Complete 7-piece tablebases are out!

[Albert Silver]

I am a bit surprised this has not been commented on, but the lead programmers of ChessOk, the makers of Chess Assistant and Aquarium, developed an algorithm that has led to the complete 7-piece tablebases to be calculated in a 6-month period on the Lomonosov supercomputer at the Moscow State University. The Aquarium interface has been modified to allow limitless online consultation of the tablebases. Bear in mind that they occupy a staggering 140 thousand GB. This is huge news.

ChessOk main page

Overview of the new 7-piece tablebases

For those who wonder, I will merely point out that Victor Zakharov, one of the founders of Convekta back in the 80s, was already a PhD in computer science before he ever worked on Chess Assistant and co.

[Cubeman]

That is exciting news, many historic games can now be judged with the aid of 7 men TB. If I remember correctly that Spassky vs Fischer where Fischer played the famous Bxh2 that experts were undecided that if best play lead to a draw or win for Spassky, some key positions required 7 men TB to correctly judge outcome.

[reflectionofpower]

awesome

[Albert Silver]

As a follow-up, I have since acquired a copy of the latest Aquarium and have been able to test the 7-piece tablebases. In a word: amazing. The interface is designed so that you can have it give results on-the-fly, meaning you do not need to constantly consult the TBs manually, the results will appear automatically, with the moves, results, and mate scores (if any).

I have run this by several GMs, including one author who is using it to revise and correct an endgame work, and all were duly impressed. Oh, and just for the fun of it:



Mate in 549

[peter]

As a follow-up, I have since acquired a copy of the latest Aquarium and have been able to test the 7-piece tablebases. In a word: amazing.

Well, Albert, but there are still some positions missing to be read by Aquarium, even if they may be stored at Lomonosov meanwhile, e.g. this one still gives no result:

1rr5/8/7Q/7P/3k4/6K1/P7/8 w - -

I'm waiting for this special position now for quite a while.
When I bought my copy of Aquarium- Houdini, there were error messages like "cannot be found in Lomonosov", now there is only, after quite a while of waiting after pressing TB6 online, something like "server did not react in certain period"

[aturri]

Mate in 549

So it is an (actual, legal) theoretical draw.

The pawn can only make 2 moves till promotion. And there can be only 4 captures (including the capture of one of the 2 white Queens). So a maximum of 50*6 = 300 moves.

So before reaching the 549 move, there should be at least a 51 move sequence without moving a pawn nor capturing... that implies draw according the 50 move rule.

That is true except if this tablebases are going to be taking into account to avoid the 50 move rule, as there exists a "know" forced mate sequence in such position.

[hgm]

Well, 7-men tablebases are not that hard to calculate, and certainly calculating one locally would be much faster than any attempt to download it. It just requires a dedicated 1TB disk, 3GB of memory, and a few days of CPU time (per end-game).

It is just that it is not clear what you should do with it after you calculated one. It would require way too much storage to go on calculating more, and collect them all. On-line consulting of tablebases as a sort of cloud-computing project is pretty useless for letting an engine use them in search; it is way too slow.

[Albert Silver]

Well, 7-men tablebases are not that hard to calculate, and certainly calculating one locally would be much faster than any attempt to download it. It just requires a dedicated 1TB disk, 3GB of memory, and a few days of CPU time (per end-game).

It is just that it is not clear what you should do with it after you calculated one. It would require way too much storage to go on calculating more, and collect them all. On-line consulting of tablebases as a sort of cloud-computing project is pretty useless for letting an engine use them in search; it is way too slow.

Depends on the piece configuration and what you call a few days. KNNN-KNN was actually done by someone on a PC, and it took 36 days. Furthermore, his 7-piece TB generator, the only public one, may not be able to calculate all of them.

In any case, the purpose of such a resource is obviously not for an engine....

[yolin]

Here is the website of the author/creator

https://plus.google.com/100454521496393505718/posts

[Albert Silver]

Here is the website of the author/creator

https://plus.google.com/100454521496393505718/posts

It is the website where he documents the results, but he is the founder of Convekta/ChessOk, makers of Chess Assistant.