English Chess Forum

Recently read about the Lomonosov tablebases, which are 7-man endings:

http://chessok.com/?page_id=27966

I wasn't aware that these were available. Has anyone here made any practical use of them during analysis?

Some old discussion here as well:

http://www.chess.com/forum/view/general/tablebases

Is there any recent news on 8-man tablebases?

I play a lot of correspondence chess . Id never heard of them...until a couple of days ago when I got this from an opponent (in a queen and pawn ending...well...we all know they are drawn ...right ?).....

"Lomonosov 7 man tablebases mate in 41 "
Most annoying...

Hi David.

But that is not fair, he used outside help.
He has obvioulsy pasted the position into a computer or perhaps has
had the whole thing ticking over in a box till it hit it's tablebase.
(or are you on one of these chess sites that allow that kind of thing.)

Don't resign yet, Buy the program and check it, he may be bluffing.

EDIT:
Just had to flick back to RHP to see if we allow this:

Apparently not.

"While a game is in progress you may not refer to chess engines, chess computers
or be assisted by a third party. Endgame tablebases may not be consulted during play
but you may reference books, databases consisting of previously played games between
human players, and other pre-existing research materials. "

He has admitted he used a tablebase.
Where you play is this allowed?

Its the ICCF webserver...computers etc perfectly allowable.....
I doubt he is bluffing . Its a Queen and 2 pawns v my Queen and 1 pawn . It looks like the sort of position I can just check forever if it was an over the board game....but clearly not at this level...

Hi David.

So, it's all fair then.

I just looked at the link Chris posted. That site does not allow computers either.
(though the post is months old - perhaps now they do.)

You can tell me to mind my own business but is allowing computers not a tad pointless.
Do they not just have computers playing computers most of the time
with a human player as the punch bag.
Not a dig David, each to his own, if you enjoy it then fair enough.
(do they post these games in a forum?)

I've have been playing with that other link Chris gave.

They have a spot where you can enter a position and it tells you within seconds
if it's won and if so how to win it.
Which is a very clever thing to have online when you have all these smurfs running
around with handheld micro phones that can access that net.

I've been playing with it.
I thought with the clues above QPP v QP and 41 moves I could juggle the
bits about till it said mate in 41, post the position and say is this it?

But it says 'No Data' (OK too many bits) But even when I go for 5 bits.

4k3/3q4/8/8/8/4P1Q1/5K2/8 w - - 0 1
It still comes back and says 'No Data' So maybe your opponent is bluffing after all.

(or, hopefully, I've broken the bloody thing.)

So somewhere in there now, presumably, is the following famous study by Troitzy (White to win):
3K4/8/p1k3N1/1p6/6Q1/1q6/8/8

John Clarke wrote:So somewhere in there now, presumably, is the following famous study by Troitzy (White to win):
3K4/8/p1k3N1/1p6/6Q1/1q6/8/8

Presumably. Maybe someone with access can look it up and see what the verdict is? Ditto other 7-man endgame studies (are there many?).

More on this here (another old thread on another forum):

http://rybkaforum.net/cgi-bin/rybkaforu ... l?tid=8477

I completely fail to understand sometimes how these mates in 545 or whatever are calculated. Sounds impressive, but I'm never sure if there isn't something slightly suspect about some of the numbers.

The game Dominguez-Nakamura from the Thessaloniki Grand Prix (round 6) reached a seven-man K+P ending on move 47 after the rooks were exchanged:

8/8/1pk5/p5P1/2P3K1/8/7P/8 w - - 0 1

Black to move, what result? I wonder what the Lomonosov tablebases say? The move played was Kd6.

The game later reached a Q+P endgame, with this position with White to move in a five-man position:

8/5Q2/3k3K/1q6/7P/8/8/8 w - - 0 1

It is a win for White, according to the tablebases, but only after three queen moves with check (Qf6, Qf4, Qg6) or Kg7. The seemingly obvious h5 only draws. I think it is to do with h5 allowing Black to play the critical move Ke5 (from d6), preventing the normal cross-checks that would force queens off and allow White to win. But Dominguez seemingly missed that. The game ended in a draw.

Does anyone here actually have access to the Lomonosov tablebases?

Does anyone here think that tablebases help understanding?

Francis Fields wrote:Does anyone here think that tablebases help understanding?

I reckon so. R+P v R is the most common and it can be instructive to follow through a game on the Shredder site to see where a draw becomes a win and vice -versa. Being able to recognise whether a position is, or should be, a draw or win, can be important in calculation.

You have to be wary of on-line engines when they aren't using the tables. In the last round of the Olympiad, the chessbomb engine was quite sure that David Howell had an advantage equivalent to about two pawns even after it simplified to Queen and Bishop v Queen and Pawn. That it got the final position wrong also suggests it was over optimistic earlier about Black's winning chances.

Francis Fields wrote:Does anyone here think that tablebases help understanding?

They can be useful for practising your play in difficult endings when used in conjunction with a chess engine, and seeing where you went wrong if you didn't get the right result.

I think this is as good a thread as any to put a link to two excellent articles from Chessbase on tablebases and how some positions defy human understanding:

http://en.chessbase.com/post/perfect-en ... tablebases
http://en.chessbase.com/post/just-one-o ... ositions-2

Play through the 262-move mate in the second article and try and understand what is going on. Anyone?

John Clarke wrote:So somewhere in there now, presumably, is the following famous study by Troitzy (White to win):
3K4/8/p1k3N1/1p6/6Q1/1q6/8/8

Yes. White mates in 20!