4 ms·

This is actually a category of moves on sites like Chess.com. They categorize moves into things like blunder, mistake, inaccuracy, good, best. But the top mov

This is actually a category of moves on sites like Chess.com. They categorize moves into things like blunder, mistake, inaccuracy, good, best. But the top move is 'Brilliant' and the second best is 'Best Move.' Best move is the move the chess engine thinks is the best, and brilliant is the category for moves that weren't considered, but turn out to be the best. They're uncommon, but if you play enough, you'll hit them occasionally. https://support.chess.com/article/364-how-does-the-game-report-analysis-work https://support.chess.com/article/364-how-does-the-game-repo...

nebulous1 5y ago
Did a little reading and it seems that chess.com are a little ambiguous as to exactly what they're doing here. Two plausible theories are 1) They run an evaluation of the position to a certain depth and come up with the Best move. Then, when evaluating your opponent's position if you didn't make the Best move, it turns out that it evaluates this position as even more in your favour than the Best move, it calls your previous move Brilliant. 2) They run an evaluation of the position up to a certain depth and note the Best move as the most advantageous one. They then search a bit deeper and if the move you actually took wasn't the best in the original eval but now is, they call it Brilliant. There would of course be a few caveats to both ideas. In 1 do they go and rerun the evaluation after the Best move to compare? Are they running the whole analysis backwards anyway (this seems to cause problems with both systems)? My suspicion would be number 2 given the way it's phrased in your link, but I've seen other descriptions that make it sound more like 1. My overall takeaway is that it's simply a bad idea to have the category to begin with as it just confuses matters for little gain, at at least they should provide a more technical description. I'd wager that close to 0% of Brilliant moves were intentionally calculated so there was no brilliancy as the term is usually used in chess.
rurban 5y ago
In the chess literature this categorization is established as !!, !, !?, ?!, ?, ?? for over a century now. Everybody loves the !! and ?! moves which do not only evade chess engines but also humans.

nebulous1 5y ago

Did a little reading and it seems that chess.com are a little ambiguous as to exactly what they're doing here. Two plausible theories are 1) They run an evaluation of the position to a certain depth and come up with the Best move. Then, when evaluating your opponent's position if you didn't make the Best move, it turns out that it evaluates this position as even more in your favour than the Best move, it calls your previous move Brilliant. 2) They run an evaluation of the position up to a certain depth and note the Best move as the most advantageous one. They then search a bit deeper and if the move you actually took wasn't the best in the original eval but now is, they call it Brilliant. There would of course be a few caveats to both ideas. In 1 do they go and rerun the evaluation after the Best move to compare? Are they running the whole analysis backwards anyway (this seems to cause problems with both systems)? My suspicion would be number 2 given the way it's phrased in your link, but I've seen other descriptions that make it sound more like 1. My overall takeaway is that it's simply a bad idea to have the category to begin with as it just confuses matters for little gain, at at least they should provide a more technical description. I'd wager that close to 0% of Brilliant moves were intentionally calculated so there was no brilliancy as the term is usually used in chess.

rurban 5y ago

In the chess literature this categorization is established as !!, !, !?, ?!, ?, ?? for over a century now. Everybody loves the !! and ?! moves which do not only evade chess engines but also humans.