Difference between revisions of "Edwards' Tablebases"
GerdIsenberg (talk | contribs) (Created page with "'''Home * Knowledge * Endgame Tablebases * Edwards' Tablebases''' '''Edwards' Tablebases''', are three-, four and some five-piece tablebases constructed...") |
GerdIsenberg (talk | contribs) |
||
(One intermediate revision by the same user not shown) | |||
Line 1: | Line 1: | ||
'''[[Main Page|Home]] * [[Knowledge]] * [[Endgame Tablebases]] * Edwards' Tablebases''' | '''[[Main Page|Home]] * [[Knowledge]] * [[Endgame Tablebases]] * Edwards' Tablebases''' | ||
− | '''Edwards' Tablebases''', | + | '''Edwards' Tablebases''',<br/> |
are three-, four and some five-piece tablebases constructed by [[Steven Edwards]] in the early 90s, initially for his chess program [[Spector]]. Edwards' Tablebases rely on [[Endgame Tablebases#DTM|Depth to Mate]] and the complete coverage for both sides, using one [[Byte|byte]] per [[Chess Position|position]], with evaluations of the forms "mate in N", "lose (get mated) in N", "draw", and "illegal". Values for the number N (measured in fullmoves, not [[Ply|ply]]) for mates range from mate in 1 upto mate in 126 and for losses in 0 (lose in 0 means [[Checkmate|checkmated]]) to lose in 125 moves. Each file is for a given class (e.g., KBNK) and for a given [[Side to move|side to move]] (e.g., White) <ref>[https://www.stmintz.com/ccc/index.php?id=11493 Steve Edward's Endgame Tablebase generator is now available for WIN95] by [[Michael Byrne|Mike Byrne]], [[CCC]], November 01, 1997</ref>. After finishing the construction procedure in 1994, Steven Edwards made the whole data, documentation and a test program written in [[C|ANSI-C]] publicly available on the Internet. Until the advent of the compressed [[Nalimov Tablebases]], Edwards' Tablebases were quite popular and used in several chess programs, such as [[Crafty]], [[Gromit]], and the commercial [[MChess|MChess Pro]], to name a few. | are three-, four and some five-piece tablebases constructed by [[Steven Edwards]] in the early 90s, initially for his chess program [[Spector]]. Edwards' Tablebases rely on [[Endgame Tablebases#DTM|Depth to Mate]] and the complete coverage for both sides, using one [[Byte|byte]] per [[Chess Position|position]], with evaluations of the forms "mate in N", "lose (get mated) in N", "draw", and "illegal". Values for the number N (measured in fullmoves, not [[Ply|ply]]) for mates range from mate in 1 upto mate in 126 and for losses in 0 (lose in 0 means [[Checkmate|checkmated]]) to lose in 125 moves. Each file is for a given class (e.g., KBNK) and for a given [[Side to move|side to move]] (e.g., White) <ref>[https://www.stmintz.com/ccc/index.php?id=11493 Steve Edward's Endgame Tablebase generator is now available for WIN95] by [[Michael Byrne|Mike Byrne]], [[CCC]], November 01, 1997</ref>. After finishing the construction procedure in 1994, Steven Edwards made the whole data, documentation and a test program written in [[C|ANSI-C]] publicly available on the Internet. Until the advent of the compressed [[Nalimov Tablebases]], Edwards' Tablebases were quite popular and used in several chess programs, such as [[Crafty]], [[Gromit]], and the commercial [[MChess|MChess Pro]], to name a few. | ||
Line 17: | Line 17: | ||
=Publications= | =Publications= | ||
− | * [[Steven Edwards]] | + | * [[Steven Edwards]] and the Editorial Board ('''1995'''). ''An Examination of the Endgame KBNKN''. [[ICGA Journal#18_3|ICCA Journal, Vol. 18, No. 3]] |
* [[Steven Edwards]] ('''1996'''). ''An Examination of the Endgame KBBKN.'' [[ICGA Journal#19_1|ICCA Journal, Vol. 19, No. 1]] | * [[Steven Edwards]] ('''1996'''). ''An Examination of the Endgame KBBKN.'' [[ICGA Journal#19_1|ICCA Journal, Vol. 19, No. 1]] | ||
Latest revision as of 15:10, 30 October 2018
Home * Knowledge * Endgame Tablebases * Edwards' Tablebases
Edwards' Tablebases,
are three-, four and some five-piece tablebases constructed by Steven Edwards in the early 90s, initially for his chess program Spector. Edwards' Tablebases rely on Depth to Mate and the complete coverage for both sides, using one byte per position, with evaluations of the forms "mate in N", "lose (get mated) in N", "draw", and "illegal". Values for the number N (measured in fullmoves, not ply) for mates range from mate in 1 upto mate in 126 and for losses in 0 (lose in 0 means checkmated) to lose in 125 moves. Each file is for a given class (e.g., KBNK) and for a given side to move (e.g., White) [1]. After finishing the construction procedure in 1994, Steven Edwards made the whole data, documentation and a test program written in ANSI-C publicly available on the Internet. Until the advent of the compressed Nalimov Tablebases, Edwards' Tablebases were quite popular and used in several chess programs, such as Crafty, Gromit, and the commercial MChess Pro, to name a few.
Contents
Index Scheme
The index scheme for pawn-less endgames exploits the fourfold symmetry of the chessboard to restrict the last identified piece to the a1-d1-d4 triangle by horizontal, vertical, or diagonal reflections, and features vertical symmetry by confining one pawn to the queen-side flank for endgames with one pawn. Other schemes as applied for instance in Thompson's Databases, enumerating all legal positions of both kings as combined index and considering pawns can't reside on the first or eighth rank, feature denser index ranges than Edwards' [2].
See also
- Bitbases
- Gaviota Tablebases
- Lomonosov Tablebases
- Nalimov Tablebases
- Scorpio Bitbases
- Syzygy Bases
- Thompson's Databases
Publications
- Steven Edwards and the Editorial Board (1995). An Examination of the Endgame KBNKN. ICCA Journal, Vol. 18, No. 3
- Steven Edwards (1996). An Examination of the Endgame KBBKN. ICCA Journal, Vol. 19, No. 1
Forum Posts
1990 ...
- Announcing forced mates: a poll by Steven J. Edwards, rgc, July 21, 1994
- KBBK tablebases: attempt #2 by Steven J. Edwards, rgc, August 13, 1994
- Updated tablebase documentation by Steven J. Edwards, rgc, October 14, 1994
- KPK tablebases now available via ftp by Steven J. Edwards, rgc, October 14, 1994
1995 ...
- Re: FICS vs. ICS: No, really, which is better? by Steven J. Edwards, alt.chess.ics, June 01, 1995
- MCP6 Endgame Tablebase. Thanks mr.Hirsch! by H.Pieters, rgcc, November 11, 1996
- Steve Edward's Endgame Tablebase generator is now available for WIN95 by Mike Byrne, CCC, November 01, 1997
- EGTB by Frank Phillips, CCC, February 10, 1998
2000 ...
- Tablebases Edwards by RR, Winboard Forum, July 25, 2001
2010 ...
- A different tablebase encoding format by Steven Edwards, CCC, August 10, 2014
Downloads
References
- ↑ Steve Edward's Endgame Tablebase generator is now available for WIN95 by Mike Byrne, CCC, November 01, 1997
- ↑ Ernst A. Heinz (1999). Endgame Databases and Efficient Index Schemes for Chess. ICCA Journal, Vol. 22, No. 1, ps
- ↑ Courtesy Steven Edwards