Difference between revisions of "Move Generation"

From Chessprogramming wiki
Jump to: navigation, search
(6 intermediate revisions by the same user not shown)
Line 77: Line 77:
 
==1980 ...==
 
==1980 ...==
 
* [[Joe Condon]], [[Ken Thompson]] ('''1982'''). ''Belle Chess Hardware'', [[Advances in Computer Chess 3]], Reprinted ('''1988''') in [[Computer Chess Compendium]]
 
* [[Joe Condon]], [[Ken Thompson]] ('''1982'''). ''Belle Chess Hardware'', [[Advances in Computer Chess 3]], Reprinted ('''1988''') in [[Computer Chess Compendium]]
* <span id="Waterloo1982"></span>[[Greg Bakker]], [[Jim Jonkman]], [[Jonathan Schaeffer]], [[Tom Schultz]] ('''1982'''). ''VLSI Implementation of a Chess Legal Move Generator''. EE755S-1, [[University of Waterloo]] <ref>[https://uwaterloo.ca/water-under-the-bridge/1983 1983 | Waking up to change] in [https://uwaterloo.ca/water-under-the-bridge/about-authors Chris Redmond and Simon the Troll] ('''1998'''). ''[https://uwaterloo.ca/water-under-the-bridge/ Water Under the Bridge]''. [[University of Waterloo]] » [[Move Generation#Waterloo1982|VLSI Move Generation]]</ref>
+
* <span id="Waterloo1982"></span>[[Greg Bakker]], [[Jim Jonkman]], [[Jonathan Schaeffer]], [[Tom Schultz]] ('''1982'''). ''VLSI Implementation of a Chess Legal Move Generator''. EE755S-1, [[University of Waterloo]] <ref>[https://uwaterloo.ca/water-under-the-bridge/1983 1983 | Waking up to change] in [https://uwaterloo.ca/water-under-the-bridge/about-authors Chris Redmond and Simon the Troll] ('''1998'''). ''[https://uwaterloo.ca/water-under-the-bridge/ Water Under the Bridge]''. [[University of Waterloo]] » [[#Waterloo1982|VLSI Move Generation]]</ref>
* [[Jonathan Schaeffer]], [[Patrick A.D. Powell]], [[Jim Jonkman]] ('''1983'''). ''[https://link.springer.com/chapter/10.1007/978-3-642-95432-0_19 A VLSI legal move generator for the game of chess]''. in [https://en.wikipedia.org/wiki/Randal_Bryant Randal Bryant] (eds.) [https://link.springer.com/book/10.1007%2F978-3-642-95432-0 Third Caltech Conference on Very Large Scale Integration]  
+
* [[Jonathan Schaeffer]], [[Patrick A.D. Powell]], [[Jim Jonkman]] ('''1983'''). ''[https://link.springer.com/chapter/10.1007/978-3-642-95432-0_19 A VLSI legal move generator for the game of chess]''. in [[Mathematician#REBryant|Randal E. Bryant]] (eds.) [https://link.springer.com/book/10.1007%2F978-3-642-95432-0 Third Caltech Conference on Very Large Scale Integration]  
 
* [[Carl Ebeling]], [[Andrew James Palay]] ('''1984'''). ''The Design and Implementation of a VLSI Chess Move Generator''. Proceedings of the 11th Annual International Symposium on Computer Architecture. [[IEEE]] and [[ACM]].
 
* [[Carl Ebeling]], [[Andrew James Palay]] ('''1984'''). ''The Design and Implementation of a VLSI Chess Move Generator''. Proceedings of the 11th Annual International Symposium on Computer Architecture. [[IEEE]] and [[ACM]].
 
* [[Stuart Cracraft]] ('''1984'''). ''Bitmap move generation in Chess''. [[ICGA Journal#7_3|ICCA Journal, Vol. 7, No. 3]]
 
* [[Stuart Cracraft]] ('''1984'''). ''Bitmap move generation in Chess''. [[ICGA Journal#7_3|ICCA Journal, Vol. 7, No. 3]]
Line 126: Line 126:
 
* [http://www.open-aurec.com/wbforum/viewtopic.php?f=4&t=331 Move Generation Speed] by [[Dan Honeycutt]], [[Computer Chess Forums|Winboard Forum]], October 21, 2004
 
* [http://www.open-aurec.com/wbforum/viewtopic.php?f=4&t=331 Move Generation Speed] by [[Dan Honeycutt]], [[Computer Chess Forums|Winboard Forum]], October 21, 2004
 
==2005 ...==
 
==2005 ...==
 +
* [http://www.open-aurec.com/wbforum/viewtopic.php?f=4&t=5623 Yet another new bitboard move generation method] by [[Zach Wegner]], [[Computer Chess Forums|Winboard Forum]], September 22, 2006 » [[Titboards]]
 +
: [http://www.open-aurec.com/wbforum/viewtopic.php?f=4&t=5623&start=6 Re: Yet another new bitboard move generation method] by [[Harm Geert Muller]], [[Computer Chess Forums|Winboard Forum]], October 01, 2006 <ref>[http://www.talkchess.com/forum3/viewtopic.php?f=7&t=52861&start=8 Re: multi-dimensional piece/square tables] by Tony P., [[CCC]], January 28, 2020</ref>
 
* [http://www.talkchess.com/forum/viewtopic.php?t=17790 Is it time for another new move generator?] by [[Michael Sherwin]], [[CCC]], November 11, 2007  
 
* [http://www.talkchess.com/forum/viewtopic.php?t=17790 Is it time for another new move generator?] by [[Michael Sherwin]], [[CCC]], November 11, 2007  
 
* [http://www.talkchess.com/forum/viewtopic.php?t=17820 Did someone mention the GNUChess move Generator?] by [[Michael Sherwin]], [[CCC]], November 12, 2007 » [[GNU Chess]]
 
* [http://www.talkchess.com/forum/viewtopic.php?t=17820 Did someone mention the GNUChess move Generator?] by [[Michael Sherwin]], [[CCC]], November 12, 2007 » [[GNU Chess]]
 +
* [http://www.talkchess.com/forum3/viewtopic.php?f=7&t=19837 compact bitboard move generator] by [[Robert Hyatt]], [[CCC]], February 25, 2008 » [[Bitboard Serialization]]
 
* [http://www.talkchess.com/forum/viewtopic.php?t=20630 Move generator] by [[Chua Kong Sian|kongsian]], [[CCC]], April 12, 2008
 
* [http://www.talkchess.com/forum/viewtopic.php?t=20630 Move generator] by [[Chua Kong Sian|kongsian]], [[CCC]], April 12, 2008
 
* [http://www.talkchess.com/forum/viewtopic.php?t=25917 Bitboards / move generation on larger boards] by [[Gregory Strong]], [[CCC]], January 09, 2009  
 
* [http://www.talkchess.com/forum/viewtopic.php?t=25917 Bitboards / move generation on larger boards] by [[Gregory Strong]], [[CCC]], January 09, 2009  
Line 152: Line 155:
 
* [http://www.talkchess.com/forum/viewtopic.php?t=56049 Questions about chess programming from a newbie] by Matt Palmer, [[CCC]], April 18, 2015
 
* [http://www.talkchess.com/forum/viewtopic.php?t=56049 Questions about chess programming from a newbie] by Matt Palmer, [[CCC]], April 18, 2015
 
* [http://www.talkchess.com/forum/viewtopic.php?t=56299 Caching generated moves list in recursive searches] by [[Rein Halbersma]], [[CCC]], May 10, 2015
 
* [http://www.talkchess.com/forum/viewtopic.php?t=56299 Caching generated moves list in recursive searches] by [[Rein Halbersma]], [[CCC]], May 10, 2015
 +
* [http://www.talkchess.com/forum3/viewtopic.php?f=7&t=58433 An approach to precomputed move generation bitboards] by [[Michael Sherwin]], [[CCC]], December 01, 2015
 
'''2016'''
 
'''2016'''
 
* [http://www.talkchess.com/forum/viewtopic.php?t=61020 speed up your engine part 4] by [[Laurie Tunnicliffe]], [[CCC]], August 03, 2016 » [[Move Generation#Staged|Staged move generation]]
 
* [http://www.talkchess.com/forum/viewtopic.php?t=61020 speed up your engine part 4] by [[Laurie Tunnicliffe]], [[CCC]], August 03, 2016 » [[Move Generation#Staged|Staged move generation]]
Line 161: Line 165:
 
* [http://www.talkchess.com/forum/viewtopic.php?t=64619 History heuristic and quiet move generation] by [[Daniel José Queraltó]], [[CCC]], July 16, 2017 » [[History Heuristic]]
 
* [http://www.talkchess.com/forum/viewtopic.php?t=64619 History heuristic and quiet move generation] by [[Daniel José Queraltó]], [[CCC]], July 16, 2017 » [[History Heuristic]]
 
* [http://www.talkchess.com/forum/viewtopic.php?t=65900 Skipping duplicat moves] by [[Harm Geert Muller]], [[CCC]], December 03, 2017
 
* [http://www.talkchess.com/forum/viewtopic.php?t=65900 Skipping duplicat moves] by [[Harm Geert Muller]], [[CCC]], December 03, 2017
 +
'''2019'''
 +
* [http://www.talkchess.com/forum3/viewtopic.php?f=7&t=70082 Opinions requested for new move gen idea]  by [[Michael Sherwin]], [[CCC]], March 03, 2019 » [[Table-driven Move Generation]]
 +
* [http://www.talkchess.com/forum3/viewtopic.php?f=7&t=70508 My newest almost bb move generator is wonderful] by [[Michael Sherwin]], [[CCC]], April 17, 2019
 +
* [http://www.talkchess.com/forum3/viewtopic.php?f=7&t=70515 Efficient capture generation in the game of Thud] by koedem, [[CCC]], April 18, 2019
  
 
=External Links=  
 
=External Links=  

Revision as of 10:03, 9 April 2020

Home * Board Representation * Move Generation

Blast Generation [1]

Generation of moves is a basic part of a chess engine with many variations concerning a generator or an iterator to loop over moves inside the search routine. The implementation heavily depends on the board representation, and it can be generalized into two types, pseudo-legal and legal move generation.

Legality

Pseudo-legal

In Pseudo-legal move generation pieces obey their normal rules of movement, but they're not checked beforehand to see if they'll leave the king in check. It is left up to the move-making function to test the move, or it is even possible to let the king remain in check and only test for the capture of the king on the next move.

Legal

In Legal move generation, as the name implies, only legal moves are generated, which means extra time must be spent to make sure the king isn't going to be left or placed in check after each move. Pins are the main difficulty, particularly when en passant is involved.

Special Generators

Most programs use special move generators for the quiescence search, sometimes supplemented by one for getting out of check. These special cases can be made more efficient than generating and testing each possible move to fit specific criteria. For example, if the king is in check, the only possible legal moves are to capture the attacking piece, block the attacker if it is a "ray" piece, or move the king to safety. Special generators for the quiescence search might want to generate checks in addition to captures and promotions. They can use the fact that a knight or bishop must start off on the same color square as the opponent king if they are to attack it. And rooks can only generate at most 2 checking moves...to the squares with the rook's column and the king's row or the king's column and the rooks row.

Similar tricks can be used for generating possible moves out of check, which must be by the king, capturing the opponent's checking piece, or blocking its attack if it is a ray piece. When in double check, only king moves are permitted.

Chunk move generation

With move ordering in mind, chess programs, while traversing pieces and their move-target sets once, store and buffer generated moves inside one or two move lists (i.e. for tactical and quiet moves), which is convenient for book-keeping and assigning scores based on MVV-LVA, SEE, history, piece square table etc., to later perform a selection sort before actually making the move.

Staged move generation

Some programs do not generate all moves at once, but do it in several stages (i.e. hash move first, then captures, then killer moves, then all the rest in a chunk) on the premise that if one of the early moves causes a cutoff, then we may save on the effort of generating the rest of the moves [2].

Debugging

It is important to ensure that the move generator works properly. Although this could be tested by playing many games, a better approach is to write a Perft function. This function recursively generates moves for the current position and all children up to a certain depth, and by counting all the leaf nodes, it can be compared to a table of values to test its accuracy.

See also

General

Board Array

0x88
Vector Attacks

Bitboards

Hardware

Chess Programs

Publications

1950 ...

1970 ...

1980 ...

1990 ...

2000 ...

2010 ...

Forum Posts

1990 ...

Re: move generators in computer chess, Tricky bit tricks by Marcel van Kervinck, rgcc, October 20, 1994 » Traversing Subsets of a Set

2000 ...

2005 ...

Re: Yet another new bitboard move generation method by Harm Geert Muller, Winboard Forum, October 01, 2006 [4]
Re: Move generation: staged vs all-at-once by Lance Perkins, CCC, April 30, 2009

2010 ...

2012

2014

2015 ...

2016

2017

2019

External Links

generate.c by Scott Gasch

References

Up one level