Difficulty curve

Your AI works out how hard each level really is from what players did, counting each player once, and compares it with how hard you meant it to be. It finds the level that is suddenly much harder than the one before, shows who tried again and who gave up, and says when there are too few players to tell.

When you need it

When a game feels too hard or too easy and you do not know which level is to blame, or when you want to check a level you just changed against your plan.

Say this to your AI

Use the difficulty curve from Spielerando on my event log. First write down, level by level, what share of players should win on the first try, and draw how a level plays out as a Mermaid state diagram with winning, losing, retrying and giving up. Then compare the real numbers with that plan. Show me players, first-try wins and sure ranges for each level, and say where the numbers are too thin to judge.

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

Settings

minPlayers (30): Under this many players, a level is called too thin to judge.
spikeFall (0.15): A fall in first-try wins of this much against the level before counts as a spike.

Your AI gets the code

The finished code and math come through the Spielerando connection, so your AI can use them right away. Connect your AI

Common questions

which level of my game is too hard

Compare the share of players who win each level on the first try. A sudden fall against the level before is a spike, if the two ranges do not overlap.

why count first tries and not all attempts

A struggling player can lose forty times and a good player once. Counting attempts lets a few people drown out everyone else, so each player counts once per level.

why do my later levels look easy

Only players who got that far are counted, and those are the better ones. Later levels would be harder for everybody than the numbers suggest.