English Chess Forum

I'm not sure if this is the right board for this issue. I have come across a technical problem which I can't resolve online and wondered if anyone else had seen it.

I have signed up to play correspondence chess in this year's county championship. Part of the documentation says that claims of a forced win or draw with optimal play in positions with a reduced number of pieces will be decided with reference to Lomonosov 7-piece tablebases. These are a recent development and only seem to be available through the Windows database programs of a company called Chess OK. I bought and downloaded the latest version (Aquarium 2022). Because the tablebases are so large (about 140,000 Gb) they are accessed through an online server to which the program is supposed to connect. But when I try to connect I get the error message "Error 10061. No connection could be made because the target machine actively refused it." Online research suggests that this might relate to file corruption or the use of a firewall. But I've investigated these and they're not causing the problem. I have contacted Chess OK but no-one has got back to me.

Yes, there's something wrong at their end. I've used the Lomonosov tablebases quite a lot over the past few years and it's only recently they've started giving the Error message. No one seems to be sorting it out either, which is annoying. There are other seven-piece databases online, but they only say "win" or "draw"; Lomonosov always gave more details.

If you have an Android device you should be able to access them for free using https://play.google.com/store/apps/deta ... omonosovtb.

Whether that'll fix the problem you report, I don't know, because I don't have any Android devices to install the app on.

Ian Thompson wrote: ↑

Fri Jan 21, 2022 3:49 pm

If you have an Android device you should be able to access them for free using https://play.google.com/store/apps/deta ... omonosovtb.

Whether that'll fix the problem you report, I don't know, because I don't have any Android devices to install the app on.

Just tried the app on my phone, which I haven't looked at for ages. It says Connection Failed.

Same problem if you try here https://www.iccf.com/message?message=641

Follow the link and people are grumbling it's not working. I thnk syzygy is though.

I took this up with the ICCF overnight. They confirmed that the syzygy route still works for positions with up to 7 pieces: go to https://syzygy-tables.info/.

Is there a technicality where you are not allowed to use a table base like the one on Lichess?
https://lichess.org/blog/W3WeMyQAACQAdf ... e-complete

I would assume that all table bases have the same outcome (win/draw/loss) but maybe different pathing to it?

"Is there a technicality where you are not allowed to use a table base like the one on Lichess?"

I assume the ones mentioned on ICCF site are the ones used. I made such a claim recently as an opponent clearly was not going to resign. The process was either automatic, or someone was hovering over the computer waiting for a claim and assessed it immediately. The Lichess one might work, all that happens if your claim is wrong is that you have to play the move (if any) you made when making the claim. (e.g. I intend to play Rf6 repeating the position for the third time.)

My opponent was very unhappy and sent me a long rant, complaining that the decision was wrong and I was unsporting to claim it anyway. In another game, the opponent played on Q+R down, until mated. Well, you've paid your entry fee, but nobody is going to mess that up at CC, unless they suffer a catastrophic internet failure for 6 weeks, or become seriously ill. The practical point of view from ICCF is that if you reach a tablebase ending, you are just going to play the tablebase moves, so why not just claim the result. I would add that the opponent was clearly using Stockfish (e.g.), but believing what it said after about 1 minute, rather than waiting to see if it changed its mind.

Kevin Thurlow wrote: ↑

Sat Jan 22, 2022 12:45 pm

complaining that the decision was wrong

How can we be sure that there aren't any bugs in the new(ish) tablebases? If your opponent thought he had found one, I assume that could have been presented as evidence in an appeal and would have been of value to the whole community.

NickFaulks wrote: ↑

Sat Jan 22, 2022 12:55 pm

Kevin Thurlow wrote: ↑

Sat Jan 22, 2022 12:45 pm

complaining that the decision was wrong

How can we be sure that there aren't any bugs in the new(ish) tablebases? If your opponent thought he had found one, I assume that could have been presented as evidence in an appeal and would have been of value to the whole community.

Bugs in what sense? Incorrect calculation? Or a database glitch? (I think tablebases are not strictly speaking databases, though quite what they are, I am not sure. Tablebases, maybe?)

https://en.wikipedia.org/wiki/Endgame_tablebase

OK, they are still database, just very specialised ones.

"How can we be sure that there aren't any bugs in the new(ish) tablebases?"

A fair point, but I can modestly say that the tablebase confirmed my analysis in this case!

Christopher Kreuzer wrote: ↑

Sat Jan 22, 2022 1:00 pm

Bugs in what sense?

In the sense that they say a position is a win with best play for both sides when it is in fact a draw.

Tablebases, by definition, are "solved" positions that engines may stlll get wrong.The only issue could be a bug in the software retrieving the database position.