How to stop blundering in chess
Nearly every article about blunders tells you to slow down and check for hanging pieces. Here is what the losing move actually looks like in 10,440 positions, and why that advice misses most of them.
- 66% of losing moves are not captures
- 12% of punishments hit the square that just changed
- 40% of punishments are a capture
- counted from 10,440 positions
What a blunder actually looks like
Every position in this trainer begins with one move by your opponent -- the move that hands the game over. There are 10,440 of them, so it is worth asking what they have in common.
They are mostly not captures. 3,571 of the 10,440 losing moves take something, which is 34%. The other 66% are ordinary developing moves, rook lifts, pawn pushes and king moves that simply left something behind.
And the punishment is rarely aimed at the piece that moved. The winning reply lands on the square the opponent just moved to in 1,272 positions -- 12%, about one in 8. So the reflex of glaring at the piece that just moved catches a small minority of what is available.
Nor is the answer usually a capture: 4,189 of 10,440 winning replies take something, 40%. Which means a blunder check built entirely out of "what did they take and what can I take back" is looking in the wrong place most of the time.
What this is and is not. These are counts over the 10,440 positions in this trainer, all from the Lichess puzzle database. A puzzle exists because somebody made a mistake worth punishing, so this describes the shape of punishable mistakes rather than every mistake ever made. It cannot tell you how often you personally hang a piece, and no page that has not seen your games can.
A blunder check that matches the data
Four questions, in this order, before every move. It takes under ten seconds once it is a habit, and it is built around what the counts above actually say rather than around the captures that are easy to remember to check.
- What changed? Not what did they threaten -- what changed. Which squares does that piece see now that it did not see before, and which squares did it stop covering when it left.
- What of mine is undefended right now? Count it. Not the pieces you are worried about, all of them. This is the one step the standard advice gets right and it is worth keeping.
- If I play my move, what is undefended then? Different question, and the one that catches the majority of blunders, because the piece you are about to move is usually defending something.
- Do they have a check, a capture or a threat I have not answered? Checks first. They are forcing and there are rarely more than three.
The pattern that does the punishing
If you want to know what a blunder costs you, look at what collects. The fork is tagged on 1,445 of the 10,440 positions and appears in 82 of the 87 skills. In the rating-drawn pools it is the most common motif in every band from 800 to 1800, and tied for the most below 800. Only above 1800 does anything overtake it, and what overtakes it is the defensive move.
The explicit hanging-piece tag is rarer -- 362 positions -- but it turns up in 75 of the 87 skills, including the master pool at the top of the ladder. Leaving a piece is not a habit you grow out of at some rating. It gets rarer and more expensive.
- Free piecesThe motif page: how a hanging piece announces itself
- ForksWhat usually collects the material
- Free pieces worksheetsPrintable, with an answer key
Where blunders come from at your rating
What you are most likely to hang, and what is most likely to punish it, changes as you climb. The band pages have the counts.
- At 600 Elo26 of 45 answers are mate
- At 1000 Elo10 of 44 are pins or deflections
- At 1800 Elo33 of 149 answers are quiet or defensive
Why do I blunder when I know better?
Because knowing a rule and running it are different actions. The counts here suggest a specific gap rather than a general one: most losing moves are quiet, so a check that only fires when something is captured never runs at all. A fixed order of questions before every move fires whether or not anything looks dangerous.
Does playing faster make me blunder more?
This page cannot measure that, and it is not going to guess. What it can say is that the check above is short by design, because a routine that costs thirty seconds is a routine you will stop running in a blitz game.
What is the difference between a blunder and a mistake?
Convention rather than anything precise: a blunder loses material or the game outright, a mistake gives away an advantage. The engine label is not the useful part. What is useful is that both usually arrive on a quiet move.
Do strong players still hang pieces?
The tag appears in 75 of the 87 skills here, including the master pool at the top of the ladder. It gets rarer and it never stops.