This hits a real pain point. The "fake precision" issue is exactly why most SEO-driven team-builder tools fall apart in practice — they treat stat weights as if they're known constants when the underlying game is a black box, so the output looks scientific but is really just guesswork dressed up with decimals. A proper multi-role knapsack framing (where each unit competes for limited slots across multiple roles, not a single dimension) is a much more honest way to model it, even if it's messier to present to users than a clean ranked list.

I've run into the same problem maintaining a small reference page for Undertale — most "guides" online present numbers with false confidence when the actual mechanics are far more situational, so I ended up keeping things intentionally simpler and clearly caveated instead.

Learn More: https://theudertale.com/