Why did Chess.com call my move brilliant?

• Chess.com's classification article (updated February 9, 2026) says a Brilliant is a good piece sacrifice, and that it is always the best or nearly best move. You should not be in a bad position after it, and you should not have been completely winning without it.
• The same article is more generous when it defines a piece sacrifice for newer players than for higher-rated players. It does not publish a rating cutoff or how often the label appears.
• Best, Excellent, Good, Inaccuracy, Mistake, and Blunder come from expected points lost. Brilliant, Great, and Miss use extra rules and are not rows in that cutoff table.
• A Brilliant on the game-over screen can change in full Game Review. Chess.com says the summary is a faster analysis, and that web and mobile engine settings can differ.
• One Brilliant is a move to review. Improvement is a pattern over many games, which is how to know if you are improving.

Chess.com's Help Center article "How are moves classified?" (updated February 9, 2026) says it replaced the old Brilliant algorithm with a simpler definition. In its words: "a Brilliant move is when you find a good piece sacrifice." The same section says Brilliant moves "are always the best or nearly best move in the position," and that they are "special in some way."

Two more conditions are on that page, quoted here in full: "You should not be in a bad position after a Brilliant move." And: "You should not be completely winning even if you hadn't found the move." The article does not give a number for "bad position," "completely winning," or "nearly best." This page does not add one.

It also says: "We are also more generous in defining a piece sacrifice for newer players compared to those who are higher-rated." That is the rating rule. There is no table of which sacrifices count at which rating.

A shorter line on the Game Review Help Center page (updated September 25, 2026) still says Brilliant is "The best move - and tricky to find too!" That line does not say "piece sacrifice," and it does not say "nearly best." Where that one-line label and the February 2026 classification article differ, this page follows the classification article, because that is the page that states the definition after the old algorithm was replaced.

Chess.com's February 2024 article "How To Play A Brilliant Move" writes a Brilliant as !! in chess notation, and a Great move as !. Those symbols are from that article. The February 2026 Help Center page quoted above does not define the symbols.

The same February 2026 article says Classification V2 uses an Expected Points model. Expected points are winning chances from your rating and the engine evaluation. On that page, 1.00 is always winning, 0.00 is always losing, and 0.50 is even. It also says that at 1.00 you have a 100% chance of winning, and at 0.00 you have a 0% chance. After the move, Chess.com looks at how those expected points changed.

Brilliant is not one of the rows below. The article puts Brilliant, Great, and Miss in a later section, "beyond expected points." The cutoff table is only for the six labels it prints. Chess.com says a move gets the label whose row contains the expected points lost, between that row's lower and upper limit.

The limits meet. For example, 0.00 is both Best's upper limit and Excellent's lower limit, and 0.02 is both Excellent's upper limit and Good's lower limit. The article does not say which label is used when a loss lands on a shared number. This page does not invent a tie-break.

Great, on that article, is a move "critical to the outcome of the game," with the examples "turning a losing position into an equal one, an equal position into a winning one, or finding the only good move in a position." Chess.com says it is more generous about what it calls a Great move for new players than for higher-rated players. That sentence is about Great moves. The Brilliant sentence above is about how a sacrifice is defined. They are not the same rule.

A Miss, on that article, is when you "fail to capitalize on your opponent's mistake and miss the opportunity to gain a winning position, often resulting in an equal or worse outcome." The evaluation that counts as winning, equal, or losing "varies according to the player's rating, similar to expected points."

The Game Review page's one-line list, separate from the cutoff table, calls Best "The chess engine's top choice" and Book "A conventional opening move." Book is not in the expected-points table on the classification article.

Two different rating rules are on the February 2026 article. First, expected points already mix the engine evaluation with your rating, so the same evaluation is not the same winning chance for every player. Second, Brilliant uses a more generous definition of a piece sacrifice for newer players.

That is why a sacrifice in a lower-rated game can receive a Brilliant when a similar idea at a higher rating does not. Chess.com does not say which material counts as "a piece," and it does not say that a lower rating produces more Brilliants by any measured share. A count like that is not on this page.

The label also does not say that you saw the idea. The three Help Center pages opened for this guide do not describe an "accidental" Brilliant. The rules above are about the move, not about your intention.

The Game Review Help Center page says the summary right after the game uses a faster analysis than full Game Review. Opening Game Review runs a more detailed analysis, "which can result in different move classifications." Its example is the one people ask about: a move marked Brilliant at first "may later be classified differently after the full review."

A second Help Center page, "Why did my move classification change in Game Review?" (dated July 25, 2025), says the game-over modal runs a quick version of Game Review on your device, and that Chess.com displays the engine's initial evaluation without altering it. Full Game Review is processed on Chess.com's servers, and that deeper pass "can sometimes update or change the data."

The same July 2025 page says Game Review can run at different strength levels. On the website, the engine choice is the cogwheel between the board and the right-side panel, then the dropdown next to Engine. On mobile, it is the cogwheel at the top right of Game Review, then the Strength dropdown. It says only Diamond members have Deep and Maximum engine strength. If the strength differs between the website and the phone, the same move can get a different label.

The September 2026 Game Review page says the same thing in shorter form: different analysis settings by device or engine strength can change a classification, and Diamond members have deeper engine settings that can differ from a faster analysis. Both pages describe a second analysis that can reclassify the move. They do not say the icon was removed because the idea was bad.

Under the February 2026 definition, a move stays off the Brilliant list when any of the conditions fail. The checks below are those conditions, not an extra scoring system.

• It was not a piece sacrifice. A strong quiet move can still be Best or Great.
• It was not the best or nearly best move. "Best" on the Game Review list is the engine's top choice. Chess.com does not define "nearly best" as a number of pawns.
• You were in a bad position after it. The article states that condition and does not attach it to the cutoff table.
• You were already completely winning even without the move. A sacrifice in a position you had already won can be the best move and still not be Brilliant.

The February 2024 article adds rules the 2026 Help Center does not. It says a Brilliant is usually a sacrifice of a piece or an exchange (a rook for a knight or a bishop). It says that in the endgame the sacrifice must be the only good move, and that if you were winning in several ways the move may be best but not Brilliant. It says that in the opening or middlegame the move must be strong, and that it is all right if it is one of several good options. It also says Brilliant moves occur in competitive positions, and that if you have already won a lot of material a later sacrifice probably will not be classified as a brilliant combination.

The February 2026 article does not split that rule by phase. It says a Brilliant is always the best or nearly best move, in any position it is talking about, and that you should not have been completely winning even if you had not found it. Where the 2024 "only move in the endgame" and "several good options in the opening" lines conflict with that, this page follows the 2026 Help Center.

The 2024 article also says a surprising sacrifice with check is very often Brilliant, and it tells you to look for checks that win material or lead to a quick mate. That is a finding tip in the older article. It is not a promise that every check is Brilliant, and the 2026 definition still requires a good piece sacrifice that passes the conditions above.

The steps below are a review habit on this page. They are not a Chess.com setting, and they are not a rating plan.

1. Write one sentence before you read the coach: what the sacrifice attacked, or what it forced the opponent to answer. If you cannot say it, the label has not taught you the idea yet.
2. Play the opponent's replies yourself. Chess.com's Game Review page says that when a key move could have been better, you can Retry: replay the position and try to find the best move. Do that before you read the line. The engine-off pass is how to analyze your own chess games.
3. If the move was not a sacrifice, stop looking for a Brilliant explanation. Ask whether it was Best, Great, or a Miss instead, using the definitions above.
4. Keep the note with the pattern, not with a tally of Brilliant icons. A weekly plan across games is Game Review versus a training plan. Puzzle ratings and game ratings measure different things: puzzle rating vs game rating.

Lichess analysis will show an evaluation and a line, and it will not apply Chess.com's Brilliant rule. What that screen does and does not explain is why Lichess analysis doesn't explain itself.

No. It means that one move met Chess.com's sacrifice rules. A game can contain a Brilliant and still be decided by a later blunder. Chess.com's Game Review page says accuracy measures how closely your moves matched the engine's preferred moves, and that it should not be viewed as an overall measure of your playing strength. Counting Brilliant icons is a narrower version of that mistake: one marked move, treated as a trend.

Judge the trend from many games in one time control, and from the mistake that repeats. That method is how to know if you are improving at chess. Chasing another Brilliant is not one of the things that page tracks.