English Chess Forum

I was reading up on various long games, including the game where someone resigned when they could have claimed a draw under the 50-move rule, and I then read that the longest theoretical winning position was 517 moves from a position involving a queen and knight versus a rook, bishop, and knight. That is a seven-man endgame, so have they worked out some of those but not all of them?

Christopher Kreuzer wrote:I was reading up on various long games, including the game where someone resigned when they could have claimed a draw under the 50-move rule, and I then read that the longest theoretical winning position was 517 moves from a position involving a queen and knight versus a rook, bishop, and knight. That is a seven-man endgame, so have they worked out some of those but not all of them?

According to Wikipedia (usual caveats apply), all of the six-man endgames have been done, and some work has been carried out on the seven-man endgames. The problem of finding adequate storage for the latter is likely to make progress very slow.

Christopher Kreuzer wrote:I was reading up on various long games, including the game where someone resigned when they could have claimed a draw under the 50-move rule, and I then read that the longest theoretical winning position was 517 moves from a position involving a queen and knight versus a rook, bishop, and knight. That is a seven-man endgame, so have they worked out some of those but not all of them?

You may be interested in the problems composed by Otto Blathy. He had notoriously long endgame puzzles.

I am not sure that I believe a mate in 517. Surely a computer couldn't 'think' this far ahead with any degree of accuracy?

Matthew Turner wrote:I am not sure that I believe a mate in 517. Surely a computer couldn't 'think' this far ahead with any degree of accuracy?

Its an endgame tablebase, so in essence, they start from the final position and work backwards. I'm sure there are people on this forum who can explain it in much more detail...

Andrew,
I understand the idea - you find a 4 piece ending which is very long, add on another piece with a long way of winning it and so on. However, when we are talking 517 moves each stage must be mammoth. I am not convinced computers are up to that.

The 517 move conversion is here: http://www.gothicchess.com/endings/ches ... n/game.htm

3-6 man endgame tablebases are here: http://kirill-kryukov.com/chess/tablebases-online/, downloadable if you have a spare 1100 GB or so of hard disk space.

Richard Thursby wrote:3-6 man endgame tablebases are here: http://kirill-kryukov.com/chess/tablebases-online/, downloadable if you have a spare 1100 GB or so of hard disk space.

How long would that take at typical home broadband speeds?!

Richard Thursby wrote:The 517 move conversion is here: http://www.gothicchess.com/endings/ches ... n/game.htm

3-6 man endgame tablebases are here: http://kirill-kryukov.com/chess/tablebases-online/, downloadable if you have a spare 1100 GB or so of hard disk space.

But Richard, the final position isn't yet mate. Anyway Rybka found an improvement for Black on move 334.

That is, as Stewart might say, what is known as a joke.

Sean Hewitt wrote:

Richard Thursby wrote:3-6 man endgame tablebases are here: http://kirill-kryukov.com/chess/tablebases-online/, downloadable if you have a spare 1100 GB or so of hard disk space.

How long would that take at typical home broadband speeds?!

I don't know if this is typical or not, but my broadband connection has a download speed of approximately 2MB/s. 1,100GB ~ 1,126,000MB, so that would take roughly 563,000 seconds - just under a week.