Media comments on chess

[Matt Mackenzie]

Chess isn't going to be "solved" that way in any of our lifetimes, so I don't worry about it too much

[Stewart Reuben]

Roer >I think there will be hundreds or thousands. All drawn unless someone changes the stalemate rule.<

'They' could also make repetition illegal as in Go or Shogi. That will not solve the ultimate problem. Chess is a finite game and eventually computers and, to a lesser extent, top players will understand it too well. Perhaps this is evidenced by the recent World Championship virtual stand-off.
This won't matter at all to players of our strength, but at the very top, it will probably lead to waning interest.
There is an analogy with running the 1500 metres and motor cars doing it quicker. That has not meant people have lost interest. Of course it is not an exact analogy.

[Geoff Chandler]

Hi Justin,

No reference form Dawkins. The bibliography list nobody I recognise chess wise.
He does cite a J. Maynard Smith; "Game Theory and the Evolution of Fighting."
but whether or not he got the chess bit from there I do not know.

Hi Stewart,

Agree there will always be chess games amongst humans but know never to underestimate
the future of what computers will achieve. They are solving it backwards by Table Bases.

When they do solve it we should not be too surprised they have discovered that White is in zugzwang.

[Ian Thompson]

Page 61 of 'The Selfish Gene' by Richard Dawkins.

Isn't this much the same as what AplhaZero is doing? It's interesting to see that it had better results against Stockfish when left to play whatever it wanted (with no games lost) compared to when it had to play popular openings (with 2% of games lost), which might suggest that one or more of the popular openings is losing/very close to losing for one player with best play.

[Roger de Coverly]

Isn't this much the same as what is doing?

Engines in the 1970s were relatively weak. I'm aware that later in the 1990s some engines were programmed to learn their opening book from previous games. If engines in the 1970s varied their play, that could have been an evasive idea to prevent exploits.

If you find an idea where the engine's evaluation gets it wrong, you can win numerous games unless there's some form of randomisation so that the same moves aren't repeated.

What I' like to know is what Alpha Zero thinks of the position after 1. e4 e5 2. f4 and how it compares to the positions after 1. d4 d5 2. e4 and 1. e4 c5 2. d4.

[JustinHorton]

Page 61 of 'The Selfish Gene' by Richard Dawkins.

Isn't this much the same as what AlphaZero is doing?

This is what makes me suspicious about the claim that this was happening in 1978.

which might suggest that one or more of the popular openings is losing/very close to losing for one player with best play.

Might not be this one, but I do have a book on the King's Indian in which the author reports a fellow-enthusiast breezily suggesting that that opening might very well lose by force.

[Roger de Coverly]

This is what makes me suspicious about the claim that this was happening in 1978.

It was a book written in 1978. Given that I was there, the claim in the book might be marginally ahead of its time, but it was a logical idea, even if yet to be fully implemented.

[JustinHorton]

It says "have been used". So I was wondering which programs Dawkins was referring to.

[NickFaulks]

When I was at school around 1965, a scientifically minded classmate brought in a computer he had built which played noughts and crosses. I seem to remember that it involved matchboxes and Smarties. Starting from scratch, you played games against it and, by learning from its mistakes, before long it never lost. There was no human intervention in the thinking process and I was greatly impressed.

The concept was certainly well established forty years ago but, like Roger, I question whether the processing power existed to get far with in a game as complex as chess.

[Michael Yeo]

I think it was earlier, probably 1962, as it was based on an article in the Scientific American in March 1962 available online at http://cs.williams.edu/~freund/cs136-07 ... xapawn.pdf. This was based on work by Donald Michie There is some more info at http://www.mscroggs.co.uk/blog/19.

[NickFaulks]

I think it was earlier, probably 1962

That would have been our first term at school, which I think it wasn't. I'm willing to compromise on something in between.

Thanks for the references. I guessed it was the Scientific American, of which your father was a great proponent.

[JustinHorton]

Chess in the media three years ago Today

[Stewart Reuben]

CHESS TELECOM
https://www.crunchbase.com/organization ... n-overview '

I read about this company in the Sunday Times 'Question of Money' section. They are providers of phone and broadcast services. They do not have a high reputation and I have no idea why Chess is in their name.

[Roger de Coverly]

and I have no idea why Chess is in their name.

The River Chess in Bucks (Chesham) is a possibility, although they don't seem to have any geographic connection.

[JustinHorton]

Also their location in Cheshire