Using a computer in correspondence chess

Discuss anything you like about chess related matters in this forum.
Kevin O'Rourke
Posts: 190
Joined: Fri Nov 20, 2015 4:01 pm

Using a computer in correspondence chess

Post by Kevin O'Rourke » Thu Jan 07, 2016 5:48 pm

Since the invention of really strong computers that can beat a Grandmaster and being available to everyone, has correspondence chess declined somewhat? It would be very easy to cheat either the whole game or at a crucial point in one.

Most online chess sites can tell if you’re using a computer (I once had one flash up a warning and I was only viewing chess sets in another window) but the long games you could do it whenever you want.


I’d hope most people would want to test their skills rather than cheat and save the computer analysis till after a game is finished.

Roger de Coverly
Posts: 18188
Joined: Tue Apr 15, 2008 2:51 pm

Re: Using a computer in correspondence chess

Post by Roger de Coverly » Thu Jan 07, 2016 6:45 pm

Kevin O'Rourke wrote: Most online chess sites can tell if you’re using a computer (I once had one flash up a warning and I was only viewing chess sets in another window) but the long games you could do it whenever you want.
They can only tell if you are using the same computer, hence the spurious message.

Correspondence chess has indeed at the higher levels become a matter of man-machine chess.

NickFaulks
Posts: 5173
Joined: Sat Jan 02, 2010 1:28 pm

Re: Using a computer in correspondence chess

Post by NickFaulks » Thu Jan 07, 2016 10:11 pm

Roger de Coverly wrote: Correspondence chess has indeed at the higher levels become a matter of man-machine chess.
A good development in my view. With a strong blunder-checker at hand, players can concentrate on the deeper elements of the game.

Brian Towers
Posts: 1228
Joined: Tue Nov 18, 2014 7:23 pm

Re: Using a computer in correspondence chess

Post by Brian Towers » Fri Jan 08, 2016 1:16 pm

Kevin O'Rourke wrote:I’d hope most people would want to test their skills rather than cheat and save the computer analysis till after a game is finished.
Nonsense!

Just as in days gone by there was a lot more to correspondence chess than Informator, Shakmatny Bulletin and MCO, so today there is a lot more to it than just computers.

Somebody armed with a powerful computer and the most powerful engine but little or no chess ability would get crushed in today's top level correspondence chess. Of course they would probably beat the likes of you and me, but not the experts.
Ah, but I was so much older then. I'm younger than that now.

Kevin O'Rourke
Posts: 190
Joined: Fri Nov 20, 2015 4:01 pm

Re: Using a computer in correspondence chess

Post by Kevin O'Rourke » Fri Jan 08, 2016 1:30 pm

Somebody armed with a powerful computer and the most powerful engine but little or no chess ability would get crushed


Why do you say that, what if they let the computer think for hours before each move? What do the other guys do that is better than the most powerful engine?

MartinCarpenter
Posts: 2454
Joined: Tue May 24, 2011 10:58 am

Re: Using a computer in correspondence chess

Post by MartinCarpenter » Fri Jan 08, 2016 1:39 pm

Well use multiple engines for starters ;)

I imagine there's still some points where human judgement helps too. Engines do still have some non trivial shared weaknesses, its just that in practical play those are absolutely overwhelmed by their tactical strengths.


Kevin O'Rourke
Posts: 190
Joined: Fri Nov 20, 2015 4:01 pm

Re: Using a computer in correspondence chess

Post by Kevin O'Rourke » Fri Jan 08, 2016 2:17 pm

So computers play completely tactical is that correct?

I thought a lot of the newer ones a had strategic side to their game

User avatar
Michael Farthing
Posts: 1828
Joined: Fri Apr 04, 2014 1:28 pm
Location: Morecambe, Europe

Re: Using a computer in correspondence chess

Post by Michael Farthing » Fri Jan 08, 2016 3:06 pm

Computers have always had a strategic side but are inevitably better at tactics. A particular position can be given a good strategic assessment, but that takes time - may be as much as a microsecond!! [That's a guess actually - but it's making a point] So the strategic assessment for a particular position can be made very strong, but to do it for all positions popping out of the tactical analysis (maybe thousands of millions) makes that microsecond seem a very long time. Refinement, of course, involves the computer knowing when a position is so bad that further tactical analysis is pointless, or where a position is fundamentally lacking tactical opportunities so that the strategic assessment at that point is what to run with. That sounds simple, but even making this decision is going to involve time!

Kevin O'Rourke
Posts: 190
Joined: Fri Nov 20, 2015 4:01 pm

Re: Using a computer in correspondence chess

Post by Kevin O'Rourke » Fri Jan 08, 2016 3:11 pm

i think i can wait a microsecond :-)

if you give the computer 30 seconds does that mean it goes really really deep or has it made up it's mind in about 2 seconds.

Adam Ashton
Posts: 78
Joined: Wed Feb 24, 2010 7:37 pm

Re: Using a computer in correspondence chess

Post by Adam Ashton » Fri Jan 08, 2016 3:26 pm

I wonder why more of these correspondence players don't venture out otb to demonstrate their deep positional understanding...

MartinCarpenter
Posts: 2454
Joined: Tue May 24, 2011 10:58 am

Re: Using a computer in correspondence chess

Post by MartinCarpenter » Fri Jan 08, 2016 3:38 pm

Kevin O'Rourke wrote:i think i can wait a microsecond :-)
Multiply it up by all the positions it could be looking at and you can't! Well not every time. The tree of possible moves expands so fast they have to use some quite harsh pruning techniques to get the search to work well. Those are brilliantly tuned nowadays of course :)

I guess with OTB/correspondence chess, there's always been people much better suited to one or the other. The skill set must have got further apart with computers though.

Things like managing the sundry engines in their searches - out of the box they're set up to be optimised for very different sorts of searches/time rates so this isn't trivial - and knowing what sorts of positions its possible to beat someone who's defending using a computer are presumably crucial for correspondence and no use at all OTB. How much 'raw' human chess understanding there still is I'm not sure.

Mick Norris
Posts: 7595
Joined: Tue Apr 17, 2007 10:12 am
Location: Bolton, Greater Manchester
Contact:

Re: Using a computer in correspondence chess

Post by Mick Norris » Fri Jan 08, 2016 4:02 pm

Adam Ashton wrote:I wonder why more of these correspondence players don't venture out otb to demonstrate their deep positional understanding...
It might be nerves with some? Others may have family/work commitments that stop them getting to OTB matches/events, it can be a good way to keep an interest, and we have seen some players return OTB eventually
Any postings on here represent my personal views and should not be taken as representative of the Manchester Chess Federation www.manchesterchess.co.uk

Paul Dargan
Posts: 514
Joined: Sun May 13, 2007 11:23 pm

Re: Using a computer in correspondence chess

Post by Paul Dargan » Fri Jan 08, 2016 4:26 pm

I played correspondence chess when I was unable to play OTB - and used it to justify to myself keeping my Informator subscription going .., it was always just a means to an end for me though, I played as a substitute for playing proper chess. I checked the OTB grades for the people I was playing and I was playing 110's who never missed a tactic - it was clear that almost everyone I played (even back in mid-late 90's) had at least a blundercheck computer in use.

Fortunately (and probably wrongly) I felt that given enough time in evenings in hotel rooms I could still outplay most of the software back then ... Me at a rusty 190-ish was good enough back then to win the British CC title and qualify for the World Cup semis. Then I bought some software, but ended-up wondering what the point of it was if the moves weren't mine. That combined with work pressure that meant I was never at home to receive the cards (this was pre server days) meant I withdrew completely from correspondence (save the odd county game I feel obligated to play).

I'm sure GM+current software is a powerful combination, but mere mortals are probably conning themselves if they think they bring much to the partnership.

I do keep meaning to go over some of my correspondence games with today's engines to see how bad they were. There's a prticular piece sac I 'invented' in a line of the From gambit that I suspect they may not approve of ...

Paul

Tim Harding
Posts: 1799
Joined: Sat Oct 23, 2010 8:46 pm
Location: Dublin, Ireland
Contact:

Re: Using a computer in correspondence chess

Post by Tim Harding » Fri Jan 08, 2016 7:00 pm

I think some of the comments so far in this thread would be considered fairly ignorant by top ICCF players. Personally I don't play serious CC any more and for 3-4 years I did not play at all. Lately I have had some interesting games where the engines don't always find the best moves. Also some games I won ridiculously easily because the opponent either didn't use an engine or didn't know how to use them properly.

Certainly 2 seconds or even 30 is rarely enough time to get a correct evaluation and main line unless the position is so clear that an engine is hardly required. That might be sufficient time for spotting blunders but not for proper analysis.

You can expect top players are mostly using two engines to cross-check with their own ideas. Also players with true chess understanding (which generally means players who have reached at least 2200 OTB) can judge what engine feedback is relevant and what is rubbish, and in particular which positions should be searched deeply (overnight plus) and doing it sufficiently far ahead so it's possible to avoid that line altogether if the outcome is not satisfactory.

It is indeed very hard to win games against opponents of a similar standard who also use their engines properly. It requires either a significant advantage from the opening or outplaying them in an endgame before it reduces to a solved 7-man tablebase position. I have won three rook endgames in the past couple of years (two from equal or only slightly advantageous positions) and in each case (Rybka 3, Houdini 4, Stockfish 6) the engines gave unreliable results.

And in complex middle games you frequently get the 0.00 evaluations because the computer cuts from its search some lines that might be promising and instead shows repetitions and perpetuals. You have the initiative but keep seeing this score so you have to look for ways to keep the game going.

From an interview by Gino Figlio with FIDE GM Leitao in 2008: https://www.iccf.com/message?message=225 I extract:

GFSubmit Most people that criticize correspondence chess after the advent of chess engines suggest that understanding of the game may not improve for the average player by simply using the powerful calculation abilities of a chess engine. How do chess engines help OTB players and do you see any negative influence in their use?

Rafael Leitao: Certainly most people do not know how to use the chess engines in a correct way. Its use can surely be very harmful for chessplayers in general. I think it is very important that you create your own ideas and form opinions about the character of the position, only using the computer to check for obvious calculation mistakes. Also, most engines do not understand positions with material imbalance. Even with computers, one can never be lazy. Properly used, the engines can improve one´s play.

GFSubmit What are the main reasons why most top OTB players do not play correspondence chess?

Rafael Leitao: I believe most professional players do not have time to play correspondence chess. Sometimes the search for the best move can take days (and even then there is no guarantee you will find a proper answer). And I know many players who enjoy the sportive and psychological aspects of the game, that are most obvious on OTB fights. But I believe that correspondence chess can have a serious impact on improving the game of an OTB player.


GFSubmit The International Correspondence Chess Federation (ICCF) does not prohibit the use of chess engines. Are you planning or perhaps already using them to help you with your correspondence chess games?

Rafael Leitao: Yes, I use all chess engines available. The search for the best move, which is what I try when playing correspondence chess, should meet no boundaries, in my opinion. As I said, blindly following the computer suggestions will usually lead to disaster. In the end, the best players will win anyway. I should add, complementing my last answer, that one of the main positive points of correspondence chess nowadays is that it teaches you (sometimes in the hard way) how to use computers to improve your analysis.
Tim Harding
Historian and Kibitzer

Author of 'British Chess Literature to 1914', Joseph Henry Blackburne: A Chess Biography', and 'Eminent Victorian Chess Players'
http://www.chessmail.com

Post Reply