Francesca
团队成员
人 | 职位 |
---|---|
Tom King | engine programmer |
参赛
赛 | 参赛者 | 得分 | 对局 |
---|---|---|---|
国际象棋, 2000, 伦敦 | 14 | 3.5 | 9 |
国际象棋, 1999, Paderborn | 30 | 4.0 | 7 |
国际象棋, 1997, 巴黎 | 34 | 3.0 | 11 |
国际象棋, 1996, Jakarta | 27 | 6.0 | 11 |
国际象棋, 1995, Paderborn | 34 | 3.0 | 11 |
介绍
- 介绍来源于 1999:
Francesca is an amateur program. It follows the trends of most chess programs since the 1970's and uses many of the state-of-the-art techniques: aspiration alpha-beta, null moves, other forward pruning mechanisms, a large transposition table, an evaluation hash table, large opening book etc.
Prior to 1998, Francesca relied mainly on piece-square values and a fast search to play good chess. In 1998, the evalution code was completely rewritten, and all evaluation is now done in the leaves of the search tree. This has produced stronger play; although 1999 Francesca is slower in nodes per second than 1997 Francesca, she plays a more interesting and challenging game.
- 介绍来源于 1997:
Francesca follows the trend of programs since the late 1970s. It uses iterative deepening alpha-beta search to determine which move to make.
The search is selective rather than brute-force. Null moves, and a home- grown forward pruning mechanism are used to help accelerate the search. In addition, a hash table is used to aid move ordering, and save searching of subtrees that have already been visited in the search.
The evaluation function is simple, but it knows something about pawn structure, king safety, and piece placement. There is always a tradeoff between knowledge and speed, and Francesca falls into the category of a "fast, dumb" searcher, rather than a "slow, smart" searcher.