In short: A composed endgame position with a task, usually White to play and win or White to play and draw. The solution should be unique and preferably surprising.
Also called study.
A study is an endgame position that someone has composed. The task is usually "White to play and win" or "White to play and draw," without saying how many moves it may take. The position often looks as if it came from a real game, but the solution is both unique and surprising.
Studies are art, but they also teach endgames. Many of the ideas the studies have shown later turn up in real games.
Réti's study
Richard Réti published this study in 1921. White's king is far from the h-pawn, and Black's king stands close to White's c-pawn. It looks hopeless.
The solution is for the king to head for both targets at once: toward the h-pawn and toward its own c-pawn. Follow it step by step on the page about the rule of the square.
The Saavedra position
One of the most famous studies was published in 1895 in a Scottish newspaper column. The solution, which Fernando Saavedra found, is to promote a pawn to a rook instead of a queen. After a queen, Black can save himself with a stalemate. Read about it on the page about promotion.
A study is a kind of chess problem, but it differs from the usual mate problems: there the number of moves is given, and the position needn't look like a game.
Frequently asked questions
What is the difference between a study and a chess problem?
A problem says how many moves the solution may take, such as mate in two. A study only gives the goal, a win or a draw, and the position can often look like one from a real game.
Which studies are the most famous?
Two of the most famous are on LaviChess: Richard Réti's from 1921, where a king chases two targets at once, and the Saavedra position from 1895, where a pawn becomes a rook instead of a queen.
Updated November 14, 2026