ChessGrader / Guides / Lichess vs Chess.com analysis

Lichess vs Chess.com analysis: which should you use?

Both run Stockfish on your games. One gives you unlimited analysis with spartan labels; the other gives you a polished report once a day. Here's the straight comparison, verdicts included.

By the ChessGrader team · Updated July 10, 2026

The short answer

If you analyze more than one game a day and don't want to pay: use Lichess. Its server-side Stockfish analysis is free, unlimited, and deeper than what most sites run. If you want the friendliest single report — move-by-move labels like Brilliant and Miss, a coach explaining your mistakes — Chess.com's Game Review is the better product, but the free tier gives you one per day and holds the coach commentary behind a paid membership. Serious volume analysis: Lichess. One nice review of today's game: Chess.com. Unlimited Chess.com-style grading without the cap: that's the third option, covered below.

What each actually gives you free

Lichess: everything, forever

Lichess gives every user unlimited server-side Stockfish analysis — roughly 2 million nodes per move, which is deeper than most sites bother with. You get an eval graph, average centipawn loss (ACPL), an accuracy percentage, and every serious error flagged as an Inaccuracy, Mistake, or Blunder. On top of that: an opening explorer built on a database of billions of games, a masters database, and Studies — a genuinely excellent free tool for annotating and sharing analysis.

What you don't get: the fun labels. No Brilliant, no Great, no Miss, no estimated game rating. Lichess tells you where you lost the thread; it doesn't celebrate the moments you found it.

Chess.com: one good review a day

Chess.com's free tier includes one Game Review per day. That review is genuinely richer than Lichess's output: the full label vocabulary (Brilliant down to Blunder), an accuracy score, an estimated game rating, and a coach persona that walks through key moments. But the full coach commentary and blunder/mistake details in self-analysis require a paid membership, and once your daily review is spent, you're waiting until tomorrow. The free self-analysis board exists but is limited compared to what Lichess hands out.

Side by side

Where Lichess wins

• Volume. Play five blitz games, analyze five blitz games. No meter.
• Depth. ~2 million nodes per move of server Stockfish is a deeper search than most commercial reviews run.
• Openness. The accuracy formula is documented, the code is open source, and the game database is downloadable. When Lichess gives you a number, you can check how it was made.
• Study tools. The opening explorer and Studies are the best free preparation tools anywhere, full stop.

Where Chess.com wins

• The report itself. Game Review is the most polished post-game experience in chess. The labels are motivating, the layout is clear, and the coach makes engine output feel human.
• The label vocabulary. Lichess only tells you what went wrong. Chess.com's Great and Brilliant labels also tell you what went right, which matters more for morale than analysts admit.
• Where your games are. If you play on Chess.com, reviewing there is one click. Convenience is a feature.

The third option

The awkward truth of this comparison: most people want Chess.com's report at Lichess's price. That gap is why ChessGrader exists. ChessGrader's free game review gives you the full label vocabulary — Brilliant, Great, Miss, and the rest — plus accuracy and an estimated game rating, unlimited, with no signup. It imports games from both Chess.com and Lichess by username, or from any pasted PGN. The engine (Stockfish 17.1) runs in your browser at a fixed 100,000 nodes per move, with critical moments verified at 600,000 nodes, and every formula is published on the methodology page.

The honest trade-offs: analysis speed depends on your device, there's no opening explorer or game database, and it's standard chess only. ChessGrader replaces the daily-capped review, not Lichess's study tools.

Why the accuracy numbers don't match

Run the same game through both sites and you'll get two different accuracy percentages. That's not a bug. Lichess publishes its method: centipawns become a win probability via a logistic curve, each move's probability drop becomes a per-move accuracy via an exponential-decay formula, and the game score blends a volatility-weighted average with a harmonic mean. Chess.com's formula is proprietary — nobody outside the company can reproduce it, and in practice it tends to read a few points more generous.

So never compare your Lichess 84% to a friend's Chess.com 87%. Compare numbers only within one site, against benchmarks for that site's formula. Same engine family, different math on top.