PTGigi

So I was thinking to myself: ‘It’d be awesome to fuse EVERY SINGLE Pokemon–bracket style. Pair off each Pokemon family, fuse them, then pair off their fusions and fuse them again! Repeat until you have one left!

And so this project was born!

Links:

How Pokemon are Paired:

I was partially inspired by the Pokemon Tree of Life project, so the aim for choosing fusions was always to pair similar Pokemon families. I didn’t want to base it off an existing Tree, and the work to make my own Tree was daunting, so I opted for the next best thing: create an algorithm to pair the most similar Pokemon for me!

To start, the algorithm calculates a similarity, or happiness, score (0-100) between all possible pairs (link can lag on older computers!). It bases this score on a comparison between the Pokemons’: body type, color, stages/forms, and elemental types. With those calculated, it pairs Pokemon who’s best matches are one another. Those paired are removed as best matches from those remaining, who move onto their second best choice. This continues until all Pokemon have been paired for that round/generation. If there’s one left without a pair, it gets a bye to the next round/generation of fusions.

FYI: I discovered some incorrect data with my initial sample too late into the process to correct without essentially having to restart. I’ve taken this as a lesson for the future and will proceed with the current dataset so I can finish this project in a timely manner.