Lichess game review is 100% free and unlimited — request a computer analysis on any finished game and Stockfish, an engine rated above 3500 Elo, annotates every move in seconds, flagging blunders, mistakes, and inaccuracies plus your accuracy and ACPL. To use it well: read the ACPL as a trend not a verdict, click each flagged move to see what you missed, and use "Learn from your mistakes" to make it active. Its one blind spot is that it reviews one game at a time — it can't tell you the weakness costing you rating across your last hundred games. For that you have to aggregate.
Lichess has one of the best free analysis tools in chess, and most players use maybe 30% of it. They request the computer analysis, glance at the accuracy number, nod at the red blunder markers, and close the tab. The engine did its job; the player learned almost nothing. This guide covers how to actually run a Lichess game review, how to read what it gives you, how it stacks up against Chess.com's Game Review, and the one thing single-game review structurally cannot do for you.
Lichess "game review" is its computer analysis feature: request it on a finished game and the Lichess servers run Stockfish — the open-source engine rated well above 3500 Elo, far beyond any human — over the whole game and hand back an evaluation for every move. Unlike Chess.com's Game Review, the full feature is free and has no daily cap, because Lichess is a non-profit, open-source, ad-free project. You get the same engine strength a titled player uses, at no cost, on unlimited games.
The output has three parts: a summary (each side's accuracy percentage and average centipawn loss), per-move classification (blunder, mistake, inaccuracy, plus "good," "brilliant," and book moves), and an interactive engine line for any position so you can see what you should have played. There's also a "Learn from your mistakes" trainer that replays your errors as puzzles.
The whole thing takes under a minute:
You don't even need to play on Lichess to use it. Paste any PGN — including games exported from Chess.com — into the Lichess analysis board and it will analyze them for free. That makes it a universal engine-review tool, not just a review of your Lichess games.
This is where most of the value is left on the table. The numbers only help if you know what they mean.
ACPL — average centipawn loss — is the average evaluation you surrendered per move, where 100 centipawns equals one pawn. Lower is better. It's a useful fitness signal over many games, but on any single game it's noisy: one wild tactical melee can spike it, and a dull, drawish game can make weak play look clean. Don't chase the ACPL number. Use it to spot trends across dozens of games, and always look at the actual mistakes underneath it.
Lichess marks moves as blunder (??), mistake (?), or inaccuracy (?!) based on the evaluation swing they cause. Blunders are where games are decided — below about 1600, a large majority of decisive amateur mistakes are one- or two-move oversights, not deep strategic failures, which is why the blunder list is the single highest-value thing on the page. Click each one, cover the engine line, and try to find the refutation yourself before revealing it. Reading the answer teaches far less than reproducing it.
The honest comparison, since this is the question most people are really asking:
| Feature | Lichess game review | Chess.com Game Review |
|---|---|---|
| Price | Free, unlimited | Limited free; unlimited needs paid membership |
| Engine | Stockfish (cloud + local) | Stockfish-based |
| Move classification | Blunder / mistake / inaccuracy | Same, plus "great/brilliant" flair |
| Coaching tone | Terse, engine-first | Friendlier narrated "coach" walkthrough |
| Active practice | "Learn from your mistakes" trainer | Retry-the-move + key-moments |
| Cross-game patterns | No — one game at a time | No — one game at a time |
For raw engine quality they're a wash — both run Stockfish. Lichess wins clearly on price and openness; Chess.com's review is a bit more hand-holding if you like the narrated style. If you specifically want the Chess.com experience without the paywall, we cover that in our free Chess.com Game Review alternative guide. But notice the last row: neither tool reviews across games, and that's the real limitation.
Single-game review answers "what went wrong in this game." It cannot answer the more important question: "what goes wrong in most of my games." Those are different problems. You can review fifty games one at a time and still miss that you hang a piece to the same knight-fork pattern every week, or that your win rate collapses whenever the clock drops under two minutes, or that one opening quietly costs you a third of your losses. The information is spread across fifty tabs you closed, so the pattern never assembles.
This is why players who diligently review every game still plateau. They fix the specific blunder from Tuesday's game and make a structurally identical one on Thursday, because the review never named the category. Improvement comes from attacking recurring weaknesses, and recurring weaknesses only show up when you aggregate. For the full manual method, see how to analyze your chess games and how to find your chess weaknesses.
To turn one-game review into real improvement, aggregate your games and group the errors by theme:
Chess DNA does this aggregation automatically. It imports your full Lichess (or Chess.com) history, analyzes it with Stockfish, and surfaces your recurring weaknesses ranked by how much rating each one costs you — so instead of reviewing one game and forgetting it, you see the pattern and drill it. It's free to start and pairs naturally with Lichess's per-game review: use Lichess for the deep dive on a single game, and let Chess DNA find the pattern across all of them. For a broader tool comparison, see our chess analysis app guide.
Yes — completely free with no daily limit. Lichess is a non-profit, open-source, ad-free site, so its full computer analysis (accuracy, ACPL, blunder/mistake/inaccuracy classification, and the "Learn from your mistakes" trainer) is available on every game at no cost. This is a genuine difference from Chess.com, whose complete Game Review is limited on the free tier and needs a paid membership for unlimited use.
Open the finished game, go to the analysis board, and click "Request a computer analysis." Lichess runs Stockfish in the cloud and annotates every move within a few seconds, flagging your blunders, mistakes, and inaccuracies. You can also turn on the local engine to evaluate any position live as you move pieces. And you don't need to play on Lichess — paste any PGN into the analysis board and it analyzes it for free.
ACPL (average centipawn loss) is the average evaluation you lost per move — lower is better. Roughly, a clean club-level game lands under 40, strong play is often 20 or below, and an error-strewn game runs well over 60. But it's noisy: one sharp game can spike it and a quiet one can flatter you. Read it as a trend across many games rather than a verdict on one, and always look at the actual blunders instead of chasing the number.
For engine quality they're nearly identical — both run Stockfish and flag the same error types. Lichess wins on price: full analysis is free and unlimited, while Chess.com's complete Game Review needs a paid membership. Chess.com adds a friendlier narrated coach walkthrough some players prefer. Neither, though, reviews across your last hundred games, so a recurring weakness stays invisible in both unless you aggregate your games yourself.
Single-game review won't reveal them — you have to aggregate. Download all your games as PGN from your Lichess profile, run analysis over the batch, and group the errors by theme: hanging pieces, lost endgames, time-trouble collapses, the same opening failing. After 20–30 games a pattern emerges. Chess DNA automates this — it imports your Lichess history, analyzes it, and ranks your recurring weaknesses by the rating they cost, so you train the pattern instead of one-off moves.
