Chess Fluency
Spaced repetition for chess
See the right move in two seconds, not two minutes. Chess Fluency turns your own mistakes, your openings and the patterns strong players see instantly into positions you play out on a real board, each one scheduled to come back just before you would forget it.
Free and open source. Works on your phone. Connects to your Lichess account.
Early version I train with it and refine it as I go.
Why
As a polyglot and a passionate language learner, I know vocabulary doesn't stick from one pass through a textbook. It sticks through spaced repetition: seeing the right thing again at the right moment, until recalling it takes no effort at all.
In chess I could solve hard puzzles if I was patient enough, but in bullet and blitz I didn't see the same ideas fast enough. What was missing wasn't calculation. It was recognition. I wondered for a long time when someone would combine spaced repetition with a real chessboard and your own games. Then it occurred to me that I could make it.
It learns from your games
Connect Lichess and Chess Fluency reads your games: every error, the positions where you slip again and again, and the patterns behind them. Then it turns them into cards.
The position, the move you played, the move you missed, tagged with what went wrong. The most instructive come first: errors that changed the result, not slips in a game you were winning anyway. In correspondence games, every inaccuracy counts.
Positions you've misplayed in several games, usually an opening line you meet again and again, come first. Fix one and it pays off every time it comes up.
The moment after the opening when everyone is developed and the plan isn't obvious. Taken from your games, especially where you then went wrong, followed by master games that reached a similar position.
Your blind spots by type of position, the openings where you go wrong most, and the habits behind quiet errors: long thinks that end badly, minor-piece trades that let the tension go, locked centres. Each one has a drill.
How you train
Build a repertoire from the Lichess opening explorer or import any Lichess study. Watch a new line once, then play it from memory. At the end of any line, see what's played next.
Every move is timed against your own history on that card. Faster than usual pushes it further out; slower brings it back sooner. Pause any time.
After each card, replay why your move fails and why the best move works, or open the analysis board with Stockfish's choices, an evaluation bar and the opening explorer.
Seventeen classic opening traps, drilled as: dodge it, punish it, and set it. The ones you've fallen for in your own games come first. Every line is checked by Stockfish.
Your missed Lichess puzzles come back on a schedule, once enough time has passed that you have to see the idea again. Easiest or hardest first, to match your focus.
In a puzzle you know there's a shot. In a game you don't. Decide first, then play: a real puzzle or a quiet position, on a coin flip every time.
Stockfish runs right in your browser and accepts any move that is just as good, so a different winning idea still counts. Your progress syncs between your phone and computer through your Lichess login, and decks can be exported to Anki or MintDeck.
Why spaced repetition
You forget on a curve. Right after you learn something, memory fades fast, then more slowly. Psychologists have measured it since the 1880s.
Recalling beats rereading. Pulling an answer out of memory strengthens it far more than looking at it again. That's why every card asks you to play the move, not just look at it.
Timing is everything. A review just before you would forget does the most good, and each successful one lets the next wait longer: a day, a few days, a week, a month. Your time goes to what you're about to lose, not to what you already know.
I learned this the hard way. I had already learned French, German and Spanish, and found I had both a passion and a knack for languages. Then I moved to Beijing and started on Chinese, and for the first time I wasn't progressing as fast as I'd hoped. I felt I needed to revolutionize how I studied. After a lot of research I found Anki and spaced repetition.
Being consistent with it made a huge difference: my Chinese grew much faster than before. Since then I've believed that timing your reviews well is an essential part of learning anything new efficiently. Chess Fluency is that idea, on a chessboard.
Every position is a card. After you play it, you rate it Again, Hard, Good or Easy, and the app suggests the rating from whether you found the move and how your speed compares with your own past times. The FSRS algorithm, the modern scheduler in Anki, then works out when your recall would drop to about 90% and brings the card back right then.
Open source
Chess Fluency stands on the work of the open source chess world: the Lichess board, rules library, puzzle database and opening explorer, the Stockfish engine, and the FSRS scheduler. It is free, with no ads and no account of its own, and its code is open under the GPL.
Your data. Your cards and progress live in your browser. To sync between devices they are also kept on this site's server, filed under your Lichess username. Your Lichess login itself is never stored there, only a one-way fingerprint of it to recognize you. Nothing is sold or shared.
Support
If Chess Fluency helps your chess, you can chip in toward the time it takes to build and improve.
Or
Listen at samuelwexler.com/music
Or on your platform: