Free · no account · every number carries its sample size
Chess stats analyzer for Chess.com & Lichess
Enter any public username. The scan reads up to 1,000 recent rated games through the official APIs and turns them into one ranked verdict: which openings leak rating, how your losses actually happen, what tilt costs you — and the single most valuable thing to fix this week. Every stat shows its sample size.
What the report contains
- Leak ledgerEvery fixable leak ranked by estimated Elo cost — opening, colour, clock, conversion — each with sample size and severity.
- Repertoire mapEvery opening you play enough to judge, plotted volume × performance against your Elo-based expected score.
- Loss anatomyBlunder, drift, clock or missed win — how your losses actually happen, classified with computer evals where they exist.
- Tilt & sessionWin rate after one loss versus two straight, and by game number within a session. Ends in a rule, not a chart.
- Healthy zonesTime trouble, flag losses and draw rate benchmarked against what is typical for your rating band.
- Chess characterFour measured style axes resolving to one of nine archetypes — grounded in your own numbers, built to share.
What it does not do
No engine verdicts yet — the current report is descriptive statistics measured against Elo expectation in your own games, and it says so on every panel. Percentages below the evidence gate are counted, not judged. If a pattern doesn’t clear its sample-size gate, the report tells you that instead of inventing a story. The methodology is public.
Frequently asked
Is the chess stats analyzer really free?
Yes. The scan is free, needs no account and no sign-up. Enter any public Chess.com or Lichess username and the report renders in your browser.
Does it work for both Chess.com and Lichess?
Yes. It reads up to 1,000 recent rated games through the official public Chess.com and Lichess APIs — nothing is scraped and no password is needed, because the game history of every account is public.
What stats does it actually compute?
Openings by colour scored against your Elo-based expected score, a ranked leak ledger priced in estimated rating points, loss anatomy (blunder / drift / clock / missed win), tilt after losses, win rate by game number within a session, and time-trouble share. No engine evaluations yet — the current report is descriptive statistics only, and says so.
Why does every number show an n= sample size?
Because a win rate over 6 games is noise. Every percentage in the report ships with its sample size, and patterns below the evidence gate (n ≥ 10) are counted but not judged. That statistical honesty is the point of the tool.
Why is my chess rating stuck?
Usually not because of one blunder — repeatable leaks cost more: an opening you keep playing that scores below what your rating predicts, losses on the clock, or playing on tilt after consecutive losses. The scan measures each of these in your own games and ranks them by estimated Elo cost.
Can I analyze someone else's games?
Any public account can be scanned — coaches scan students, and you can scan a titled player to see how the report reads. Only public game data is used, and scan pages are not stored on the account you scan.
Run the scan on your own games.
Free for any public Chess.com or Lichess account. No sign-up required.