Let's start with the part that matters most:
in the overwhelming majority of positions we agree with the other platforms. What is a
there is a
here, the best move is the same best move, and your accuracy usually lands within a point or two of what you would see elsewhere. What can differ is the label on a handful of moves in a game. We might hand out a
one time more — or one time fewer — than you are used to, or flag a
that another site recorded as simply the best move. Here is why. Game Analysis is built on
Stockfish, the strongest chess engine in the world, and every raw evaluation comes from it. But the badges that sit on top of those evaluations come from our own
chessrank.org mathematical model. We developed that model ourselves and deliberately did not train it on other platforms' output, so it is not a copy of anyone else's labels. Two things follow from that. First, somebody has to decide where a "good move" ends and a "great move" begins — and every platform draws that line in a slightly different place. Second, an engine is not an oracle: give Stockfish a little more or a little less time on the same position and even it can change its mind. So a different badge here does
not mean our analysis is wrong, or that the other one is. It is the same engine evidence read through a different lens — and the conclusions that actually matter, namely where you lost the game, which move threw away the advantage and how accurately you played, hold up either way. That said, we are not infallible. If you ever see a result that is genuinely strange and makes no chess sense, please report it to us using the form below. Every report reaches us and helps us make the model better.
