This project started as PokeIdleon. A fan game. If you'd opened the game folder a year ago, you'd have found 323 Pokémon sprites, the PokéAPI as the data source, the usual 18 types with their 18×18 matrix, Kanto/Johto/Hoenn as the map, Eevee and the Eeveelutions, Charizard, Aerodactyl, the four legendary birds, Mewtwo, Mew, an Elite Four, three clans called Rubra, Cerúlea, and Áurea designed on top of the gym color palettes. A pre-built corral. Instantly recognizable structure.
And I didn't drop it out of fear of a lawsuit, even though the IP risk was real and had been documented since v0.5. I dropped it because, when I tried to justify every system built on top of that skeleton, the answer was always the same: "because that's how Pokémon does it." Not "because it makes sense for my game." That's not design, that's decorating a house that isn't mine.
The heavy decision wasn't swapping names. It was realizing I'd have to rebuild the entire roster from zero — 120 species, 52 concepts, with no recognizable silhouette to lean on — and that it was going to hurt acquisition. Losing instant recognition has a real cost: nobody searches "ESVA" on Google in January 2026. They searched "Pokémon idle." That's the first metric that dropped, and it dropped on purpose.
What died, listed out, because listing hurts more than summarizing
I went through line by line to figure out what had to go, because "let's just rebrand it" is easy to write and folds the second you have to decide whether the Fossil Dungeon survives. It didn't survive. Everything went: the 323 sprites, the PokéAPI as the canonical data layer, the 18 types, the 18×18 matrix, all the geography (Kanto, Johto, Hoenn), every proper creature name, the Fossil Dungeon built on top of the Amber Fragment/Aerodactyl, the Evolution Dungeon with the Eeveelution Stones, the four legendary birds, Mewtwo, Mew, the "Elite Four," the three clans as they'd been designed, Butterfree Alpha, the entire Gen 1-through-9 roadmap, and even the fan-game disclaimer — because without the IP, the IP-risk paragraph doesn't make sense either.
None of it stayed. Not as an aesthetic rule. Because a "recognizable creature" is, by definition, someone else's property, and the entire game I wanted to make is about a creature that breaks a law of physics, not about an animal you already recognize from somewhere else.
What survived, and the mistake I almost published
The only thing that made it through was a pack of 200 attack animations — explosions, generic elemental effects I'd bought in a pack called pokeidleon-skills-gen1-3-complete-v1. Of those 200, 142 already came typed (candidates to become the 13 new classes) and 58 were marked "Normal" in the original manifest — I wanted to reuse those as the classless moves in the new system (§6.4 of the GDD, if you want the section number).
Here's the mistake I almost let slide. In some earlier document — I won't pretend I remember which — I wrote that "190 of the 200 animations fit the 13 classes, plus 58 marked Normal." Add that up: 190 + 58 = 248. In a pack of 200 pieces. I published that math, nobody on the team (which is me, alone, so "nobody" is literal) caught it, and it sat there until I went back to actually audit the pack for real and the arithmetic simply didn't close. There's no way 200 animations produce 248 candidates. One of the two counts was wrong, or both were.
I redid the count manually, piece by piece. The real number is 142 typed + 58 "Normal" = 200. Matches exactly. The "190" never existed — it was a typo or a memory slip that I treated as fact for way too long, because the sum "felt right" without me ever actually having added it up.
This isn't a little story of gratuitous embarrassment. It's the reason why, starting that day, every count that goes into a canonical ESVA document gets a sum check before it turns into a decision — a dumb rule, but it's the rule that exists today because of this exact mistake.
What's still broken
The final breakdown of which of the 142 typed animations I actually keep, versus which I discard during the visual audit and the renaming pass into the 13 new classes, still isn't settled. That's pending as of today, not devblog rhetoric — it's real work that has to happen before I can say "these are the game's attack animations."
What the game gains from this
The Registro becomes an obligation, not a collection — you're not gathering critters, you're documenting an ongoing harm. The public odds become the Ordem's scientific method, not a shop's fine print. The Market becomes a real consequence of genuine rare specimens, not a decorative system. Brand, lore, and system now explain each other, because they were born together — not because one got glued on top of the other afterward.
The payoff for dropping instant recognition is that now, when someone asks "why does this creature exist," I have an answer. Before, the answer was always "because the original game had one like this too." That was never an answer. It was an echo.