Rank maps by how often players give up on them when some maps have very few games

A map with two games, both abandoned, looks like the worst map in the game. A map with a hundred games and one in five abandoned is the real problem. This entry shows how to rank maps so that thin data does not shout louder than thick data.

Why this works

Little data is mostly luck. A place with two games and two give-ups looks terrible but has barely been tested. Pulling each rate toward the overall rate, harder when evidence is thin, lets only places with real history rise to the top of the list.

When not to use it

Do not use it to decide whether one particular map is truly bad, or when every map has plenty of games and the pull changes nothing. Then use the plain rate and read the maps themselves; the adjusted value only helps order a list.

In the full entry

• What to build, step by step
• Starting values and the reason for each
• 3 common mistakes and the fix
• 15 lines of example code with 3 checks that pass
• How your AI shows you that it works

Say this to your AI

Use the Spielerando entry "Rank maps by how often players give up on them when some maps have very few games" (read_entry map-forfeit-ranking-small-samples) in my game. Build it step by step, and show me how you checked that it works.

Or skip this and connect the MCP: your AI then finds all of this by itself.Connect the MCP