Applying Proof-number Search to Chess 0s0z0z0z Z0z0z0ok 0z0z0o0z Z0z0o0z0 0a0o0z0m Z0z0z0z0 Pzpzpo0z Zkz0zbz0

Proof-number search (pn-search) has shown its merit in contributing to the solution of Connect-Four, Qubic and Go-Moku. In this contribution we show that pn-search is a highly capable searcher for mates in chess. Pn-search achieves its results without using any heuristic chess knowledge. We present the results of comparing the performances of pn-search and a number of other mate searchers on a large number of problems. Finally, we discuss how pn-search can be applied to general tactical searches. 1 Background A major attraction of chess is the number of diierent skills necessary to play all stages of the game at a high level. To play the opening well, one needs to select an opening repertoire, study the lines in it, understand the positions resulting from it, and constantly keep up to date with enhancements introduced by top-level players. In the middle game, a deep understanding of the strategical features of positions, as well as recognition and thorough calculation of the combinatorial opportunities are necessary. The end-game again has its own rules. Some end-games are fully analyzed and only require some experience. For others deep analysis, either performed at home or during a game, constantly unveils chess secrets to the chess-playing community. In short, diierent parts of the game require diierent skills. This is partly reeected in the way chess programs are composed. To play the opening, large opening books are installed, and constantly updated with the newest inventions of grandmaster play. For the end-game, databases,