How is work on 8-man tablebases progressing?

Discussion of chess software programming and technical issues.

Moderator: Ras

Rein Halbersma
Posts: 749
Joined: Tue May 22, 2007 11:13 am

Re: How is work on 8-man tablebases progressing?

Post by Rein Halbersma »

hgm wrote: ↑Thu Sep 19, 2024 7:58 am Apparently not, because I cannot find the word 'correct' in that posting at all...

The point is that in a generally won end-game most P-slices can be solved as total wins even under rules where opponent promotion is an immediate loss, because in practice you won't be able to beat that extra Queen. Only the small fraction of P-slices for which this is not the case you won't be able to solve.

Another way of looking at it, is that instead of a WDL bitbase, one could use a WUL bitbase, where U means 'undefined'. A probing engine would then have to search on when it hits a U position. And guess what, they are already doing that when they hit a D position. So not much changes for the user. Only some of the W or L would now be a U too. Mostly for positions that a game would never reach, such as multiple Pawns on the pre-promotion rank for both sides.
The Chinook team pioneered this approach of partial information endgame databases having bounds instead of exact values: https://link.springer.com/chapter/10.1007/11922155_2
Top
syzygy
Posts: 5696
Joined: Tue Feb 28, 2012 11:56 pm

Re: How is work on 8-man tablebases progressing?

Post by syzygy »

hgm wrote: ↑Thu Sep 19, 2024 7:58 amAnother way of looking at it, is that instead of a WDL bitbase, one could use a WUL bitbase, where U means 'undefined'. A probing engine would then have to search on when it hits a U position. And guess what, they are already doing that when they hit a D position. So not much changes for the user. Only some of the W or L would now be a U too. Mostly for positions that a game would never reach, such as multiple Pawns on the pre-promotion rank for both sides.
So you basically have no information for drawn positions?
(I think you do have information for them, but it is only "usually correct" information.)

It can have its uses. In chess there is no chance that partial or complete 8-men tables will help you solve the game, but being able to "cheaply" generate a partial table that may help you analyze a particular position could be very useful.

But if you generate enough partial tables, you will end up with enough complete tables that you can replace partial tables with less-partial, more-complete tables, until everything is complete. Unless you limit yourself beforehand to a restricted set of partial tables and then stick to that, it seems to me you could as well go for complete tables from the start (and have less complexity in the generator).
Top
Uri Blass
Posts: 10798
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: How is work on 8-man tablebases progressing?

Post by Uri Blass »

Viz wrote: ↑Tue Sep 17, 2024 9:39 pm
Jouni wrote: ↑Thu Sep 12, 2024 5:31 pm I think 8-man can give 0-5 elo gain like 7-man. No practical benefit. Only academic interest. I don't use even 6-man. If there is interesting position I check DTM from Shredderchess :) .
You are wrong.
With current technology 8 man will lose elo because they will slow down engine a lot during probing because of their enormous size.
No.
It is not going to lose elo because engines are not going to use it if it lose elo.
I see no reason to use it in every node and engines can use it only in the first 1% of the time that they search.
Top
Uri Blass
Posts: 10798
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: How is work on 8-man tablebases progressing?

Post by Uri Blass »

syzygy wrote: ↑Tue Sep 17, 2024 10:44 pm
Leo wrote: ↑Tue Sep 17, 2024 7:33 pm I found this about 10 years ago.

Many show interest in what is to expect from 8-man endings. First, take note that the longest 6-man mate took 262 moves (KRN-KNN). Moving to 7-man endings doubled this value. Second, 8-man tablebases include much more endings with both sides having relatively equal strength. All this gives us a strong hope to discover a mate in more than 1000 moves in one of 8-man endgames. Unfortunately the size of 8-man tablebases will be 100 times larger than the size of 7-man tablebases. To fully compute them, one will need about 10 PB (10,000 TB) of disk space and 50 TB of RAM. Only the top 10 supercomputers can solve the 8-man problem in 2014. The first 1000-move mate is unlikely to be found until 2020 when a part of a TOP100 supercomputer may be allowed to be used for solving this task.
The idea that the longest mate must double with every extra piece (as someone has "predicted" in a paper) is really based on nothing.

I believe most pawnless candidates for longest mates have already been computed, and the longest mate record was increased only by a bit. See here:
https://en.chessbase.com/post/8-piece-e ... -interview

Mate in 584. It seems the forum's pgn viewer stops at 300 moves.
I see the position is with blocked pawns(a pawn block another pawn) so I wonder what is the record for longest mate with block pawns for 4,5,6,7 pieces tablebases and how much it is smaller than the record for non blocked pawns.

We can also talk about all positions with 2 pairs of blocked pawns when it is easier to calculate the 8 piece tablebases and maybe it is possible to use them to find some 9 piece tablebases(for example white pawns at a6 h6 and black pawns at a7 h7) so it may be interesting first to have the records for 2 blocked pawns for 6,7 piece tablebases and hopefully also full tablebases for 8 piece tablebases with 2 pair of blocked pawns).
Top
Uri Blass
Posts: 10798
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: How is work on 8-man tablebases progressing?

Post by Uri Blass »

posting the full pgn(previous post did not work because of some "{" after 300 moves that I simply deleted in editing)

Top
Uri Blass
Posts: 10798
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: How is work on 8-man tablebases progressing?

Post by Uri Blass »

Uri Blass wrote: ↑Fri Sep 20, 2024 6:43 am posting the full pgn(previous post did not work because of some "{" after 300 moves that I simply deleted in editing)


I do not see a mate in 584 and the last move 584.Re6 is not mate(based on chest the last position in the pgn is mate in 9 that means the original position is mate in 592 assuming the moves are the best).

Edit:The line is not the best moves and 581.Kh4 is not mate in 13 based on chest.
Chest continues to calculate move 581 to find the distance to mate at that point.

I guess 584 is the distance to conversion when the distance to mate is not clear at this point of time and I wonder if people can calculate it.
Top
Guenther
Posts: 4718
Joined: Wed Oct 01, 2008 6:33 am
Location: Regensburg, Germany
Full name: Guenther Simon

Re: How is work on 8-man tablebases progressing?

Post by Guenther »

Uri Blass wrote: ↑Fri Sep 20, 2024 7:02 am
Uri Blass wrote: ↑Fri Sep 20, 2024 6:43 am ...snip
{A new record position where White needs at least 584 moves to achieve
checkmate or transition to a won 7-man ending. This ending was checked with
two independent programs. The original result was obtained with Marc
Bourzutschky's general 8-man generator, the 2nd result from Yakov Konoval's
7-man generator, modified to handle the special situation of 8-man endings
withone pair of blocked pawns. Play near the end of the winning line may look
a little strange, because in the DTC metric the algorithm prefers positions
where Black loses the queen in N+1 moves to positions where he loses his pawn
in N moves. That does not affect the overall correctness of the line.} 1. Rf8

I do not see a mate in 584 and the last move 584.Re6 is not mate(based on chest the last position in the pgn is mate in 9 that means the original position is mate in 592 assuming the moves are the best).

Edit:The line is not the best moves and 581.Kh4 is not mate in 13 based on chest.
Chest continues to calculate move 581 to find the distance to mate at that point.

I guess 584 is the distance to conversion when the distance to mate is not clear at this point of time and I wonder if people can calculate it.
No need to guess it seems you have not read this part of your own quotes?
A new record position where White needs at least 584 moves to achieve checkmate or transition to a won 7-man ending.
https://rwbc-chess.de

[Trolls n'existent pas...]
Top
Uri Blass
Posts: 10798
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: How is work on 8-man tablebases progressing?

Post by Uri Blass »

Guenther wrote: ↑Fri Sep 20, 2024 7:51 am
Uri Blass wrote: ↑Fri Sep 20, 2024 7:02 am
Uri Blass wrote: ↑Fri Sep 20, 2024 6:43 am ...snip
{A new record position where White needs at least 584 moves to achieve
checkmate or transition to a won 7-man ending. This ending was checked with
two independent programs. The original result was obtained with Marc
Bourzutschky's general 8-man generator, the 2nd result from Yakov Konoval's
7-man generator, modified to handle the special situation of 8-man endings
withone pair of blocked pawns. Play near the end of the winning line may look
a little strange, because in the DTC metric the algorithm prefers positions
where Black loses the queen in N+1 moves to positions where he loses his pawn
in N moves. That does not affect the overall correctness of the line.} 1. Rf8

I do not see a mate in 584 and the last move 584.Re6 is not mate(based on chest the last position in the pgn is mate in 9 that means the original position is mate in 592 assuming the moves are the best).

Edit:The line is not the best moves and 581.Kh4 is not mate in 13 based on chest.
Chest continues to calculate move 581 to find the distance to mate at that point.

I guess 584 is the distance to conversion when the distance to mate is not clear at this point of time and I wonder if people can calculate it.
No need to guess it seems you have not read this part of your own quotes?
A new record position where White needs at least 584 moves to achieve checkmate or transition to a won 7-man ending.
Correct
I see also
"Play near the end of the winning line may look a little strange, because in the DTC metric the algorithm prefers positions where Black loses the queen in N+1 moves to positions where he loses his pawn in N moves. That does not affect the overall correctness of the line. "

In other words it may be also mate in significantly more moves because after losing the pawn instead of the queen the mate can be still very long.

I guess it is possible to have a bound for the distance to checkmate based on the longest line of 7 piece tablebases that is a subset of the original position.
Top
Uri Blass
Posts: 10798
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: How is work on 8-man tablebases progressing?

Post by Uri Blass »

chest find that 581.Kh5 is the only solution for mate in 15 in the following position(581.Kh4 is in the pgn and is the smallest distance for conversion).
8 
7 
6 
5 
4 
3 
2 
1 
abcdefgh

8/8/4q3/6B1/6K1/2B4p/3R3P/2k5 w - - 0 1
Top
Viz
Posts: 223
Joined: Tue Apr 09, 2024 6:24 am
Full name: Michael Chaly

Re: How is work on 8-man tablebases progressing?

Post by Viz »

Uri Blass wrote: ↑Fri Sep 20, 2024 6:03 am
Viz wrote: ↑Tue Sep 17, 2024 9:39 pm
Jouni wrote: ↑Thu Sep 12, 2024 5:31 pm I think 8-man can give 0-5 elo gain like 7-man. No practical benefit. Only academic interest. I don't use even 6-man. If there is interesting position I check DTM from Shredderchess :) .
You are wrong.
With current technology 8 man will lose elo because they will slow down engine a lot during probing because of their enormous size.
No.
It is not going to lose elo because engines are not going to use it if it lose elo.
I see no reason to use it in every node and engines can use it only in the first 1% of the time that they search.
Engines will use it or not use it solely dependent on how they are configured by people who run them.
I've myself seen stockfish having 1,5x slowdown in midgame because of slow 7 men TBs, for example, and person who ran it said it's "stockfish bug" and not him putting TBs on really slow SSD.
As per usual you confidently write smth in area which you have literally 0 clue about.
Top

Return to “Computer Chess Club: Programming and Technical Discussions”