25,161
edits
Changes
no edit summary
'''Endgame Tablebases''', (EGTBs or EGTs)<br/>
precalculated [[Endgame|endgame]] [https://en.wikipedia.org/wiki/Table_%28database%29 tables] generated by an exhaustive [[Retrograde Analysis|retrograde analysis]]. During its [[Chess Game|game play]] and/or [[Search|search]], [[Interior Node Recognizer|recognizing]] a specific [[Material|material composition]], a chess program can probe, or in principle compute these tables to determine the outcome of positions definitively and act as an [[Oracle|oracle]] providing the optimal moves. The tables are usually divided into separate [https://en.wikipedia.org/wiki/Computer_file files] for each material configuration and [[Side to move|side to move]]. What all different [[Endgame Tablebases#Formats|formats]] of tablebases have in common is that every possible piece placement has its own, unique index number which represents the position where the information about the position is located in the file. Positions with non-null castling rights could be represented in an EGT EGTB but typically are not.
The term '''Tablebase''' opposed to endgame database was coined by [[Steven Edwards]] by means of a data vector and a single file access for both sides to move, when he introduced his [[Edwards' Tablebases]], addressing several shortcomings of [[Ken Thompson|Ken Thompson's]] earlier [[Thompson's Databases|database]] of positions with the weaker and hence potentially losing side to move only <ref>[[Ernst A. Heinz]] ('''1999'''). ''Endgame Databases and Efficient Index Schemes for Chess.'' [[ICGA Journal#22_1|ICCA Journal, Vol. 22, No. 1]], [http://people.csail.mit.edu/heinz/ps/edb_index.ps.gz ps]</ref> .
=Purposes=
When a [[Chess Game|chess game]] goes to the [[Endgame|endgame]] phase, it may complete soon with the result as a draw or a win/loss. Both human/chess engines can finish the endgame by continuing searching, applying some endgame rules. The game can finish in this way but it is not in a perfect way: it takes much time to make move, use more moves than necessary. Sometimes the winning side may lose the change since the rule 50 moves.
EGTB can help to solve this period. It is a kind of dictionary of all endgame positions which can answer directly or indirectly for a given position:
* it is a draw or a win/loss
* how far/how many move from matting/being mated
* what is the best move
The main advantages of using EGTBs:
* It does not search but retrieves data thus it can move instantly
* The move is the perfect one, leading the shortest line to win
The main disadvantages of using EGTBs:
* Data too large for downloading and storing
* If use when searching, it may slow down the search
* Require a complicated code for probing
=Data=
An EGTB is a set of endgames. Each endgame is a set of positions of that endgame that (they must have similar the same material). Each position has associates with an integer number which informs how far that position is from mating/being mated or converting (depends on the type of its metrics). All numbers of an endgame are simply organized as an array two arrays of integers(one array for one side/color). In other words, each item in that array those arrays is an integer and its index on an array is mapped to a unique position.
The size of an endgame depends on:
* The number of involving pieces: more pieces , the bigger size
* Type of pieces: for standard chess, more Pawns is the smaller size
* Type of metric
* The largest value: the integer number: should be large enough to cover the largest value. Different endgames will have different largest values. Thus some endgames require need only one byte per number/position item but some need 2 bytes* Indexing: The algorithms to map between an array item and a position* Compress ratio So far the size of an EGTB is the most important factor of success.
=Metrics=
* The side to move may not give check.
* The distance between the two Kings must be greater than one.
=Retrieving=
==Probe==
When a chess program (e.g., a chess engine or a chess GUI) works with an EGTB, it needs to retrieve data (integer numbers) for some given chess positions. The process named probe and has below steps:
* convert a given chess position into an index via indexing algorithms
* access data (typically organized as an integer array) using the index to retrieve the integer of that chess position
If data of that endgame is not ready in memory, usually the probe process has to do some extra work:
* From data index calculate to block index
* Read data of that block from storage into memory
* Uncompress the block (if it is compressed)
* Convert the endgame index into block index and retrieve the right data
==Search==
After probing the chess engine may have the result as an integer number for the queried position. Depending on the metric, the number may be a distance to mate or distance to conversion, the position is a draw or a win or loss for the querying side. Sometimes that value (state of draw/win/loss) may be enough for searching. However, sometimes that state is not enough and it needs to know which is the best move to make. The solution is to generate all legal moves from that position, make one by one, probe that EGTB for values of each new position. After all, compare the values of all children's positions to find which one is the best to make.
==Speed==
Typically the probe process requires some computing and it may access storage which is usually so slow, compared with one stored in memory. All may make it be a slow one. An engine that probes EGTB when searching may make the search be slower. The good point is that the engine can stop searching on the branch probed the EGTB.
=Brief Info=
* [[Christoph Wirth]], [[Jürg Nievergelt]] ('''1999'''). ''Exhaustive and Heuristic Retrograde Analysis of the KPPKP Endgame.'' [[ICGA Journal#22_2|ICCA Journal, Vol. 22, No. 2]]
* [[Eugene Nalimov]], [[Christoph Wirth]], [[Guy Haworth]] ('''1999'''). ''[http://centaur.reading.ac.uk/4564/ KQQKQQ and the Kasparov-World Game]''. [[ICGA Journal#22_4|ICCA Journal, Vol. 22, No. 4]]
* [[Ulrich Thiemonds]] ('''1999'''). ''Ein regelbasiertes Spielprogramm für Schachendspiele''. [https://en.wikipedia.org/wiki/University_of_Bonn University of Bonn], Diploma Diplom thesis, [httphttps://www.iaiidb.uni-bonn.de/~idbpublications/diplomarbeitenda/thiemonds99.psda_thiemonds_1999.gz zipped pspdf pdf] (German)
* [[Chrilly Donninger]] ('''1999'''). ''Der Fortschritt ist nicht aufzuhalten - Chrilly Donninger über Endspieldatenbanken''. [[Computerschach und Spiele|CSS]] 6/99, [http://www.mustrum.de/chrilly/tbs.pdf pdf] (German)
==2000 ...==
==Online Lookup==
* [https://www.chessdb.cn/queryc_en/ Chess Cloud Database Query Interface] by [[Bojun Guo|noobpwnftw]] [[Syzygy Bases|Syzygy]] DTZ50 probing
* [http://tb7.chessok.com/ Lomonosov - online 7-man tablebases]
* [http://chessok.com/?page_id=361 Endgame Nalimov Tablebases Online] by [[ChessOK]]
