pokemon he played a role in creating tended to stay almost exactly the same throughout development, possibly because he put more work into them before spriting them, or possibly just because as the team lead, he had a better idea of what he was looking for. Another reason Ledyba probably didn’t change that much is because its design inspiration was so simple. At risk of stating the obvious, Ledyba is a ladybug. That’s about it. While its evolution, Ledian, has some Super Sentai inspiration to it, Ledyba’s design seems to be completely absent of those same references. It has a distinctive mouth and cartoonish hands, but otherwise Ledyba is a design that does what it’s supposed to do: be a Pokemon version of a common insect. At a stretch, there's one other aspect of Ledyba's design worth mentioning. Ledyba has a some design similarities to the scrapped original evolution of Weedle, Kasanagi, a weird beetle Pokemon with gloved hands ready to punch somebody. We don’t know who designed Kasanagi, but it was a very early design in Generation I, and so it's possible that Ledyba took some of those design elements and put them on a different shell. (Speculative art by @RacieBeep) I am intrigued, however, by why Ledyba shows up where it does in the Korean Index. Period 4c is already only a weakly coherent string of seven Pokemon designs, but Ledyba seems to have even less to do with the other six. It’s designed as a single stage in a section filled with Pokemon which were made to evolve into known Pokemon; it’s a bug Pokemon design surrounded by much more exotic designs. It could be the case that Ledyba was part of the larger string off throwaway designs that were part of the Animal + designs of Period 2d, and that it happened to get lucky and be developed further than the rest were. It could also be the case that Ledyba was made later and overwrote something else disposable in this slot, making it seem like it's part of 4c and not 4d. My bigger suspicion, however, is that Ledyba’s origins have to do with the two bug families in Period 2b, in slots 382, 383, 384, and 385. My hypothesis about Period 2b was that it was a space where the development team knew they needed a specific type of Pokemon to fill a niche needed in the game; 2b was a run of two competing designs, each fitting the niche they needed to fill, made by different designers based on different inspirations. The team would, in theory, look at both designs, choose the one that fit the niche best, and discard the other one. If I’m right, my bet is that even though the team chose the Spinarak line to make it into Spaceworld ’97, they realized they wanted more early game bug Pokemon and they were unhappy with the moth line that was competing with Spinarak. Ledyba, then, was a quick addition to the games before they had the initial Pokedex filled in, giving them another early game Pokemon to play with and experiment with. Thus, if this was the case, they probably already planned for Ledyba to have an evolution but didn’t get to it immediately. After all, almost all bug Pokemon have evolutions, and an early game bug that didn’t evolve would be awful. Given that Ledian is at nearly the end of the Korean Index, it seems they didn’t get around to it for a long time, and even when they did, Ledian’s sprite and concept was really rough and probably still needed working out. Ledyba, then, was probably conceived with Spinarak in mind, which is even more fitting considering how they were eventually made version-exclusive counterparts in the final games. Saying that, Ledyba was never given a signature move like Spinarak got in Spider Web, and given that Ledian was created so late in the index, it seems as though more attention was put into Spinarak and Ariados. Again, maybe this was because Ledyba started out as a pretty okay design, so it became less of a priority for the team to fiddle with or perfect. The one thing that was moderately changed by the final game was Ledyba's moveset. Take a look at how different Ledyba’s Spaceworld ’97 moveset was from the final: The two sets only share three moves with each other: Reflect, Agility, and Swift. Swift was probably in both versions because of a Japanese pun: the word “Hoshi” can mean "star" or "lady bug spot," a pun that’s more pronounced in the early Ledian design which has stars on its back instead of spots. With that pun in mind, Swift is probably a joke about Ledyba throwing its spots at the opponent, which appear as stars in the animation. Otherwise, the original Ledyba is a bit more aggressive than the final: it learns Swift at a much earlier level and also learns Fury Swipes (a move that was originally a signature move for Meowth, oddly enough making an appearance here) and Tri-Attack, for some reason that isn’t clear to me. It also looks as though the original moveset implied that Ledyba had claws: Fury Swipes and Scratch both imply something sharp. The final version, on the other hand, gets rid of any claw references, and instead Ledyba uses Tackle and Comet Punch; Comet Punch, in particular, better matches the gloved hands that Ledyba sports. The final moveset is also more defensive, allowing Ledyba to use Light Screen, Safeguard, and Supersonic to protect itself. It also gained Baton Pass (a move that was the signature of Mr. Mime in SW ’97) but lost Double Team; losing Double Team was probably no accident given how powerful the two could be together. Overall, Ledyba’s fine. It started out cute, the team lightly refined its sprite, and then they made it even cuter in Crystal. Pokemon needs starter bug Pokemon, and Ledyba fits that niche well; why mess with success? ID 400: ??? And we’ve hit it: lucky number 400. Totaling all the designs up until now, we officially have 101 new designs to be introduced in Gold and Silver: the team could rest easy somewhere around here, knowing that they have a full roster (plus or minus the designs already discarded). Many of the previous ones were probably already ones the team had decided not to use, and the ratios of different types, early and late game Pokemon, and even diverse animal types are very off. But starting from about this point in the Index, the team had something they could work with, chip away at, and slowly reorganize as they edited it into a coherent new lineup of Pokemon. And the Pokemon they chose to be their final addition? A…weird looking warthog guy. With a face so funny looking it almost seems like a joke, ID 400 is strange guy that we know almost nothing about. There’s two theories we can entertain with this guy, neither of which is fully persuasive. So let’s look at both, and see which one you find more plausible. Option one is to just take 400 as it comes. This guy is probably based on a warthog or a feral pig (the difference is that one is found in Africa and the other in North America). At first, he might seem to be a hippopotamus or like a strange deformed dog, but his hair is the real clue. Feral hogs have a distinct plume of fur riding down their back, giving them their nickname of “razorback.” This guy’s punk mohawk is clearly inspired by that razorblade of hair common to warthogs, making it very probable that that’s what he was meant to be. With that in mind, we can draw a few more conclusions. Its not perfect, since of course warthogs usually have tusks, so this guy is probably a merger of a few similar animals with comparable traits. But at the very least, the hair on his back really makes me see this connection. He's also quite distinct from the other warthog in the Korean Index, ID 319, so maybe they didn't give him tusks to differentiate the two of them. Warthogs like this are not common in Japan, which suggests that ID 400 is another example of an Animal+ design based on an animal that Japanese people would find exotic. We’ve seen a lot of these, primarily in Period 1; Period 2d is about to show us a lot more of this type of design. Given ID 400’s placement, it is very possible that it fits better in Period 2d, alongside these other discarded Animal+ designs. Saying that, the second theory I’ll explain, if true, would put 400 decisively in Period 2c, so that’s where we’re stick him for now. On top of the warthog theming, it looks to me (and also @OrangeFrench, the main proponent of this theory) that the team played up the hair on its back to make 400 into something of a punk rocker. The hair looks a lot like the mohawk associated with an eighties punk band, and the rest of this guy gives a distinctive “bad ass” vibe to the design. It’s very silly, for sure, but the mouth and the eyes strike me as a rude or grotesque design, better fitting the idea of a creature that does what it wants, when it wants. My only doubts about all this is that ID 400’s sprite is 44x44 pixels, the smallest sprite size. While lots of Pokemon in the Korean Index use this size, it's often associated with first form Pokemon who evolve. And this guy, at least to me, looks like he was designed to evolve into something. It’s small and squat; it has tiny little hooves for feet that could easily grow into something more; it has strange black pants that don’t seem to go with its design, but might have a connection to another one of its family members. This is mostly a matter of opinion, but to me it looks as though this guy was designed with other Pokemon in mind. Which brings me to the second theory: ID 400 was an attempt at designing one more baby Pokemon to go along with Pichu, Cleffa, Elebebii, Mikon, Gyopin, and Puchikoon. I think it’s plausible that ID 400 is in fact a rough draft of a 1st stage evolution for Drowzee and Hypno. Given that Spaceworld ’97 is simply littered with baby Pokemon designs—even for unpopular Pokemon like Grimer—it would be strange if the Korean Index didn’t have any unused designs for babies. And while it’s impossible to be definitive about it, ID 400 is the best candidate for a discarded baby design. First of all, 400’s place in the Index is suspicious. It’s right next to three Pokemon which act as first stages for Generation I Pokemon. If 400 really is a pre-Drowzee, then it fits perfectly alongside Mikon, Puchikoon, and Gyopin in the Index. Secondly, look at the similarities 400 has with Drowzee and Hypno. 400’s weird black pants make perfect sense when he evolves into Drowzee, which already has a black lower body (mimicking a tapir). As well, the razorback mohawk that 400 sports admittedly looks out of place next to Drowzee, but seems perfectly in line with the mane of fur surrounding Hypno’s neck. On a more general scale, Drowzee and Hypno are pretty grotesque looking Pokemon, defined by their menacing presence, the fact that neither is “cute” in any sense, and by their distorted features, such as Hypno’s large nose. 400 also shares all of these features, in that it is purposefully ugly looking, not cute in any traditional sense, and has an out of proportion mouth. Looking at all of these factors together, it’s plausible to believe in some connection here.
On the other hand, there are reasons to argue against this hypothesis. First, notice how the Spaceworld ’97 sprites for Drowzee and Hypno are not created by the same artist who did ID 400: they both feature dithering in their shading, while 400 lacks anything like that. If this was created as a pre-evolution, it was made separately from those sprites. Secondly, there’s the obvious argument that 400's lack of tapir-like features is evidence that it isn't related. If you were going to design a baby Drowzee, the argument might go, why not give it at least some tapir-like characteristics? One particular trait that’s missing is a nose like Drowzee’s. Obviously, 400 need not have a trunk, but the hippo-like snout it does have doesn’t seem like it would grow into the Drowzee's. Also look at the eye designs: Hypno and Drowzee’s are both thin ovals, while 400’s are more bug-eyed. Finally, the general mood of the Pokemon is different: Drowzee and Hypno look malicious and scheming, while 400’s just excited to be alive. The obvious retort to this line of argument is that 400’s a rough draft: the team never went forward with a Drowzee baby Pokemon, and so this first draft was never redrawn to align more fully with Drowzee. Maybe so. That could explain the lack of dithering on the sprite, as well as the lack of tapir-ness. But as easy as this is to say, it doesn’t prove anything; it’s just a hypothesis to explain away the inconsistencies stated above. At the end of the day, I don’t think we can prove a relationship with the Drowzee family here, though I certainly see the resemblance. Whatever he was supposed to be, this cool dude is one of the unused designs I wish we could know more about; I’d love to understand how this idea came about and what plans the creators had for him. This guy, related to Drowzee or not, is a cool dude, and I'd love for some punk rock to find its way into Pokemon.
12 Comments
9/14/2022 11 Comments Period 2c: Final Touches to the First Draft (ID 394: Puchikoon to ID 397: ???)Period 2c is a shorter sequence of Pokemon designs. Compared to 2a and 2b, which had twenty-one and sixteen entries respectively, Period 2c is only composed of seven Pokemon designs, ending on ID #400. This sequence features four baby Pokemon, three of which were repurposed from unused Generation I designs, a pre-evolution to Girafarig, an additional bug Pokemon design (probably made to complement the options from 2b), and an odd frog design, which may fit alongside the three repurposed Generation I designs (I’ll make that case when we get there). While the three reused Generation I designs all demonstrate an obvious pattern, the others feel very disconnected, making it harder to discern an obvious theme for this section. The key puzzle piece, however, is that Period 2c ends at ID #400, which is not coincidentally 101 designs since the beginning of the Index. The team had said in interviews prior to the release of Gold and Silver that they were shooting for 100 new Pokemon designs (plus, I suspect, one more hidden design, like Mew). If they were serious about that number, ID #400 would be the very last design before the team had to start cutting. Thus, at a very simplistic level, by the time Period 2c is over, the very first draft of Gen II’s Pokedex would be complete. With this puzzle piece in place, it becomes clear that the point of Period 2c was to find seven more designs that would put the final touches on the team's first draft of the new roster. The three reused Generation I designs were likely pulled in because the team wanted to have a few more baby Pokemon to test out, especially given that by this point the team only had Pichu, Cleffa, and Elebebii to show off that new mechanic. Twinz was added because the team wanted to give Girafarig an evolution (that might explain why it doesn't look very much like Girafarig; it was made quickly to fill that role), Ledyba because the team was still looking for the right bug Pokemon to add to the early game, as we could see in Period 2b. The last two--the unused frog and hippo 'mons-- are harder to divine, but I suspect one may have been a riff on an unused Gen I design, and the other was another quick attempt at a baby Pokemon. In all of these cases, the designs filled one more niche the team wanted the game's roster to have, and/or was pulled from older unused designs as a matter of convenience before they had a roster they could test. It probably wasn't quite this neat: I’m sure by this point, the team had already decided against a few of the earlier designs we've seen in the Korean Index, and a few were certainly abandoned by this point (305 through 308 seem like the first to go, in my opinion). So while the roster may have been technically complete, loads of these Pokemon were probably unused slots at this point. Plus, by this point, most of these Pokemon were probably nothing more than dummy movesets and dummy stats, so it’s exaggerating to say that the team had a complete library of new Pokemon; more accurately, they had 101 sprites they could start riffing off of. But at the very least, the team had one Pokemon sprite filling up each of the extra Pokedex slots, and a rough draft of what the new roster would at least look like. Every Pokemon design after Period 2c in the Korean Index was either never adopted into the games, or had to replace something else. Alternatively, at least the first three (Puchikoon, Gyopin, and Mikon) might have been part of the same design process as Period 2b: they might all have been pulled into the Korean Index so the design team could choose one of them to fill a particular slot. If that were the case, obviously plans changed, as all three were found in Spaceworld ’97, before all three were then deleted after development on Gold and Silver was restarted. More likely, these last seven were just the finishing touches to fill out the roster, before the team entered a new phase based upon balancing, adding evolutionary relatives, and creating Pokemon needed for their new world. Period 2c wasn't the end of Pokemon design, but instead more like a beginning: from this flawed original roster of 101 'mons, the team could decide what the games would really look, and start chiseling away at some of the weaker ideas. ID #394: Puchikoon Puchikoon, our first of three returning designs from Generation I, is a cute little pony. Its name, according to TCRF, probably derives from “Petite” and “Unicorn.” The word "Puchi" means small or petite, so this seems pretty plausible. Puchikoon is a reused design from Generation I, now repurposed to show off Game Freak's new idea of introducing cuter, lovable baby versions of well-know Generation I Pokemon. Originally, Puchikoon wasn't a baby Pokemon at all, but instead just the first evolution of Ponyta. We don't have a front sprite of its original design, but here's what the family looked like all together, in both the Generation I incarnation and the version found in Spaceworld '97: It's uncanny how well that Generation I backsprite fits in right next to the backsprites for Ponyta and Rapidash, making it clear it was always supposed to be there. The only major difference is that Puchikoon seems to have darker fur than its evolutions; however, given that this difference in fur color is reproduced in Spaceworld '97, it seems likely that was a feature of the lost Generation I backsprite as well. As for the Spaceworld sprites, I don't have a clue what's going on with that weird pose for Rapidash, but I really like how Ponyta looks like it's starting to run from the stationary pose of its preevolution. It's a nice bit on continuity between the sprites. One of the interesting aspects of Puchikoon is that it (and its evolved family) originally formed a Electric/Water/Fire triad with two other three stage Pokemon families. The Gyopin/Goldeen/Seaking family was the Water-type counterpart of this tried, while the Kotora/Raitora line was it’s Electric-type counterpart (originally in Generation I, Kotora and Raitora also had a third stage evolution). All three are found in the internal index right next to each other, and all three have three stages, based around the idea of elemental real-world animals. Here they are, all together. Originally, Generation I was full of these Water (or Ice)/Fire/Electric triads; as I’ve said before, we have good evidence that these three types originally formed an elemental triangle, much like they do in the Final Fantasy series. In fact, we have some evidence that originally, the Special Stat was divided into three separate stats: A Pokemon’s affinity with Fire, with Ice or Water (they were probably not separate elements yet), and with Electric. When the team came up with more types, this original system was abandoned, but until very close to the end of development, the roster heavily leaned upon this trio of types. Eventually, of course, Grass replaced Electric as the third type in the Grass/Water/Fire Rock/Paper/Scissors of the games, and the designers cut a number of Electric Pokemon and replaced them with the Oddish and Bellsprout lines to reflect the new importance of Grass types. You can’t really tell that Goldeen and Ponyta were part of one of these trios in the final game, of course, because of how many Pokemon were deleted from this section. Kotora’s entire family was removed; so was Puchikoon; as well as Goldeen’s pre-evolution, Gyopin. This purge left just four of the original nine Pokemon, hardly enough to see a pattern. Part of the reason these were deleted was probably because the games had too many Electric and Water types. In general, though, the team had too many Pokemon designs to fit in the final game and they needed to be picky. At some point, late in the development of Pokemon Red and Green, the team needed to cut about 40 of their 190 (or maybe 187) Pokemon designs, and the team started taking a hacksaw to the roster. There’s no real mystery about why Puchikoon was one of those ~40 Pokemon dropped. In general, as the team was finding ways to trim the fat of the original roster, they heavily cut down on three-stage evolutionary families. The three we’ll talk about in this article—Gyopin, Mikon, and Puchikoon) are all examples of three stage families being shortened to just two stages; in addition to these three, the Pikachu line lost its third-stage evolution (Gorochu), the entire three stage Kotora line was deleted, the first stages of the Zubat families and the Meowth families were deleted, and the middle stage of the Psyduck family was lost to time. In the draft stages of Pokemon Red/Green, three stage lines were much more prominent, while in the final, usually three stage lines denoted either a starter Pokemon or a particular fearsome and powerful final stage (Nidoking, Nidoqueen, Dragonite, Machamp, etc). In addition, if we view the lines that were shortened in this way, a continuing pattern we find is that the deleted Pokemon were usually visually very similar to their second forms. In fact, usually the deleted form was only different in that it had less elemental or fantastical features, and was more akin to a real-life animal. Puchikoon, for instance, looks very similar to Ponyta, even down to them both having a mane and a tail made of fire. The main difference between the two of them is that Puchikoon is a bit smaller (at least compared to Ponyta's Red/Green sprite, as the Spaceworld '97 one looks about the same size), Puchikoon is orange and not primarily white, and Ponyta’s flames are slightly more prominent. Otherwise they’re identical. You see this as well with the other discarded first forms. First-stage Zubat, from what we can tell, had a smaller main body than Zubat, but otherwise shared the same bat wings: Zubat and Golbat’s more fantastical elements, like their gigantic mouths, likely did not come through from the first form. Gyopin is a regular looking fish that only gets a distinct sillouette when it evolves into Goldeen; Mikon looks very similar to Vulpix, except it’s tail looks more normal and less distinct than Vulpix’s. This reminds me of a pattern I've noticed before when examining the early moveset data we have for Red and Green. The drafts are difficult to parse for a number of reasons: they’re very incomplete and only cover a fraction of the Pokedex for one. They also often don’t always follow a logical progression: Deer, for instance, learns “Jump Kick”—a powerful fighting move—at level 16 but waits until level 37 to learn the relatively basic “Leer.” While Pokemon generally learn more powerful moves as they level up, a lot of these movesets feel strange and relatively scrambled. There is one consistent difference between these movesets and the final ones though: each Pokemon only learns a portion of its moveset, while the rest is learned only by its evolved relatives. What I mean by that is that--for instance--in the final, Bulbasaur, Ivysaur, and Venusaur all learn the same moves, just at slightly different levels. In the early movesets, however, Bulbasaur learns no moves after the level it evolves, and Ivysaur only learns moves at levels after it evolved from Bulbasaur and before it evolved into Venusaur. Usually the third form just learned on or two “super moves” to finish off the moveset. This is presumably because the idea of halting an evolution wasn’t in the games until very late, and so there was no need for Bulbasaur to learn moves at levels higher than it evolved. However, because this was the case, it meant that the designers could use differences in movesets between evolutions to actually define the identities of those individual evolutionary relatives. Take Growlithe, for instance. Here is its early moveset, next to Arcanine’s: Notice anything? Growlithe doesn’t learn any fire moves; only Arcanine does. In fact, to show the player that Arcanine is now capable of fire attacks, Arcanine's very first level up move is Ember, immediately differentiating it from its younger form. The implication is clear: evolution allowed a relatively mundane dog--Growlithe--to learn magical abilities, like the ability to breath fire. In this early stage, at least, Pokemon might change their entire properties upon level up. This is an idea that mostly was scrapped in the final, though it lived on in a couple places (the Eevee family, Abra and Magikarp, and how Exeggcutor can learn stop, primarily). This change was probably due to the innovation of B-button evolution canceling, which gave the player greater choice in exactly when their Pokemon evolved. Once this feature was put in place, evolutions needed to have basically the same movesets, so that no matter when a Pokemon evolved, it gained the same moves. A good mechanic for gameplay, but unfortunately one that made it harder to give each Pokemon a unique identity. In this conception, evolution seems to have a subtly different flavor: evolving allowed a Pokemon to unlock some additional fantastic abilities that it would not otherwise have, making evolutionary relatives more versatile and powerful not just due to their stats, but because of their new abilities. Growlithe may not have even been fire-type in this initial stage. We see this to a more subtle degree with some of the other Pokemon in this initial draft as well: Zubat’s initial pre-evolution learns mostly physical moves, while Zubat and Golbat learn Confuse Ray, Psychic, and Confusion, suggesting that they get better at using mind games as they evolve. Ekans learns no Poison-type moves until it evolves into Arbok, which gets Acid and Toxic. The Squirtle line learns Blizzard as a super move once it evolves into its final form, suggesting that its third form conferred upon it a super move that gave it access to Ice-type abilities. And originally Abra learned only Doubleslap as a damaging move before it became Kadabra; only then did it perfect its psychic powers and learn Psybeam and Psychic. Interestingly, given that the final games took away almost all of Abra’s moveset, the relationship between Abra and Kadabra might be a remnant of this earlier philosophy of evolution. The conclusion I draw from this is that, in many cases, the earlier evolutions of Pokemon were less likely to have any fantastical powers, but be mainly physical attackers who could only do the same things their real world counterparts were capable of. Which puts the more basic, realistic designs for Mikon, Gyopin, and Puchikoon into context: they were meant to be closer to real world animals before they evolved into elemental beasts. By cutting these Pokemon, the designers were essentially cutting to the chase. Every Pokemon would start out fantastical—except for a few exceptions, like Magikarp or the bugs—and just grow into something even more supernatural. Certainly, by cutting out these early evolutions, the Pokemon that are left over are more distinctive as a result. So if Puchikoon was removed, then why add it back in to the Korean Index, and even include it in Spaceworld ’97? Well, context is everything. By this point in Period 2, the Game Freak team had decided that the concept of baby Pokemon would be one of the marquee new features of Gold and Silver. As a result, they were happy to show it off as much as possible. By Period 2c, the team was probably trying to fill out the last bits of the initial draft of their roster, and including more baby Pokemon made sense; up until this point, the Korean Index only has Cleffa, Pichu, and Elebebii. So the team reached back into their old designs to quickly insert some more examples of babies into the roster. And discarded first evolutions, dropped because they were too simple or simply smaller versions of the second form (as Puchikoon was to Ponyta), made perfect sense. By adding in three of their best discarded first form designs, the team was able to bring the total number of baby Pokemon to six, showing off the potentials of this new idea. By the time Spaceworld ’97 was ready, the team even brought back the discarded first-stage of Meowth as well, leaving the baby Zubat as the only completely unused first form not to make an appearance in Spaceworld ’97 at all. It’s important to remember that Puchikoon and Gyopin were not originally designed as babies. Notice how Kotora wasn’t added to Spaceworld ’97 as a baby Pokemon at all, despite Kotora being a direct counterpart to Puchikoon and Gyopin. In part, that demonstrates how these designs were repurposed for a new idea when they were translated to Generation II; I wonder if the original Gen I designs were made smaller or cuter when they were redrawn for Gold and Silver. Furthermore, this provides more evidence that the Korean Index is a chronological index of designs. Kotora, which appears much earlier in the Index, was not made into a baby Pokemon because it was added before the team came up with the concept. Mikon, Puchikoon, and Gyopin were made into baby Pokemon because they were added at a point in development when the team were actively looking for more babies to show off the new mechanic. I have more to say about these baby Pokemon, but much of this concerns Gyopin as much as it concerns Puchikoon. So let’s move on to Gyopin and continue to explore the development of these cuteys. ID #395: GyopinGyopin’s an itty-bitty fish dude, the original first form of Goldeen. Gyopin’s name is probably a portmanteau of two different words, though it’s up for debate exactly which two words. The “Gyo” part could be part of the Japanese word for Goldfish (Kingyo), or it might be a transliteration of “Gulp” or “Guppy.” "Pin" might be a shortened form of "pinpin," meaning lively. Gyopin also resembles the words “Gulping” and “Guppy.” The latter would make sense, as Goldeen and Seaking are also named after real breeds of fish, those being the Tosakin and the Azuma Nishiki respectively. For awhile, I though that "pin" could also have been a transliteration of the English word for “Prince.” If that was the case, then the three stages of the Goldeen family actually form a punny sequence: Gyopin is a guppy prince, Goldeen a Goldfish queen, and Seaking is a combination of Sea and king. As much as I like this, it's probably not right: Seaking’s Japanese name preserves the “king” reference, but Goldeen’s doesn’t, so it’s a sequence that only makes sense in English. Like Puchikoon, Gyopin was originally the earliest evolution of a three stage line in Generation I, and Gyopin probably got removed for the same reasons: compared to Goldeen, Gyopin is more or less just a smaller and less fantastical version of his (admittedly, pretty bland) second stage relative. Brought back, likewise, it was supposed to be used to show off the new “baby Pokemon” mechanic of Gold and Silver. But also, like Puchikoon, Gyopin didn’t survive past Spaceworld ’97, and it’s completely gone in the next set of data we have, mid-1999. Unlike Puchikoon (and Mikon), however, we actually have moveset data for the version of Gyopin from Generation I. It’s…rough, like the rest of the draft movesets we have from Generation I. I'll put it here, right next to its later (and much more consistent) Spaceworld '97 moveset: Despite there being at least three years and tons of design innovations between these movesets, they are remarkably consistent. Gyopin learned Supersonic at an early level in its original conception, and continues to learn Supersonic at level six in Spaceworld '97. The first thing that popped out to me was that learning Horn Drill at level 19, before any other attack moves (maybe the Unknown move is a damaging move) is a completely bizarre design for a moveset. The idea of this tiny ‘lil guy one-hit KOing a Rhydon with a carefully timed Horn Drill is a bit hilarious to me. However, upon comparison to the Sapceworld '97 moveset, this isn't actually that strange: in Spaceworld, Gyopin learns Horn Drill at level 28, only nine levels later. Gyopin doesn't learn Horn Attack and Fury Attack in the original drafts, presumably because these stronger physical moves were set aside for Goldeen to use in the R/G early draft (maybe for the same reasons we talked about above). It's possible, however, that that Unknown move it learns first was an equivalent of Peck, maybe one designed for fish and not for bird beaks. And there are a lot of other oddities in this early moveset: Leer is learned late, after even Horn Drill; Bubblebeam is learned early but Water Gun is never learned; Goldeen gets Thrash despite not having claws to Thrash with, etc. Still, a lot of the final ideas for the Goldeen line are already in place. Overall, this early moveset shows that while a lot changed between the development of Generation I and the final, Gyopin's concept itself remained interestingly consistent between Generation I and Spaceworld '97. With a little work, we can actually learn quite a bit more about these reused designs as well: in particular, we have all the tools we need at our disposal to uncover what the original stats of Puchikoon, Gyopin, and Mikon were in Generation I. I originally discovered this while reading comments on Helixchamber’s analysis of the Generation I source code; someone in the comments had discovered a pattern in the way that stats were allocated to Pokemon in Generation I, and found that this pattern was broken in precisely all of the cases where an evolutionary stage had been discarded (plus a few more suspicious cases). Rather than explain the pattern, I’m just going to quote the comment, to give credit where it’s deserved: “Here’s something interesting regarding some of those deleted evolutions. Pokémon get stat increases when they evolve, but for gen 1, Satoshi (who I’m pretty said was the one in charging of determining how strong a Pokémon would be) seemed to have had a pattern. For Pokémon that evolve twice, the first evolution gets around a 75 stat increase (15 avg), and the exceptions still follow that 15 pt increase for most of their stats (the big exceptions are Dragonair, and the two cocoons, and those make sense because they’re special being the dragon and two very early evos.). For the second evolution, that stat increase will either be still 75 pts (15 avg again pt) or 100 (a 20 avg; this is true for the starters, the nidos, and the Pidgeot). The dragon line and the two bugs are exception again (Vileplume only increases by 70, but most of its stats beside speed increase by 15 pts, so it follows the pattern for the most part). For Pokémon that evolve once, the stat increase is 125 pts. But there are exceptions, and when I look at them I find something interesting because quite a few are Pokémon with removed evolutions. Rapidash and Marowak have stat increases of 75 pts in the final version. That’s low for a two stage evolution line, but it fits the pattern of a three stager. Ninetales and Golbat have an increase of 175, and that makes sense if Vulpix and Zubat were given the stats of their pre-evolutions (Zubats stats are particularly low as is). Golduck increases by 150 pts, meaning that two 75 stat increases are possible, as if the middle evolution between it and Psyduck was cut out. Seaking has a stat increase of 100, like the final stage version in a 3 parter. If those back sprites are fake, then the person making them did their homework. Granted, Magneton and Persian follow the pattern for a 2 stage line and they have some other forms it seems. But Magneton and Persian are odd because while other lines were inserted next to each other, these two aren’t. My guess is that either these other stages were never given stats to begin with or they were deleted rather early and Satoshi made adjustments. I can’t explain Gorochu either because Raichu gains 135 more points upon evolving. There are also a few also have stat increases that suggest a deleted form but likely don’t have one for the missing nos. Fearow, Arcanine, Tentacruel, Exeggutor, and Starmie, and I have no answers for them besides guesses. I’m making a lot of assumptions, but if my hypothesis is correct, than Game Freak had already had a lot of stats written done before they started removing many Pokémon, and its kind of disappointing you weren’t able to get them. Also that first encounter table, if it is real, might the oldest thing you got given it only contains monsters from the Capumon pitch.” -Paul, February 24th, 2019 This comment, by the way, was what proved to me beyond a doubt that the source code leak was legitimate and not an elaborate fake. As far as I can tell, no one had noticed this pattern before Paul, or at least before the leak of the Gen I source code. But the fact that this stat pattern lines up so perfectly with the cut Pokemon that we found in the link strongly suggested that these were real sprites. Afterall, how could a faker have accurately predicted a pattern that no one had yet noticed? The other implication of this pattern is that, had anyone discovered it earlier, before the leak, they could have used it to predict a number of deleted ‘mons; in other words, we could have discovered some of this information much earlier! That’s incredibly cool to me; there was hints and signs of the deleted Pokemon in the final games, and we just never quite noticed. But back to Gyopin, Puchikoon, and Mikon. Using this pattern, we can do a little math to figure out the original, stats. Ponyta, Vulpix, and Seaking all fit very neatly into the three-stage evolution stat pattern Paul mentioned above, especially once you realize they had a discarded first form. In the case of Mikon, it looks like Vulpix got Mikon’s original stats once Mikon was deleted; in the case of the other two, they kept their original stats. Ponyta was especially buff for a first stage Pokemon; Goldeen’s already about in line with other two-stage Pokemon families as is, which means that either Gyopin was very weak (as a common fish Pokemon, it’s possible) or Goldeen and Seaking got some slightly revised stats after it was deleted. Without further ado, then, here are the original stats of these three stage lines: If those stats are correct, and I see no reason to doubt them, then we have some sense of how powerful these Pokemon originally were in Generation I. Mikon was originally in line with Charmander, Squirtle, and Venonat, while Vulpix was initially much more formidable, fitting in with second-stage Pokemon like Gloom, Ivysaur, Charmeleon, Wartortle, and Kadabra. Puchikoon was a bit stronger than Mikon, fitting in more with Geodude, Machop, Magnemite, Bellsprout, and Horsea, but with higher Speed than any of them and Attack comparable to Geodude. Gyopin, on the other hand, had stats that resembled a lot of early game trash encounters: It’d have been about as strong as Pidgey, Rattata, and Zubat. This is good evidence that Gyopin was probably originally considered an early game encounter; it even has the very basic “Animal+” design that these early encounters all possess. I wonder if another reason it got removed was because an early game Water Pokemon wasn’t really needed. To encounter Gyopin, you’d need at least the Good Rod, which is received about halfway through the game, far after Rattatas and Pidgeys are outclassed. Since Water-type Pokemon were mostly tied to fishing, maybe the layout of the game made a Water-type this weak simply unnecessary. It goes without saying, of course, that these stats wouldn’t have been at all similar to their stats in Generation II, had they gotten to that point. In Generation II, these Pokemon fit a very different niche, and their stats (especially Mikon’s) would have been much lower. Gyopin didn’t make it past Spaceworld ’97; like most of the baby Pokemon introduced after Period 2b, the team decided that they didn’t really fit into the gameplay of the final games and took up valuable slots in the Pokedex. It’s not a huge loss: Goldeen is hardly the world’s most popular Pokemon and no one was really clamoring for it to have a pre-evolutionary relative. At this point in the Korean Index, the team was probably just drawing on old designs to fill up the Pokedex, so it’s no surprise they would abandon this design later: why reuse old ideas that you already discarded when you can come up with new ones? ID #396: MikonOur third baby Pokemon, drawn from the past to have a second life in Gold and Silver is the fan favorite Mikon! Mikon’s adorable, and even though it is visually similar to Vulpix, it’s probably the one baby Pokemon that I’m most sad to lose. Like its friends Puchikoon and Gyopin, Mikon appeared in Spaceworld ’97, but was purged from all the future versions of the game we have, along with most of the other baby Pokemon. (Credit for this Mikon art to the wonderful @RacieB) It’s name is pretty straightforward: “Mi” is three in Japanese, while “kon” is a shorted onomatopoeia for the sound a fox makes when it barks. So “three woof.” While I don’t normally pay attention to the unofficial localized names that Nob Ogasawara came up with for the lost Spaceworld ’97 Pokémon, his idea for the Mikon family is really cute. As the original translator of all mainline Pokémon games up to Platinum, it's neat to have his input on what he would have called those guys. If Nob had had his way, he would name named the three Vulpiii (Mikon), Vulpvi (Vulpix), and Vulpix (Ninetales), with the last letters of each name being the Roman Numerals for the amount of tails each one has. Those are great names! I’m absolutely in love with them. While Puchikoon and Gyopin were designed late into Generation I’s development, Mikon, Vulpix and Ninetales were a much earlier addition to the game, added right as the designers started experimenting with the concept of evolution. Period 3 of Red and Green’s development is mostly composed of two-stage evolutionary families, including Pikachu and Raichu, Horsea and Seadra, Kabuto and Kabutops, as well as two deleted families: two fighting fish Pokemon, and two Pokemon that were probably collectively reworked into Politoed. The Mikon family appears early in this sequence and is the first example in the Internal Index of a three-stage evolutionary family in the games; in fact, it was probably an experiment in the idea of three-stage evolutions, much like Eevee and friends were an experiment with a split evolution family, just twenty entries later. If you were going to design a Pokemon concept as the proof-of-concept for three stages, the Mikon family seems designed for that. All three are based on the supernatural Japanese creature called a “kitsune,” which is a fox spirit, known for its wisdom and beauty. The most distinctive feature of the Kitsune is its tails. Kitsune look like foxes but have multiple tales; the more they have—up to nine—the longer lived and wiser they are. Thus it makes perfect sense to have a three-stage evolution where each form gets stranger as it grow more tales: and if you were making your first three-stage evolution, making them kitsune with three, six, and nine tails lends itself perfectly to three stages. Together, they denote a very straightforward progression: as the Pokemon gets older, it grows more tails and thus becomes wiser. Certainly, if I were taked with coming up with the first three-stage Pokemon family, I couldn’t do much better than a Kitsune that evolves as it grows more tails. Unfortunately, the family loses some of that flavor after the team removed Mikon: Vulpix does grow more tails when it evolves, but the progression is less accentuated. There are many stories of kitsune; in most of them, the Kitsune is a trickster figure who punishes humans who are cruel or unkind. In some, kitsune can transform into beautiful women and may marry humans. Next to the Tanuki (discussed in a previous article), they’re one of the most distinctive creatures in Japanese folklore. Whether intentional or not, Mikon does seem to be referenced in the final game, obliquely, through Vulpix’s Green and Blue Pokedex entry: “When it is born, it has just one snow-white tail. The tail splits from its tip as it grows older.” Now, obviously, this is a pretty obvious piece of lore to extrapolate from Vulpix: if it has six tails, then it follows that it would start with only one and more tails would grow from it. But at the same time, it’s hard to ignore that Mikon has three tails, and in its Gen II sprite, those tails seem to be sprouting out of one main trunk, just as the Pokedex entry said. It isn’t the first time a Pokedex entry has referenced a cut Pokemon: Raichu’s original entry infamously mentioned an Indian Elephant, which was probably originally a reference to the elephant Pokemon that didn’t make it to the final game. So it’s very possible that this entry was written with Mikon in mind. When it was added to Generation II, Mikon—bizarrely—was added as a Water-type. This is almost certainly an error; it’s easy to get typing wrong when programming things in, and the Fire and Water types are literally right beside one another in the code (being type IDs 14 and 15 respectively). Less likely, but still possible, is that Mikon was a last minute replacement of a Water-type Pokemon that had previously held that slot; maybe ID 344 (Viking Ship) or 380 (Drill Squid) were originally here (I suspect 344 was developed pretty far and dropped right before Spaceworld ’97). Overall, it’s a pretty insignificant detail, but it’s worth pointing out. Unfortunately, there’s not much more to say about Mikon. It has the same movelist as Vulpix, of course, just learned at an earlier level; the only significant note about it’s moveset is that it learns Flamethrower at the pretty early level of 37, as opposed to Vulpix’s 53. In Spaceworld ’97 this is a pretty significantly lower level than everything else (it’s normally learned in the forties or fifties). But in the final game, they seem to have changed their minds; most Pokemon learn the move at the same time Mikon would have, and Vulpix in particular now learned it at level 31! While I’m sure we all would have liked to see more of this little cutie, it was sadly not to be. Like the rest of baby Pokemon introduced here, Mikon was gone by June of 1999, replaced by newer designs. ID #397: ???Here’s a strange one. This little guy looks like an oddly posed frog with a horn on his head. A completely unused design, at first glance, his basic animal nature makes 397 look a bit like the unused (and underdeveloped) designs we've seen before in Period 1e and 1b (and will see again in Period 2d). It'd be easy to dismiss 397 as another throwaway design, which would make sense in the middle of Period 2c, since the team was probably just filling up the last couple slots of the new Pokedex. And yet, something about this guy--his location right next to three throwbacks from Gen I, his pose being familiar, and why he is wedged between other used designs--makes me wonder if he has more significant origins than he looks like. To be clear, the following theories I'm about to propound have a lot of speculation built into them, more than is usual for the Cryptodex. So please join me for the journey, but I'm aware that they're founded on less than usual. First, what we do know: 397 is probably designed after a pun. If you'll notice, the top of 397's head has a tiny horn; it's a bit hard to see, and easier to see on the backsprite. In fact, this is probably 397's only distinguishing feature: without it, it's literally just a cartoony frog. But why a horn? It's probably a joke on the idea of the "Horned Toad," a common type of lizard in the United States (especially Texas), and ironically not an actual toad. The visual design of 397 is probably a literal rendering of the name: what if we had an actual frog (toad) with a horn on its head? (On the left is the logo for Texas Christian University's whose mascot is a horned toad. I've always said that it looks like a rejected Pokemon design) There are other peculiarities to read into the sprites as well. @OrangeFrench suggests that its arms look a bit like they're sticking onto a wall, much like a treefrog might. And 397's tongue brings to mind a passing resemblance to Lickitung, which also had a long tongue. Unlike Lickitung, which looks ready to slurp something or lick someone on the face, this tongue looks more stringy and maneuverable. Maybe it was originally supposed to use its tongue like a lasso? It's also seemingly based on a tree frog, fitting in with a lot of the other exotic animals that seemed to inspire a lot of the unused designs found in the Korean Index. The team came up with a design for a Sunfish 'mon, all kinds of exotic birds, a scorpion, and even a Manta Ray and a puffer fish. If the team was trying out designs of distinctive foreign animals, a tree frog from the rainforest fits right in with the brainstorming we've seen so far. I’ve stared at this guy a lot, trying to divine his mysteries. My first theory was that he was, in fact, a heavily redrawn version of the long-lost Crocky, an unused crocodile Pokemon from Generation I which had bulging eyes and a goofy expression: (Art is, like always, courtesy of the great RacieB) There are a few reasons to think this. First, the previous three Pokemon were all unused designs from Generation I. That is some evidence to me, admittedly not a lot, that the team might have reached back into unused designs when creating 397. If it was the case , Crocky is the old design that most resembles this little frog guy. They both have the same thin arms and legs, and they have a weird half-sitting pose, suggesting a similar artist or inspiration. Look at the left leg on each of them: it's hard to see in the blurry sprite we have for Crocky, but it’s more or less identical in style and pose. Crocky has hair in the sprite we have, but we know from the source code that he was reworked to lose the hair, which would leave his face looking a lot like 397. It’s not impossible. In the end though, I don't think this theory is right. 397 is clearly a frog with a long tongue; Crocky is a crocodile, as evidenced by its pretty on-the-nose name and its snout. Maybe Crocky was reworked, but if so, there's almost nothing of the original design. The more I considered it, the more unlikely it became. At the same time, there’s something going on with 397. First of all, it is found in a block of baby Pokemon repurposed from Generation I. That may be a coincidence…but look at the size of the sprite for 397. 397 has the smallest sprite size (44x44), which is the same size as the three preceding baby Pokemon. A lot of the designs in the Korean Index use this small sprite size, so it’s not definitive at all, but the next two designs, Twinz and Ledyba, both do not, suggesting that 397 has more to do with Mikon, Puchikoon and Gyopin than it does with the following Pokemon designs. It could also be potentially misplaced, and really fit as part of the upcoming Period 2d, which is characterized by a lot of unused, strange, simple Animal+ designs. In fact, if Twinz and Ledyba were actually added to the Index later and overwrote two more basic designs--or if they were updated versions of some simple designs that fit in better with Period 2d—then 397 might just be the beginning of that Period of discarded designs. So I could be making something out of nothing with 397. But let’s assume for the moment that 397 has something in common with the three entries preceding it. If that’s the case, there are a couple possibilities. First what if it was another attempt at a baby Pokemon? The obvious problem is that there isn’t anything in the Korean Index that is obviously an evolution for this guy. Potentially, 397 is a redesigned version of the preevolution of Politoed found in the Generation I backsprites, and that it was supposed to evolve into the earliest sprite of Politoed we have. Unlikely, given that Politoed looks like it was brought to Generation II in order to pioneer the concept of a split evolution, but it is definitely possible. The coloring of the two of them is actually quite close; at the very least, they look more similar than this Politoed sprite did when compared to the Poliwag line: I removed the spiral on Politoed's belly because it was probably added after it was connected to the Poliwag family, and well, damn. These two look much closer than I expected. We know that Politoed had a pre-evolution in Generation I, and we don't have the front sprites for it. Could this be it? There's some obvious differences--the eyes are different, 397 has a different style of shading in its hands, etc--but that could be an artifact from two different artists redrawing these sprites into the Korean Index. Honestly, I think I convinced myself just look at this. (Could the mystery of these two backsprites from the Pokemon Blue source code finally be solved? Maybe, but if so, 397 must have been heavily redesigned, and its top-knot changed into a horn.) My other theory has been that 397 is not a redesign of Crocky, but the pre-evolution of Crocky, repurposed and dropped in the Korean Index. This would make sense of their visual similarities, but also explain why they are such different beasts from each other: one is an extension of the other! The obvious rebuttal to this idea is that there never was a pre-Crocky Pokemon in Generation I, at least in the source code leak that we have to look at. And it's a good point. Crocky was created in Period 2 of Red/Green’s development, before evolution was a concept, and another deleted member of the Crocky family appears nowhere in the later parts of the index. But there are three unaccounted for Index slots, two of which could potentially be hiding a relative to Crocky. Take a look at the poster I made on the development of Generation I, if you’re rusty on Generation I’s development (and you probably are, if you’re not obsessing over this stuff like me). Or, conversely, just go read Helixchamber’s great articles! (This one, in particular, discusses the idea I'm about to suggest). Internal Index slots 115 and 121 (In Period 4 on that poster) are both unaccounted for; the source code for Pokemon Blue that was leaked had no data in either of those slots, suggesting that they were deleted much earlier than the rest of the deleted 'mons. These two are both found in Period 4a, which is composed of Pokemon designs that are all evolutions of Pokemon between #57 and #70. This means that, whatever 115 and 121 were, they were very likely to be evolutions from that range. However, there's not a lot of options in that range that don't already have evolutions: Tauros, the Elephant, Crocky, Farfetch'd, Cactus, and Jagg are basically the only options that could have a hidden evolutionary relative (Dragonite, Koffing, Magenton and Deer are also possible, but very unlikely). Helixchamber thinks that one of the two was probably Madame, the evolution of Farfetch'd found in Spaceworld '97, and I basically concur with that assessment. The other one could be an evolution of any of the above, but the most likely candidates are either Cactus or...Crocky. So it's feasible that something related to Crocky was hiding here, and just deleted earlier than the other unused 'mons.
In the end, I think this is basically unprovable. While I see some similarities in their designs, there's not really any evidence that can support this theory. It's possible, certainly. So, to sum up, we have three options. First, 397 could be the original pre-evolution of Politoed, before it was merged with the Poliwag family. On the one hand, the sprite comparison I did feels pretty convincing to me; on the other hand, 397 looks almost nothing like the backsprite we have, so it'd have to have been redesigned. Second, it could be the relative of Crocky. There's some similarities in their designs, but admittedly, the theory relies on a lot of leaps of logic. Or thirdly, and most likely, 397 was just a frog, based upon a too-literal interpretation of the Horned Toad. As Freud would say, "Sometimes a frog is just a frog." |
AuthorMy name's Aaron George, and I'm both a historian and a fan of Pokemon, especially of development. Reach me at @Asmoranomardic ArchivesCategories |