605
edits
Changes
→Quotes
I create both WDL and DTZ in one go, so I don't use WDL in the creation of DTZ. The algorithm used is the [[Retrograde Analysis#Algorithm|grandfather algorithm]] with 2 plies per iteration (I think [[Harm Geert Muller|HGM]] calls this leapfrogging, but I might be wrong). I tried the outcounting method, but it didn't seem to be competitive (and it makes things more complicated). <ref>[http://www.open-chess.org/viewtopic.php?f=5&t=779 Retrograde tablebase methods] by [[Mark Watkins|BB+]], [[Computer Chess Forums|OpenChess Forum]], November 26, 2010</ref> <ref>[http://home.hccnet.nl/h.g.muller/EGT7/retro.html Leapfrog: Retrograde Analysis] from [http://home.hccnet.nl/h.g.muller/EGT7/7-men.html Leapfrog tablebase generator] by [[Harm Geert Muller]]</ref>
A pure WDL/DTZ pair is not of much use for creating WDL50+/DTZ50+. I create tables in RAM that have all the information necessary for WDL50+ and DTZ50+, then permute them to different indexing schemes and compress. I do test runs on subsets of the data to find good permutations. (The idea to try permutations is from [[Jesper Torp Kristensen|Jesper Torp Kristensen's]] master thesis.) <ref>[[Jesper Torp Kristensen]] ('''2005'''). ''[https://issuu.com/jespertk/docs/master_thesis Generation and compression of endgame tables in chess with fast random access using OBDDs]''. Master thesis, supervisor [[Mathematician#Miltersen|Peter Bro Miltersen]], [https://en.wikipedia.org/wiki/Aarhus_University Aarhus University]</ref> <ref>[https://en.wikipedia.org/wiki/Binary_decision_diagram OBDD - Ordered Binary Decision Diagram from Wikipedia]</ref>
[[Ronald de Man]] explained some advantages of Syzyzy over [[Nalimov Tablebases]], July 06, 2019 <ref>[https://www.talkchess.com/forum3/viewtopic.php?p=804007#p804007 Re: DTM50] by [[Ronald de Man]], [[Computer Chess Forums|CCRL Discussion Board]], July 06, 2019</ref> :
They are smaller, so more is cached in RAM (via the OS's system/page cache), and they are compressed in small blocks which allows decompression on the fly (so no need to use RAM for caching large decompressed blocks). This makes them much faster on the same hardware, which allows probing much closer to the leaves. They also initialise a lot faster (initialising 6-men Nalimov on startup of the engine seems to have taken about half an hour in the past, probably it is faster on modern hardware but still).
In addition, 7-men syzygy TBs are now available. With SSDs becoming cheaper and cheaper, having 7-men WDL TBs on SSD starts to become a practical possibility (unfortunately RAM prices are lagging behind).
[[Ronald de Man]] in a reply to [[Pham Hong Nguyen|Nguyen Pham]], April 15, 2020 <ref>[http://talkchess.com/forum3/viewtopic.php?f=7&t=73598 Re: Almost perfect DTM tablebase] by [[Dann Corbit]], [[Computer Chess Forums|CCRL Discussion Board]], April 15, 2020</ref> :
