Most games below master level are not won. They are lost, usually by one or two moves that hand over material or the king. The good news is that blunders are not random. They come in a small number of recurring shapes, and once you can name a shape you start seeing it before you play into it.
This article walks through the patterns that show up again and again in engine reviews, how to recognize each one at the board, and a short routine that catches most of them before the move is made.
Loose pieces drop off
The oldest tactical slogan in the book is LPDO: loose pieces drop off. A loose piece is one that is not defended by anything. On its own it is not a problem. It becomes a problem the moment the opponent can attack it and something else at the same time, because you can only move one piece per turn.
FEN
r2r2k1/pp3ppp/2qp4/3Np3/4P3/3P4/PPPQ1PPP/R4RK1 w - - 0 1Black's queen on c6 is not attacked, and Black has no immediate worry. But it is loose, and it shares a knight's distance with the king on g8 through the square e7. 1.Ne7+ is check, so Black has no time to save the queen.
How to recognize it: before you move, glance at every one of your pieces that has no defender. Count them. Each loose piece is a possible fork target, and two loose pieces on the same line are a possible skewer.
Training method: play through your lost games and, at each capture you missed, ask whether the piece was loose. The Blunder Checker marks these for you and shows the engine's refutation, so the pattern stops being abstract.
The back rank nobody looked at
Castled kings with three untouched pawns in front of them are safe from almost everything except a rook or queen arriving on the back rank. The blunder is nearly always the same: a defender that was covering the first rank wanders off to grab something.
FEN
6k1/pp3ppp/1q6/8/8/2r5/PP1Q1PPP/4R1K1 w - - 0 1Black won a pawn and lost the game in the same move. The rook that stood on c8 was the only piece guarding the back rank, and nothing replaced it.
How to recognize it: if your king has no flight square (every pawn in front of it is still on its starting square), treat the back rank as a piece that needs defending. Before any capture, check which of your pieces is leaving the first or eighth rank.
Training method: give your king an escape square early in the middlegame when nothing better is going on. One tempo spent on ...h6 or h3 has saved countless games. Back-rank mates also make up a large share of the checkmate puzzles on this site, which is a fast way to drill the pattern.
Forks, and the two-targets rule
A fork is just one piece attacking two things at once. Knights are the most famous forkers because their attack is hard to see, but queens and pawns do it just as often. The queen fork is usually a check combined with an attack on a loose piece on the same rank, file, or diagonal.
The practical rule is simple: a fork needs two targets. If you keep the number of loose pieces at zero and your king out of check-able lines, there is nothing for a fork to land on. When you see a knight within two moves of both your king and a piece, the alarm should go off.
Training method: when the opponent has a knight, before each move name the squares it could reach in two jumps. It feels slow for a week and then becomes automatic.
Overloaded defenders
Sometimes every piece is defended, and the position still falls apart because one piece is doing two jobs.
FEN
2rq2k1/1p3ppp/1b1p1n2/p3p3/4P3/3PBN2/PP1Q1PPP/2R3K1 w - - 0 1The queen on d8 protects the rook on c8 and, through the empty c7 square, the bishop on b6. White exchanges rooks to drag the queen away from one of the two jobs, then collects the bishop.
How to recognize it: when you count defenders, ask whether the same piece appears in two counts. If so, one of the two pieces is effectively loose.
Training method: after each game, look at the moves where your evaluation dropped and check whether a defender was pulled away by an exchange. The Game Review shows the drop in win probability move by move, so these moments are easy to spot.
Pins that quietly become real
A pin is harmless until something changes behind it. The classic case is a bishop on g5 pinning a knight on f6 to the queen on d8. For ten moves nothing happens, then you play ...Qc7, and now Bxf6 wins a piece because the knight was being held together by the queen behind it.
How to recognize it: when a piece of yours is pinned, write down (mentally) what is being pinned to what. Any move by the piece behind the pin needs a second look.
Discovered attacks
A discovered attack happens when you move one piece and reveal an attack from another one behind it. The danger comes from the opponent's side: a piece standing on the same line as your queen or king, with just one enemy piece between them, is a discovered attack waiting to happen. Those positions deserve special attention because the moving piece can go almost anywhere, often with a capture or a check.
How to recognize it: look for lines from your king and queen through a single enemy piece to an enemy rook, bishop, or queen. If one exists, assume the discovered attack is coming and move out of the line first.
Moving the piece that was doing a job
This is the most human blunder of all. A knight on d4 is blocking a file, defending e2, and guarding against a check. You see a nice square for it on f5 and go. Three things that were held together come apart.
How to recognize it: before moving any piece, ask what that piece is currently doing. If the answer is "nothing," move it freely. If the answer is "defending something," check whether the something still holds.
Automatic recaptures and intermediate moves
The habit of recapturing instantly costs a lot of games. When the opponent captures, the first question is not "how do I take back" but "is there something better." Often there is a check or a bigger capture first. The opponent's capture may also have been a bait: the piece you recapture with lands on a square where it gets forked.
Training method: slow down specifically after a capture. Make it a rule that a recapture is never a pre-move in anything slower than bullet.
Time trouble and the blitz tax
In fast games the blunder rate rises sharply in the last minute, and premoves and mouse slips add their own category. Nothing clever helps here except playing a little faster earlier so that the final moves are not played on instinct. If your review shows that most of your blunders come after move 30 in blitz, that is a time management problem, not a tactics problem.
The pre-move routine
Here is a routine that takes a few seconds once it is habit and catches most of the patterns above.
- What did the last move change? Which lines did it open, which squares did it stop defending, what does it attack now.
- Checks, captures, threats for the opponent. Before every move, list the opponent's checks and captures after your intended move. Not all of them, just the forcing ones.
- What is my piece doing right now? If it is defending something, confirm that the something is still covered after the move.
- Count loose pieces. If you have two or more, consider a move that fixes one of them.
Find your own pattern
Everyone has a signature blunder. Some players leave knights loose, others forget about diagonals leading to their king, others collapse in time trouble. The fastest way to find yours is to review a batch of lost games and group the decisive mistakes by type.
Chessblood's Blunder Checker does this grouping for you: import your games, let Stockfish run in your browser, and look at the moves that cost 20 percentage points of win probability or more. If three of your last ten games were lost to the same shape, you know exactly what to train. For a fuller method, see how to analyze your chess games.