What changes once most of the pieces are gone, and how to actually convert.
The endgame rewards precision and technique more than flashy calculation. With fewer pieces left on the board, small advantages — an extra pawn, a slightly better king position, a passed pawn — become much easier to turn into a full win than they would be earlier in the game, but only if you know the right technique. Missing it can just as easily turn a winning endgame into a draw, or a draw into a loss.
The single biggest shift moving into the endgame: your king stops being a piece you protect and becomes one of your strongest pieces. With most of the material that could attack it gone, there's usually little risk in bringing it toward the center and into the action. A centralized king can support its own pawns, block the opponent's pawns, and directly contest key squares — in many endgames, the side with the more active king simply wins.
Opposition is the key technique in king-and-pawn endgames: when the two kings face each other with one square between them (on the same rank, file, or diagonal), the player who does not have to move holds the opposition — and that's usually the side with the advantage, since the other king is forced to give ground. Distant opposition (kings three or five squares apart on the same line) follows the same logic, just from farther away. Recognizing and fighting for the opposition is often the difference between winning and merely drawing a king-and-pawn ending.
Zugzwang describes a position where a player would prefer to simply pass, but chess doesn't allow that — any legal move they make weakens their position. Endgames are full of zugzwang because there are so few pieces left to create waiting moves with. Triangulation is a technique for exploiting this: maneuvering your own king in a small triangle to deliberately "lose a tempo," handing the move back to your opponent at the moment it hurts them most.
A passed pawn — one with no opposing pawns able to block or capture it on its way to promotion — is one of the most powerful assets in an endgame, because your opponent has to permanently commit a piece just to stop it from queening. A fast way to judge whether a lone king can catch a passed pawn is the "rule of the square": draw an imaginary square from the pawn to its promotion square: if the defending king can step inside that square, it can catch the pawn; if not, the pawn queens on its own.
A single weakness in your opponent's position — one weak pawn, one weak square — can often be defended by a single well-placed piece. Two separate weaknesses on opposite sides of the board usually can't both be defended at once, since defending one pulls pieces away from the other. This is one of the most reliable ways to convert a small, static advantage into a real, winning one.
King and queen versus lone king, and king and rook versus lone king, come up constantly, and both follow the same core idea: restrict the enemy king to an edge or corner of the board using your queen or rook, then bring your own king up to support the final mate. These patterns are worth being able to execute automatically, since reaching a winning endgame only matters if you can actually convert it.
If you're up material, trading pieces (not pawns) generally works in your favor — fewer pieces on the board means fewer tricks and swindles available to whoever's behind, and it moves the game closer to a straightforward, technical win. If you're the one behind, the opposite logic applies: keeping more pieces on the board keeps more complications and chances alive.
Once a game reaches an endgame you're unsure how to convert, run it through Analyze to see exactly where the technique held up or broke down.