#6 piece endgame tablebase update#
#6 piece endgame tablebase Pc#

Wrong #1: The resources for doing the eg KQRR-kqrr (with only 2T positions, or other pawnless 8-men tablebases) were available 10 years ago (the hardware would have been about 12.000 EUR in those days). s# tablebases (and h# tablebases) are probably easier because they're a lot less deep (compare max depth of the 7-piece s# TBs to the max depth of # TBs.), and therefore require less resources. The idea of 8-man tablebases exists ever since the first idea of tablebases, it's just that until recently the resources needed for 8-men TBs weren't available. The material I picked (without Queen) was just particularly suitable.

#6 piece endgame tablebase Pc#

In 1996, when generating P1014870 and P1014888 I used a PC with at most 8MB main memory. It gets more interesting when pawns are on the board. These days that's doable on a medium-sized workstation or even an upper-end desktop PC.

That's less than the complete 6-men tablebases. Bourzutschky picked the material KQRR-kqrr with at most 2222255026372 position. There are subsets of the tablebase that can be generated independently of the rest of the tablebase. What computer resources did you need to create 7-men tablebases? Torsten, I think such tablebases require a supercomputer, which I unfortunately don’t have (I created the above problems without using any tablebases). (3) Posted by Torsten Linß īTW, the idea of 8-men tablebases is not very original:Įven 7-men-tablebase problems had been published 16 year prior to the Lomonossov tablebases: Marc Bourzutschky provides some incredible variations and welcome explanations at the link above. One of the interesting new outcomes obtained from 8-Men Tablebase: While 15% seems like a small subset at first blush, most other piece configurations have large material differences between White and Black so that long lines are unlikely. After generating about 15% of the pawnless endings I’m quite confident to have captured the longest ones. My results suggest that we may already be at or close to this saturation point: the longest winning line for 8-man endgames without pawns appears to be “only” 400 moves. This is worth reading in detail, but one point demands quotation:Īn important question is at what point the chess board becomes so crowded that adding more pieces does not lead to longer winning lines due to the increased likelihood of shortening captures.

#6 piece endgame tablebase update#

(1) Posted by Andrew Buchanan Īn important update on 8-Men Tablebase by Marc Bourzutschky.