In short
To solve a chess puzzle, look first at every check, capture and threat for the side to move, list the candidate moves they give, and calculate each forcing line to the end, including your opponent's best reply. Choose the move only when you can see why it works against every defence. Solving carefully, by theme and with repetition, builds the pattern recognition that wins real games.
A chess puzzle is a position from a real or composed game where one side has a strong move, usually a tactic that wins material or gives mate. The puzzle tells you that something is there. Your job is to find it and, just as importantly, to be sure it works before you move.
That last part is where most solvers go wrong. It is easy to spot a tempting move and play it on instinct. Sometimes it is right, and you learn nothing. Sometimes it is wrong, and you learn even less, because you never worked out why. A method turns puzzles into practice for the thinking you need in your own games.
Step 1: look at checks, captures and threats
Start with the most forcing moves, in this order: checks, captures, then threats. Forcing moves limit your opponent's replies, which makes them easier to calculate and more likely to be the solution. List every one, even the ones that look silly. Many puzzle solutions begin with a move that gives material away, and you will only find them if you look at every check and capture rather than just the safe ones.
- Checks: every move that gives check, including sacrifices.
- Captures: every capture, even of a defended piece.
- Threats: moves that attack something valuable or threaten mate next move.
While you scan, note what looks weak in your opponent's position: an undefended piece, a king with no escape squares, two pieces a knight's move apart, a piece pinned to the king. The weakness usually tells you which forcing move to look at first.
Step 2: list your candidate moves
A candidate move is one worth calculating. From your scan, pick the two or three that create the most problems for your opponent. Writing them down, or saying them to yourself, stops you from fixing on the first idea and missing a better one. If none of the forcing moves works, then consider quiet moves that set up a threat, but only after the forcing ones are ruled out.
Step 3: calculate forcing lines to the end
For each candidate, play the line out in your head. After your move, find your opponent's best reply, not the reply you hope for. Then your next move, and so on, until the position is clearly won: mate, a piece up with no counterplay, or a pawn about to promote. If a line stops at a position you cannot judge, it is not finished.
Three chess puzzles to try
Each puzzle below gives a question, then the answer after it. Cover the answer, use the three steps, and decide on your whole line before you read on.
Answer to puzzle 1: Re8#. Scan the checks first: the rook on e1 can give check on e8. Black's king is boxed in by its own pawns on f7, g7 and h7, nothing can capture on e8, and the black rook on c7 cannot reach f8 to block. This is the back-rank mate, the most common mating pattern in real games.
Answer to puzzle 2: Nc7+. It is a check, so it goes on the list first, and it also attacks the rook on a8: a knight fork. No black piece can capture the knight on c7, and Black has to move the king to d7, d8 or e7. Then Nxa8 wins the rook. The weakness to notice was the king and rook standing a knight's move from the same square.
Answer to puzzle 3: 1.Qg8+ Rxg8 2.Nf7#. The queen check on g8 looks like it simply loses the queen, which is why it is easy to skip. But it is a check, so it belongs on the list. The king cannot take on g8 because the knight on h6 guards it, so the rook has to capture. Now the king is surrounded by its own rook and pawns, and the knight gives mate from f7. This is a smothered mate.
Learn the tactical themes
Most puzzles are built from a small number of recurring ideas. Once you know them by name, you recognise them faster, and when a review shows you keep missing one, you can practise it on its own. The puzzle themes page sorts puzzles by idea.
| Theme | The idea |
|---|---|
| Fork | One piece attacks two at once. |
| Pin | A piece cannot move without exposing a more valuable one behind it. |
| Skewer | A valuable piece is attacked and moves, leaving the one behind it. |
| Discovered attack | Moving one piece opens an attack by another. |
| Deflection | Lure a defender away from what it guards. |
| Remove the defender | Take the piece that holds the position together. |
| Back-rank mate | Mate a king trapped behind its own pawns. |
| Smothered mate | A knight mates a king hemmed in by its own pieces. |
Speed or accuracy: which matters more?
Both, but at different times. Rated puzzles are for accuracy: there is no clock pressure on the rating, so take the time to calculate fully. A wrong move counts as a miss even if you then find the right one, and so does asking for a hint, which is a good reason to check before you move.
Puzzle Rush is for speed. You solve as many as you can in three or five minutes, or in Survival with no clock, starting easy and getting harder; a wrong move ends that puzzle and the third miss ends the run. It trains fast recognition of simple patterns, which is valuable in blitz and in time trouble. It is not rated, and it should not replace careful solving: speed built on guessing does not carry over to your games.
What is a puzzle rating?
Each puzzle has a rating of its own, worked out from how often players of different strengths solve it. When you attempt one, the trainer treats it as a game between you and the puzzle: solving counts as a win and missing as a loss, and your rating moves with Glicko-2, the system Lichess uses. A new player starts at 1500 with a question mark next to the rating until enough attempts have made it reliable.
A puzzle rating measures how well you solve puzzles, which is related to your playing strength but is not the same. Puzzles tell you that a tactic exists; in a game nobody does. That is why puzzle ratings and game ratings often differ, and why both are worth having.
How do you get better at chess puzzles?
Three things work: solving regularly, solving carefully and repeating what you have solved. Regular practice keeps patterns fresh. Careful solving trains the calculation habit. Repetition turns a pattern you once worked out slowly into one you see instantly.
Repetition is the idea behind Cycles, in the style popularised as the Woodpecker Method: one fixed set of puzzles, solved again and again in the same order, faster each time. The first cycle is slow and careful; later cycles are about speed and accuracy on positions you have met before. Seven cycles is the usual aim, with a rest day between them, and the set should be puzzles you can mostly solve, since the gain comes from repeating rather than struggling.
- Solve rated puzzles most days, slowly enough to be sure of each line.
- Spend a week at a time on a weak theme, chosen from your own reviews.
- Run a Cycles set alongside to make the common patterns automatic.
- Use Puzzle Rush now and then for speed, not as your main training.
- Train positions from your own games in Key moments, where the pattern is one you actually missed.
Common mistakes when solving puzzles
| Mistake | What to do instead |
|---|---|
| Playing the first move that looks good | List every check, capture and threat first |
| Ignoring the opponent's best reply | Ask what checks and captures the defender has at each step |
| Stopping the line too early | Calculate until the result is clear |
| Skipping sacrifices because they lose material | Treat every check and capture as a candidate |
| Only doing timed puzzles | Make careful rated solving the core |
| Never looking back at missed puzzles | Replay the ones you missed and find what you overlooked |