Computer Chess

[pianolearner]

I found this online version and it seems good. I managed to beat it on "Easy" but I have never been a great player. What is it that makes a chess grandmaster? Is it a great memory or  is it skill & strategy?

https://www.shredderchess.com/play-chess-online.html

[soliloquy]

Memorizing openings and end-games is like 50% of it.  The rest is a combination of practice and being a savant lol

[prometheus]

Shredder is a stong chess engine. You won't be able to beat it unless you hit a bug or if you are are IM or GM.

[ronde_des_sylphes]

The version on the website is not especially strong, but it's quite a nice interface. Sadly, the very best programs have now got to the stage where, if they are running on decent hardware, it is now debatable whether even the world's strongest players can defeat them (eg Fritz 4, Kramnik 2 in the recent match - even if one of the human losses was a gross one-move blunder).

To be a top grandmaster, you need encyclopaedic opening knowledge, chess intuition, deep strategical knowledge, excellent calculation and visualisation skillls and endgame understanding (a bit of genius probably helps too). It's not something many people are likely to acquire.  But on the bright side - the grandmaster title has been significantly devalued in the last 20 years and is now a lot easier to attain than it used to be.

[prometheus]

It is amazing that people have not been smart enough to program instructions for a computer to beat humans.


Just let it get over with that computers are better than humans. Then we can go back to actual playing chess.

[ronde_des_sylphes]

One of the problems with programming computers to play chess is tha "event horizon" which is an issue where, in certain types of position, the computer is trying to solve an issue which involves lengthy analysis. It can occur that, for example, the computer can only see 8 moves ahead by either side, but the real crux of the matter occurs 10 moves ahead. One of the reasons chess has been of such interest to computer scientists is because of the implications regarding brute force calculation v "human understanding".

Computers are now manifestly stronger than all bar the very best humans, but there are still clear imperfections in the programming which are not likely to be solved in the near future.   However, interestingly, programs are still in no danger of beating the best Go players. 

[pianolearner]

One of the problems with programming computers to play chess is tha "event horizon" which is an issue where, in certain types of position, the computer is trying to solve an issue which involves lengthy analysis. It can occur that, for example, the computer can only see 8 moves ahead by either side, but the real crux of the matter occurs 10 moves ahead. One of the reasons chess has been of such interest to computer scientists is because of the implications regarding brute force calculation v "human understanding".

Computers are now manifestly stronger than all bar the very best humans, but there are still clear imperfections in the programming which are not likely to be solved in the near future.   However, interestingly, programs are still in no danger of beating the best Go players. 

This is interesting:

https://www.research.ibm.com/deepblue/meet/html/d.shtml

[ronde_des_sylphes]

Yes, there was huge controversy over that match. In one game, Kasparov infamously accused the Deep Blue team (which was assisted by several very strong American players) of having had human help. (He alleged that a particular move which the computer found was not a move that a computer could have found - and as he is probably the strongest player in history, his opinion should count for something). As to whether it was sour grapes, it is hard to tell - but the incident was not helped by IBM's refusal to allow access to the game logs.

[prometheus]

Kasparov may have been the strongest GM in history. He is not a computer expert, right? I don't know if he can judge accurately the difference between a 'human move' and a 'computer move'.

[counterpoint]

Kasparov may have been the strongest GM in history. He is not a computer expert, right? I don't know if he can judge accurately the difference between a 'human move' and a 'computer move'.

That's exactly right!

A computer program will find any possible moves in the first second - and then, it will decide, which one of these possible moves is the best one. That can take a very long time, since all consequences of all possible moves have to be looked at.

But there is nothing like a move, a computer program can not find.

[prometheus]

Actually, there are Chess puzzles that are presently unsolveble by computers.

[dazed]

De-Groot did an experiment with chess 'players' that has always semed to me relevant to sight reading/memorising music...  He showed GMs, IMs, club players, occasional players and non-players (or something like that)  boards from real games for N seconds and asked them to rememeber the positions of all the pieces.  The results followed the list above because the more advanced players were remembering more structure (less data) and the novices absolute positions (more data).  When shown boards with the pieces randomly scattered , everybody scored the same kindasortaish...     

[soliloquy]

Actually, you can beat Deep Shredder consistently by playing from white and running a Sozin Attack whenever it opts for a Najdorf, which is Shredder's opening of choice from black.  Deep Shredder will also OCCASIONALLY run a Modern Steinetz; its mid game from Steinetz Defense is sub-par at best, and you can squeeze out a win here and there from that.  Fischer's Poisoned Pawn also usually will give you a huge advantage early in the game whenever you get a chance to play it; shredder also sometimes makes very amateur mistakes against Centre/Counter.

[ronde_des_sylphes]

De Groot's experiment was very interesting. The one-time world champion Max Euwe took part in this and, on the test position that I have seen the results for, he not only recalled the position almost perfectly on seeing it for a very short time, he also found the correct next move for White. What was noteworthy was how all the players, no matter what strength, produced the same low level of recall when confronted with what was, in effect, a nonsense position with the pieces randomly scattered, suggesting that pattern recognition plays a significant part for strong players.

  Fischer's Poisoned Pawn also usually will give you a huge advantage early in the game whenever you get a chance to play it

lol.. I would be very surprised: opening theory on the Poisoned Pawn runs very deep. You must have stolen Radjabov's notebooks

[dazed]

What was noteworthy was how all the players, no matter what strength, produced the same low level of recall when confronted with what was, in effect, a nonsense position with the pieces randomly scattered, suggesting that pattern recognition plays a significant part for strong players.

Thus atonal music is more difficult to sightread than (say) Chopin, because of fewer familar structures...  or is that a generalisation too far...

Chess story: on day one of a new job I suggested to my office-mate that we set up a chessboard on the filing capinet and move when we feel like it.  Fine.  This we did and while at the time I thought I was an OK player, I got anhiliated  time after time.  I was moaning (good naturedly) about this to another collegue a month or so later and he said "ah, well that's because Mark's a grandmaster and you're not" - I laughed - good joke - and carried on getting stuffed for a few more weeks; it never occured to me it may be true.  Mark (Orr) eventually fessed up - not a GM but an IM and at the time Irish Chess Champion.  I think I beat him once - and that is my chess claim to fame 

[ronde_des_sylphes]

Chess story: on day one of a new job I suggested to my office-mate that we set up a chessboard on the filing capinet and move when we feel like it.  Fine.  This we did and while at the time I thought I was an OK player, I got anhiliated  time after time.  I was moaning (good naturedly) about this to another collegue a month or so later and he said "ah, well that's because Mark's a grandmaster and you're not" - I laughed - good joke - and carried on getting stuffed for a few more weeks; it never occured to me it may be true.  Mark (Orr) eventually fessed up - not a GM but an IM and at the time Irish Chess Champion.  I think I beat him once - and that is my chess claim to fame 

That's a funny story: I happen to know Mark Orr, not well, but I've had contact with him on occasion - he used to be a regular on the Scottish tournament scene and played for a club in Edinburgh at one point.

[dazed]

That's a funny story: I happen to know Mark Orr, not well, but I've had contact with him on occasion - he used to be a regular on the Scottish tournament scene and played for a club in Edinburgh at one point.

Nice guy.  He'd come back to the house for food and after the wife and kids had gone to bed we'd watch a colourful film, drink lots of booze and and play some chess.  He sat on the floor without a board and I'd sit on the sofa in front of one.  Even after feeding him lots of alcohol I still lost.  He said he pictured the board in two halfs in his head and worked on each independently - I tried - absolutely hopeless.  I've never been sure if that abilty only comes with chess expertise or is something more generic...   

[ronde_des_sylphes]

Nice guy.  He'd come back to the house for food and after the wife and kids had gone to bed we'd watch a colourful film, drink lots of booze and and play some chess.  He sat on the floor without a board and I'd sit on the sofa in front of one.  Even after feeding him lots of alcohol I still lost.  He said he pictured the board in two halfs in his head and worked on each independently - I tried - absolutely hopeless.  I've never been sure if that abilty only comes with chess expertise or is something more generic...   

Ah, sounds like "blindfold chess" - which was banned in Stalinist Russia due to the alleged mental stress it causes. All master-level players should be able to do it, but what really impresses me are such feats as then world champion Alekhine's blindfold exhibition where he played thirty-two simultaneous games of blindfold chess. I suspect the ability to do such things is a combination of chess ablity, excellent memory and visualisation skills. I can do it with one game at a time (and with practice, maybe two or three) but Alekhine's example is in another world altogether.

[dazed]

Ah, sounds like "blindfold chess" - which was banned in Stalinist Russia due to the alleged mental stress it causes.

I'll google for more information; sounds fascinating.

I can do it with one game at a time (and with practice, maybe two or three)...

Very impressive - all I see is a blurry haze with some indistinct blobs floating around; mind you my eyesight is -8 so...