Anyone play chess?

[georgey]

Hi Pianostreet members.  Just wondering if there are any chess enthusiasts here.  I might be posting in the wrong site but there are chess grandmasters that were world class pianists (e.g. Taimanov who was USSR world champion chess player candidate and is listed as one of the great pianists of the 20th century).  French composer Philidor was the greatest chess player of the 1700's. Although my first love is music, I view chess to be an equal to classical piano as a pursuit.   I would say there are as many people passionate about chess as classical piano, maybe more.  Chess.com has I think 15 million members with 30k on line at any time.

[lostinidlewonder]

Yes I do I like bullet chess 1minute per side a lot of fun, standard times can be all too serious sometimes! I used to run a chess club in my area and held a few competitions with prize money for fun. The state champ visited our club and I beat him haha I never done chess seriously though in Australia it is quite unpopular. I'm out of practice now but used to beat players around 2000-2200 during my highschool years. Endgame theory is what improved my game the most.

[georgey]

Yes I do I like bullet chess 1minute per side a lot of fun, standard times can be all too serious sometimes! I used to run a chess club in my area and held a few competitions with prize money for fun. The state champ visited our club and I beat him haha I never done chess seriously though in Australia it is quite unpopular. I'm out of practice now but used to beat players around 2000-2200 during my highschool years. Endgame theory is what improved my game the most.

Wow!  I feel like I hit the jackpot!  Here is my background:  In 1990 I spent 80 hours learning chess rules and basic strategy just for the purpose of writing a chess computer program to increase my programming skills.  I then spent about 1000 hours in 1990-1991 writing from scratch a program in Turbo Pascal that plays at about 1000 Elo.  I played my program against ed’s chess (a free piece of software available in 1991) and lost the first 4 games but drew the 5th game. (I set Ed’s chess to only think 1 minute per move compared to my 5 minutes – i.e., I cheated!).  I was happy anyway when it offered my program a draw on game 5 lol!  That was it until about 10 weeks ago when I bought 22 chess books and Komodo 10 software with Fritz 15 user interface.  I can’t believe how powerful chess computers are now!  My goal is to get to 2000 Elo in next 2 to 4 years as a player (no more programming).  I’m 58 years old so my learning is going to be slower than a young person.

I agree with you completely on end game importance.  From what I read, some good players are very bad at end games. One of the books I bought is Silman’s complete endgame course.  It is great in that it has chapters starting from Elo 1000, then Elo 1200, etc, up to Elo 2399.  In the next year I plan to go through the first 220 pages of this book that will take me up to 1799 Elo plus go through about 13 other books and play games against Komodo 10.  Then I will join the USCF and compete while continuing to study.

[eldergeek]

I retired a few years ago and took up piano as a retirement hobby, having played classical guitar for many years. In my younger days I played chess quite seriously, although never professionally - in the UK I used to play for various universities and various county chess teams.

Eventually, I quit chess for a number of reasons, the main ones being (a) the amount of time it takes to keep up-to-date with modern opening theory, and (b) I stopped enjoying the stress involved in playing a match that lasts 4-5 hours, especially when the team result depends on you not making an error.

I would back 100% the idea that the endgame is the most important aspect of the game to study, and I believe that every single world champion there has ever been was a great endgame player.

Somewhat related to the thread on mental practice:

At one stage in my chess development, I created a lot of exercises for improving the ability to visualise future positions on a chessboard without moving the pieces (clearly a vital skill for normal play) by utilising endgame studies composed by a little-known bunch of chess enthusiasts who dedicated a big part of their lives. Some of them are really rather beautiful and they really do helpwith the ability to calculate chess variations in your head while playing. I must admit to having tried to do the same sort of thing with mental practice with the piano, but not really had much success ...yet.

If you would be interested in some of my endgame study examples for chess improvement, feel free to let me know - maybe by pm or email.

[Bob]

Did a little a long time ago.  Probably not very good.

[ronde_des_sylphes]

Chess was probably my first love and piano came second. I don't play competitively nowadays: I got to a level by the time I was 18 where I was good enough to beat titled players on occasion (there is a game of mine on chessgames.com where I defeat a former UK champion) but not good enough to be a serious contender in important events, and I got frustrated and stopped improving significantly. Still, I got to play in junior international events and retain fond memories of this.

[georgey]

Eldergeek: You and I have some amazing similarities.  I also retired recently (1.5 years ago) and also took up piano as a hobby in retirement.  Unfortunately I developed osteoarthritis of the hands and decided 10 weeks ago to take up chess instead.  I also played classical guitar for many years.  I attended in 1982-1983 the University of Toronto for a Master’s degree in classical guitar performance and studied with Eli Kassner and Norbert Kraft there.  Before finishing I decided to go back to my college in USA to get a math degree.  I can understand the team pressure you talk about for chess.  Nice thing with me is I am just doing chess for fun and it is fun!  As you say, chess is a game of visualizing.  The person that can visualize the farthest and the most accurately will have a huge advantage in chess.  Sort of like racing at night on a curvy track with obstacles.  The one with the best headlights to see in the distance will have a huge advantage.

Bob:  Hope all is well.  Glad you had the chance to play chess. 

Ronde: I hope all is well.  I am amazed that chess may have been your first love over piano!  I tried to find your game on chessgames.com.  When I search for Andrew Wright it shows games for an Andrew Greet.  I must be doing something wrong lol.  If you post a link I would love to look at the game.  Studying games is a great way to learn.  I checked amazon.com today to see if your 2nd CD is available.  I don’t see it listed.  Any idea when it will be available?  Thanks.

[eldergeek]

Georgey: we have more similarities than you think. My degrees are also in mathematics

[lostinidlewonder]

Wow!  I feel like I hit the jackpot!  Here is my background:  In 1990 I spent 80 hours learning chess rules and basic strategy just for the purpose of writing a chess computer program to increase my programming skills.  I then spent about 1000 hours in 1990-1991 writing from scratch a program in Turbo Pascal that plays at about 1000 Elo.  I played my program against ed’s chess (a free piece of software available in 1991) and lost the first 4 games but drew the 5th game. (I set Ed’s chess to only think 1 minute per move compared to my 5 minutes – i.e., I cheated!).  I was happy anyway when it offered my program a draw on game 5 lol!  That was it until about 10 weeks ago when I bought 22 chess books and Komodo 10 software with Fritz 15 user interface.  I can’t believe how powerful chess computers are now!  

Wow that is incredible, writing a program that plays chess well is no easy feat. Nowadays there are insanely difficult programs out there no human in the long run can beat them these days without some kind of odds! Still I like that chess has not been solved like Checkers (draughts) has, apparently that game is a draw if both sides play perfect.

https://www.sciencebuzz.org/blog/checkers_is_solved_on_to_the_next_problem

I believe at the moment Komodo is the strongest engine, but I really like Fritz and all its analysis. I used to use Rybka a lot it was one of my favorites still is even after that big controversy but its not updated no more Anti computer openings is my favorite approach vs computers, closed position, back in the 1980s-2000s I could beat any computer with the hippopotamus opening which pretty much befuddled them all and you could slowly build an attack on whatever side it castled.

My goal is to get to 2000 Elo in next 2 to 4 years as a player (no more programming).  I’m 58 years old so my learning is going to be slower than a young person.

A realistic goal I think is possible with some hard work! Once you have the general gist of opening theory and middle game put all energy to endgame and your level will increase a huge amount. Endgame books I liked was Laszlo Polgars (father of the famous Polgar sisters) Endgame, was so damn thick and literally would take a lifetime to fully appreciate! Endgame Secrets by Lutz is also exceptionally instructive, smaller book but will take years to fully appreciate. I also loved to study lots of mate puzzles or other combinations, this helps sharpen tactics a great deal, there is another book from Laszlo Polgar, 5334 Problems Combinations and Games which where written in neatly organized themes.

I agree with you completely on end game importance.  From what I read, some good players are very bad at end games.

I wouldn't say good players are very bad at it but certainly it is within the endgame where you must beat most of the knowledgeable players. The opening and middle game usually you beat people who are less experienced, or who you've book bash to death aahaha ^__^ I like blitz/lightning so that the game isn't so complicated and the win doesn't occur so often in the endgame, its nice to also see the tables turn round! Fast chess is though quite bad for your overall playing skill level, it does help with instincts and calculation speed but for deep thought you need for standard games of course its a bad idea.



This is something I posted on pianostreet a while back Ill copy paste it here, it was a game I played which I was proud of ahaha:

I was going through my old files and found one of my old favorite games. Here is a position where a real pretty mating combination ends with an amusing fianchettoed bishop mate Smiley The game was played 5 minutes a side, there are blunders but black made some strange moves in the opening which I exploited and thus my pieces where just so much more active.

1: d4  e6
2: e4  d5
3: e5  c5
4: c3  Nc6
5: Nf3  Nge7
6: Be2  Nf5
7: 0-0  Qb6
8: dxc5  Bxc5
9: b4  Bf8
10: Bd3 Nfe7
11: b5  Nb8
12: Be3  Qc7
13: Qa4  b6
14: Nd2  Nd7
15: Rac1 Ne5
16: Nxe5  Qxe5
17: Bxb6  Bb7
18: Nf3  Qh5
19: Be3  Nf5
20: b6+ Kd8
21: Ne5  Ke7
22: Bg5+  Qxg5
23: Qd7+ Kf6
24: Qxf7+ Kxe5
25: Rfe1+ Kf4
26: g3+ Kg4
27: Be2+ Kh3
28: Bf1+ Kg4
29: h3+ Kf3
30: Bg2++       1-0

The funny thing about the position is that if black had his pawn on g6 (instead of g7) this mating combination would have not worked because then whites Queen would not be able to cut off the Black king on the h5 square (after white plays 29 h3+). This is one of the best mating combinations I ever played.

[ronde_des_sylphes]

Ronde: I hope all is well.  I am amazed that chess may have been your first love over piano!  I tried to find your game on chessgames.com.  When I search for Andrew Wright it shows games for an Andrew Greet.  I must be doing something wrong lol.  If you post a link I would love to look at the game.  Studying games is a great way to learn.  I checked amazon.com today to see if your 2nd CD is available.  I don’t see it listed.  Any idea when it will be available?  Thanks.

Andrew Greet is an English international master or grandmaster, offhand I don't recall which.

https://www.chessgames.com/perl/chessgame?gid=1381867 is the game I had in mind. I was just a kid here. My opening play didn't quite work and my opponent equalised fairly easily: if he had recaptured on c6 with the pawn I think he was better. An exciting opposite side castling attack race followed.

My cd is very much in the pipeline but I'm not releasing it independently; it will be coming out on the Divine Art label and I anticipate their release schedule will imply the start of April or May. They have the master tracks and contracts have been signed but they have a lot of projects for release and there is a queue!

As for chess books I liked - Kotov's Think like a Grandmaster and the best game collections of Alekhine, Tal and Fischer I recall with particular affection. I agree with LIW about blitz and bullet chess: very good for training tactical relexes but can be insidious in terms of also conditioning you to think in clichés. There are good tactics trainers available online as an alternative and some decent online commentary channels on youtube (eg Matojelic, Will Stewart - both get the essentials over without being verbose - they are suitable for the less experienced player). LIW - looks like a nice combination. I don't have a board in front of me but 5. Nge7 and 10. Ne7 presumably, Nd7 being a typo.

[lostinidlewonder]

...LIW - looks like a nice combination. I don't have a board in front of me but 5. Nge7 and 10. Ne7 presumably, Nd7 being a typo.

Gah you're right how did that happen ahaha. Thanks Ill change it.

[georgey]

Wow that is incredible, writing a program that plays chess well is no easy feat. Nowadays there are insanely difficult programs out there no human in the long run can beat them these days without some kind of odds! Still I like that chess has not been solved like Checkers (draughts) has, apparently that game is a draw if both sides play perfect.

I want to go through the games given here and other comments, but this will take a day or so.  I am a beginner in chess.  I’m going through Ronde’s game right now.  My head is spinning lol. I just wanted to comment on the above for now (for those interested. if any.  This can be boring!  ).

I had only 1 programming course in college and it was in FORTRAN.  I learned Pascal for the chess project.   Turbo Pascal was an amazing compiler with incredible debugging features back in 1990-1991.  It allowed me to step through the program 1 line at a time (even backwards lol) and see how all the variables changed 1 line at a time.  I have never done any research on writing chess programs but I knew that I was going to do examine every single possible legal move through exhaustive search. 

The biggest challenge I had was having the computer move after 3 minutes average thinking time.  Computers were very slow back then and I was only able to exhaustively go out 4 ply with a short cut to simulate 5 ply.  At the start of the game, the computer would look at the following sequence of moves.  1. a3 a6, 2.  a4 a5 3.b3 .  As silly as this was, it was 1 of about 3 to 6 million moves that had to be looked at.  To pick out the relevant sequences of moves from the garbage sequences, I pondered for 2 weeks to come up with the “min of the minmax” solution.  I googled this just now and see that this part of game theory.  I sort of reinvented the wheel.  But this is something that anyone with a little math background and an intuition could do. 

Say the computer is looking at all of the 30 legal moves that it can play next.  It finds that 4 of the 30 moves have the min of the minmax equal to 3 points (that is the computer calculates it can gain a piece with each of the 4 moves, 1 point is a pawn) and the other 26 moves result in scores less than 3.  The computer would then do a post-analysis of the 4 moves to see which would be best for the computer to do.  The post analysis would look at the number of squares controlled by the computer less the number of squares controlled by the opponent (with extra weighting given to the central squares).  It would also look at pawn structure and a few other basics.

I always felt that if I had a computer that was many trillion times faster than the fastest computer of today, I OR ANYONE could write a program to play the perfect game and solve chess once and for all.  If 2 perfect computers played each other, would chess always end in a draw?  Or does white have the advantage and could always win?

[eldergeek]

With regard to your last question - if two "perfect" players played against each other, would the game always end in a draw:

I believe the answer to this must be yes, for the following reason(s).

Over the last few decades, the percentage of draws between grandmasters has increased (or so I believe - I do not have precise statistical data here!), as has the ELO grades of the top few hundred players in the world. It seems most probable that as humans (with the assistance of computer analysis for opening theory) get closer and closer to the "perfect player", then the percentage of draws will increase, sort of implying (for a professional mathematician, I know I am being pretty vague at this point!) that in the limit, two perfect players will always draw against each other.

[eldergeek]

More ramblings, this timeabout the differences between humans and machines playing chess:

As I understand things, with my somewhat limited knowledge of computer chess playing algorithms, it seems that all chess algorithms more or less rely on exhaustive searches, involving examining all possible moves in a given position, and then examining all possible replies to those moves. I believe that modern algorithms weed out quite a lot of the "tree" by deciding that many moves do not deserve much analysis, and can be discarded, thus pruning the tree of analysis sufficiently that what remains can be analysed more thoroughly in the time available.

In contrast, an expert human looking at a normal chess position (ignoring the extremely implausible positions which arise in "mate in 2" type problems) will almost immediately narrow the search down to a small handful of moves which need to be looked at carefully - often, only 3 or 4 moves really deserve deeper analysis.

So, one of the most significant differences between amateur and professional players is the speed with which the professional can narrow the search down to a very small number of variations which (s)he can then examine in detail, and this is a skill that can be learned with thousands of hours of practice on standard tactical processes. I believe Carlsen got to his remarkable level by doing just this, aided by playing thousands of games online each year.

In a way, this process seems to be similar to the ability to sight-read piano pieces - the more hours (thousands of hours?) you put in, the more you come to recognise at a glance, the frequent chord sequences and melodic sequences. I.e. a good sight reader does not look at each note individually, but see blocks of notes at a single glance, and plays those almost on auto-pilot while looking at what comes next ...

[timothy42b]

In contrast, an expert human looking at a normal chess position (ignoring the extremely implausible positions which arise in "mate in 2" type problems) will almost immediately narrow the search down to a small handful of moves which need to be looked at carefully - often, only 3 or 4 moves really deserve deeper analysis.

Chess masters don't analyze all available moves and the consequences.  I don't know what computer programs do. 

Chess masters have stored memories of positions, spatial arrangements of pieces that they know are favorable.  They eventually have many thousands of these stored, apparently. 

What they do is manipulate play to attempt to force the current board into a memorized position.  Position gives them an advantage of strength that eventually can force an advantage of pieces.  At their level (avoiding mistakes) a very small advantage can yield a win, provided time allows. 

That may relate to sightreading, in which recall of stored patterns is far faster than prima facie reading. 

I lost to a grandmaster in a simul once, it was quite enlightening.  And amazing, I was thoroughly schooled in a way I hadn't expected.  Nothing I did made any difference, and he was playing 25 of us, so time was on our side.  Guess I shouldn't have played Ruy Lopez. 

[georgey]

With regard to your last question - if two "perfect" players played against each other, would the game always end in a draw:

I believe the answer to this must be yes, for the following reason(s).

Over the last few decades, the percentage of draws between grandmasters has increased (or so I believe - I do not have precise statistical data here!), as has the ELO grades of the top few hundred players in the world. It seems most probable that as humans (with the assistance of computer analysis for opening theory) get closer and closer to the "perfect player", then the percentage of draws will increase, sort of implying (for a professional mathematician, I know I am being pretty vague at this point!) that in the limit, two perfect players will always draw against each other.

I think ultimately it will be a mathematician that proves this.  Computers may never be fast enough to answer the question by brute force.  My intuition agrees with yours but as mathematicians, we know intuitions cannot be relied on.  It may be intuitive to some that there are twice as many points on the number line on [0,2] as there are on [0,1] but we can prove very easily that these 2 closed intervals have the same number of points.  The harmonic series 1+ 1/2 + 1/3 + 1/4 + ....  may appear to converge by intuition to some but we can easily prove that it diverges. 

[georgey]

More ramblings, this timeabout the differences between humans and machines playing chess:

As I understand things, with my somewhat limited knowledge of computer chess playing algorithms, it seems that all chess algorithms more or less rely on exhaustive searches, involving examining all possible moves in a given position, and then examining all possible replies to those moves. I believe that modern algorithms weed out quite a lot of the "tree" by deciding that many moves do not deserve much analysis, and can be discarded, thus pruning the tree of analysis sufficiently that what remains can be analysed more thoroughly in the time available.

In contrast, an expert human looking at a normal chess position (ignoring the extremely implausible positions which arise in "mate in 2" type problems) will almost immediately narrow the search down to a small handful of moves which need to be looked at carefully - often, only 3 or 4 moves really deserve deeper analysis.

So, one of the most significant differences between amateur and professional players is the speed with which the professional can narrow the search down to a very small number of variations which (s)he can then examine in detail, and this is a skill that can be learned with thousands of hours of practice on standard tactical processes. I believe Carlsen got to his remarkable level by doing just this, aided by playing thousands of games online each year.

In a way, this process seems to be similar to the ability to sight-read piano pieces - the more hours (thousands of hours?) you put in, the more you come to recognise at a glance, the frequent chord sequences and melodic sequences. I.e. a good sight reader does not look at each note individually, but see blocks of notes at a single glance, and plays those almost on auto-pilot while looking at what comes next ...

I agree with everything you say here.  I did check Wikipedia on computer chess programming and it appears you are correct on "I believe that modern algorithms weed out quite a lot of the "tree" by deciding that many moves do not deserve much analysis, and can be discarded, thus pruning the tree of analysis sufficiently that what remains can be analysed more thoroughly in the time available.".  Thanks.

[lostinidlewonder]

I had only 1 programming course in college and it was in FORTRAN.  I learned Pascal for the chess project.   Turbo Pascal was an amazing compiler with incredible debugging features back in 1990-1991.  It allowed me to step through the program 1 line at a time (even backwards lol) and see how all the variables changed 1 line at a time.  I have never done any research on writing chess programs but I knew that I was going to do examine every single possible legal move through exhaustive search.  

The biggest challenge I had was having the computer move after 3 minutes average thinking time.  Computers were very slow back then and I was only able to exhaustively go out 4 ply with a short cut to simulate 5 ply.  At the start of the game, the computer would look at the following sequence of moves.  1. a3 a6, 2.  a4 a5 3.b3 .  As silly as this was, it was 1 of about 3 to 6 million moves that had to be looked at.  To pick out the relevant sequences of moves from the garbage sequences, I pondered for 2 weeks to come up with the “min of the minmax” solution.  I googled this just now and see that this part of game theory.  I sort of reinvented the wheel.  But this is something that anyone with a little math background and an intuition could do.  

Say the computer is looking at all of the 30 legal moves that it can play next.  It finds that 4 of the 30 moves have the min of the minmax equal to 3 points (that is the computer calculates it can gain a piece with each of the 4 moves, 1 point is a pawn) and the other 26 moves result in scores less than 3.  The computer would then do a post-analysis of the 4 moves to see which would be best for the computer to do.  The post analysis would look at the number of squares controlled by the computer less the number of squares controlled by the opponent (with extra weighting given to the central squares).  It would also look at pawn structure and a few other basics.

Was interesting to read! Nowadays very strong computers use endgame tablebases which makes calculation of deep endings unnecessary, this has made beating computers next to impossible as it was in the ending where we could catch out the best computers. The huge database for opening theory can also be stored so then they are only calculating when out of book or outside of tablebase positions. The strength of their opening moves needs to be carefully considered by very strong human players, a computer is a slave to whatever opening you choose it to take, a lesser programmer will not use current chess expertise in opening theory and probably make its program choose lines which can be exploited with accurate play.

I always felt that if I had a computer that was many trillion times faster than the fastest computer of today, I OR ANYONE could write a program to play the perfect game and solve chess once and for all.  If 2 perfect computers played each other, would chess always end in a draw?  Or does white have the advantage and could always win?

I have a feeling if chess is played perfectly white should win, the extra half tempi must be an edge, also statistically it is white which wins more often than black when masters play. I wonder what it is with checkers, if the 1st player wins more often or 2nd?

It is impossible to calculate all moves in chess with our current technology, its a stupid large number of possibilities.

[georgey]

Was interesting to read! Nowadays very strong computers use endgame tablebases which makes calculation of deep endings unnecessary, this has made beating computers next to impossible as it was in the ending where we could catch out the best computers. The huge database for opening theory can also be stored so then they are only calculating when out of book or outside of tablebase positions. The strength of their opening moves needs to be carefully considered by very strong human players, a computer is a slave to whatever opening you choose it to take, a lesser programmer will not use current chess expertise in opening theory and probably make its program choose lines which can be exploited with accurate play.

I have a feeling if chess is played perfectly white should win, the extra half tempi must be an edge, also statistically it is white which wins more often than black when masters play. I wonder what it is with checkers, if the 1st player wins more often or 2nd?

It is impossible to calculate all moves in chess with our current technology, its a stupid large number of possibilities.

When I watched the 2016 world championship a couple weeks ago, one of the commentators I think said ALL possible 6 piece board positions (or was it 7??) are in a table base to show mate or draw.  My program was terrible at end games! That is why I was so glad when Ed’s chess offered my 5th game a draw.  That’s when I copied the source code on a 5.25” floppy and threw it in a box.  I think I might still have it.  Ed did not know how bad my program was at end games or it would have kept playing lol.  I did put in a modest opening book though.

If I could exhaustively calculate 150 ply using minmax only, I would need no opening or end game knowledge, just brute force calcs that anyone could program.  I think the number of theoretical chess positions is maybe something like 10^80 so this will never be possible.

Checkers??  What’s that?  You may be correct about chess.  My guess is one of the following would happen if 2 perfect computers played.  1)  White would win 100% of the time, OR 2) A draw would be reached 100% of the time, OR 3) If an infinite # of games were played, white would win X%, black would win Y% and a draw would be reached Z% of the time such that X + Y + Z = 100%  

[eldergeek]

Checkers??  What’s that?  You may be correct about chess.  My guess is one of the following would happen if 2 perfect computers played.  1)  White would win 100% of the time, OR 2) A draw would be reached 100% of the time, OR 3) If an infinite # of games were played, white would win X%, black would win Y% and a draw would be reached Z% of the time such that X + Y + Z = 100%

I hadn't ever considered that last possibility with X + Y + Z = 100%, but I believe that possibility does not actually arise between "perfect" players, for the following reason:

If we look at the 6 piece tablebases, they show that for any starting position with 6 pieces on the board, each position leads to either a win, loss or draw (100% of the time). Now consider that perfect play on both sides is equivalent to both players following a 32-piece endgame tablebase ... which includes the starting position of the whole game

[ronde_des_sylphes]

With regard to your last question - if two "perfect" players played against each other, would the game always end in a draw:

I believe the answer to this must be yes, for the following reason(s).

Over the last few decades, the percentage of draws between grandmasters has increased (or so I believe - I do not have precise statistical data here!), as has the ELO grades of the top few hundred players in the world. It seems most probable that as humans (with the assistance of computer analysis for opening theory) get closer and closer to the "perfect player", then the percentage of draws will increase, sort of implying (for a professional mathematician, I know I am being pretty vague at this point!) that in the limit, two perfect players will always draw against each other.

Yes, I'm certain that perfect play results in a draw. To a certain extent you can approach simulating this by examining games of top GMs; there is a very high percentage of draws and it is very rare indeed that computer analysis of decisive games  (which I hypothesise are by definition imperfect) does not pick up clear-cut error(s).

[ronde_des_sylphes]

Chess masters don't analyze all available moves and the consequences.  I don't know what computer programs do. 

Chess masters have stored memories of positions, spatial arrangements of pieces that they know are favorable.  They eventually have many thousands of these stored, apparently. 

Not fully true. What they have in mind are the positional characteristics of certain typical structures and an understanding of what strategies usually apply given these structures. For example,  in your quoted Ruy Lopez example, the d5 and f5 squares are often strategically important and white will try to control them. Similarly an a4 counterthrust against a black queenside pawn structure of a6; b5 and c5 can be a standard motif.

[ronde_des_sylphes]

Was interesting to read! Nowadays very strong computers use endgame tablebases which makes calculation of deep endings unnecessary, this has made beating computers next to impossible as it was in the ending where we could catch out the best computers. The huge database for opening theory can also be stored so then they are only calculating when out of book or outside of tablebase positions. The strength of their opening moves needs to be carefully considered by very strong human players(...)

I agree with the vast majority of this. It's worth pointing out the tablebases facilitate calculation of continuations which resolve into six or seven piece positions - thus a programme can evaluate instantly whether a couple of exchanges leading into a Q + P v Q ending, for example, constitute a winning or drawing continuation and then play the result optimally, something which is excruciatingly difficult for humans to do precisely. (Likewise, masters will probably know the "Troitzky line" and its application in 2N v P, a notoriously difficult ending to play well - even for grandmasters - but a computer can play it with total precision given tablebases.) I'm not convinced endgames were the best chance for humans to beat computers - I think it's more a matter of positions where the "event horizon" is beyond the search ply depth of the computer. One of Kasparov's victories against Deep Blue [edit: mea culpa - it was v Fritz in a four game match] occurred in such a position - a closed position where long-term strategic planning and "understanding" were of prime importance and dynamic considerations were minimal. Opening theory is imo now, at the highest level, developed by strong humans in tandem with computer assistance.

[georgey]

Yes, I'm certain that perfect play results in a draw. To a certain extent you can approach simulating this by examining games of top GMs; there is a very high percentage of draws and it is very rare indeed that computer analysis of decisive games  (which I hypothesise are by definition imperfect) does not pick up clear-cut error(s).

I agree also with draw as seeming most likely, although it would need to be proven with mathematical certainty for me to be sure.  It would be strange if it could be proven that white would always win.  Say you had such a perfect computer and were playing white.  At the start of the game before any moves are made the computer says “mate for white in 142 moves”.  You make the first move 1. d4 then the computer says “mate for white in 187 moves” showing that d4 is not the best first move, but it was not a blunder.  

I’m still working on your game.  It was beautifully played!  I may have a few questions later today.

[georgey]

https://www.chessgames.com/perl/chessgame?gid=1381867 is the game I had in mind. I was just a kid here. My opening play didn't quite work and my opponent equalised fairly easily: if he had recaptured on c6 with the pawn I think he was better. An exciting opposite side castling attack race followed.

My cd is very much in the pipeline but I'm not releasing it independently; it will be coming out on the Divine Art label and I anticipate their release schedule will imply the start of April or May. They have the master tracks and contracts have been signed but they have a lot of projects for release and there is a queue!

Great game!!  I can’t believe how well you played.  One thing I decided to add to my program of study was analysis of master or GM games like I did of your game.  This is the first time I tried this. After EACH move of your game, I thought to myself: “Where would I move next?”.  After deciding, I had the computer show me if it was a good move or not.  If it was bad, the computer showed me how by kicking my **** within a few moves or so.  I then had the computer calc its best move and I also compared the computer move with you or your opponent’s move.  Anyway I learned a lot from this 1 game!  Here are a couple things I found in this game.  

The computer did agree with your opponent’s move of  15. … Qxc6 instead of 15. … bxc6 leaving an isolated pawn.  Also, both sides played great but the game came down mostly to a single blunder made by your opponent:  30.  ... Bc4??  The correct move was 30. ... Rxg6.  You would still have been ahead by 0.9 pawns had the correct move been made, but after your opponent made his move the computer had you ahead 16 pawns.

Thanks for the update on your CD.  This helps.  I will buy my copy on the Devine Art Label website when available.  It looks like they are set up to take USA customers.

EDIT: One other item in your game.  Great kingside attack!  The computer did show a way to slow the attach down by 22. ... Kf7 instead of 22. ... Be6 made by your opponent.  I thought Be6 was a good move at first because it connected his rooks and blockaded your e5 pawn and attacked the a2 pawn.  He did not need to worry about this bishop being trapped if he took your a2 pawn since the c2 pawn was pinned.  Unfortunately the move did not prevent 23. Qh6!

[georgey]

I'm not convinced endgames were the best chance for humans to beat computers - I think it's more a matter of positions where the "event horizon" is beyond the search ply depth of the computer. One of Kasparov's victories against Deep Blue occurred in such a position - a closed position where long-term strategic planning and "understanding" were of prime importance and dynamic considerations were minimal.

Here is a link to the 2nd game of Kasparaov and Deep blue rematch in 1997.  I thought this was a great discussion as a beginner player.  Minute 33:25 of the video shows a move by deep blue that Kasparov was convinced was aided by humans.  Kasparov was black in the game. Beep blue moved his bishop to e4 to blockade black’s e5 pawn and so not allowing black to launch an attack into white’s territory.  Computers were starting to play great in 1997.

[georgey]

Wow that is incredible, writing a program that plays chess well is no easy feat. Nowadays there are insanely difficult programs out there no human in the long run can beat them these days without some kind of odds! Still I like that chess has not been solved like Checkers (draughts) has, apparently that game is a draw if both sides play perfect.

https://www.sciencebuzz.org/blog/checkers_is_solved_on_to_the_next_problem

I believe at the moment Komodo is the strongest engine, but I really like Fritz and all its analysis. I used to use Rybka a lot it was one of my favorites still is even after that big controversy but its not updated no more Anti computer openings is my favorite approach vs computers, closed position, back in the 1980s-2000s I could beat any computer with the hippopotamus opening which pretty much befuddled them all and you could slowly build an attack on whatever side it castled.

A realistic goal I think is possible with some hard work! Once you have the general gist of opening theory and middle game put all energy to endgame and your level will increase a huge amount. Endgame books I liked was Laszlo Polgars (father of the famous Polgar sisters) Endgame, was so damn thick and literally would take a lifetime to fully appreciate! Endgame Secrets by Lutz is also exceptionally instructive, smaller book but will take years to fully appreciate. I also loved to study lots of mate puzzles or other combinations, this helps sharpen tactics a great deal, there is another book from Laszlo Polgar, 5334 Problems Combinations and Games which where written in neatly organized themes.

I wouldn't say good players are very bad at it but certainly it is within the endgame where you must beat most of the knowledgeable players. The opening and middle game usually you beat people who are less experienced, or who you've book bash to death aahaha ^__^ I like blitz/lightning so that the game isn't so complicated and the win doesn't occur so often in the endgame, its nice to also see the tables turn round! Fast chess is though quite bad for your overall playing skill level, it does help with instincts and calculation speed but for deep thought you need for standard games of course its a bad idea.

This is something I posted on pianostreet a while back Ill copy paste it here, it was a game I played which I was proud of ahaha:

I was going through my old files and found one of my old favorite games. Here is a position where a real pretty mating combination ends with an amusing fianchettoed bishop mate Smiley The game was played 5 minutes a side, there are blunders but black made some strange moves in the opening which I exploited and thus my pieces where just so much more active.

SEE GAME POSTED ABOVE.

The funny thing about the position is that if black had his pawn on g6 (instead of g7) this mating combination would have not worked because then whites Queen would not be able to cut off the Black king on the h5 square (after white plays 29 h3+). This is one of the best mating combinations I ever played.

Thanks for the suggestions and words of encouragement!  I am currently working form the 5334 problems book (one of 13 books in my starter program).  I went thru the first 414 problems so far.  I am amazed that not a single typo have I found so far and all the puzzles have exactly 1 solution.

Great fast game!  I took a quick look at it just now.  The one thing that immediately caught my eye is the move by your opponent 9. ...  Bf8???  How can you castle with this kind of move?  This game shows the importance of castling EARLY!  But castling does not guarantee safety. The king chase reminded me of a game a saw in one of my books:  Edward Lasker vs Sir George Thomas, London Chess Club 1912.  Black (Thomas) had his King safely castled on g8 (so he thought) before the final assault started with a queen sacrifice on h7.  8 moves later the black king was mated on g1!!  7 checks in a row before black was mated on move 8 on the spot you would expect to find white's king.  Black's army was left with 6 pawns and 5 pieces (including queen) all on rank 6, 7 and 8.  They all watched and were unable to help as their king marched to his execution lol.

[ronde_des_sylphes]

Great game!!  I can’t believe how well you played.  One thing I decided to add to my program of study was analysis of master or GM games like I did of your game.  This is the first time I tried this. After EACH move of your game, I thought to myself: “Where would I move next?”.  After deciding, I had the computer show me if it was a good move or not.  If it was bad, the computer showed me how by kicking my **** within a few moves or so.  I then had the computer calc its best move and I also compared the computer move with you or your opponent’s move.  Anyway I learned a lot from this 1 game!  Here are a couple things I found in this game.  

The computer did agree with your opponent’s move of  15. … Qxc6 instead of 15. … bxc6 leaving an isolated pawn.  Also, both sides played great but the game came down mostly to a single blunder made by your opponent:  30.  ... Bc4??  The correct move was 30. ... Rxg6.  You would still have been ahead by 0.9 pawns had the correct move been made, but after your opponent made his move the computer had you ahead 16 pawns.

Thanks for the update on your CD.  This helps.  I will buy my copy on the Devine Art Label website when available.  It looks like they are set up to take USA customers.

EDIT: One other item in your game.  Great kingside attack!  The computer did show a way to slow the attach down by 22. ... Kf7 instead of 22. ... Be7 made by your opponent.  I thought Be7 was a good move at first because it connected his rooks and blockaded your e5 pawn and attacked the a2 pawn.  He did not need to worry about this bishop being trapped if he took your a2 pawn since the c2 pawn was pinned.  Unfortunately the move did not prevent 23. Qh6!

 

Thanks.. it's so long since I even thought about this game that I've never checked to see how well it holds up under computer analysis and I don't even have a program on my pc. I don't agree with the computer wanting to recapture on c6 with the queen and speculate that this is an example where human understanding trumps silicon calculation. The reason is that the defining strategic characteristic of the position is the opposite side castling means Black attacks on the queenside and White on the kingside. The way play went is I think indicative that White's attack with the resulting structure after Qxc6 is faster than Black's. This is partly because the open file Black has for attacking purposes is the c file and as things stand c2 is quite well protected. b2 is a more susceptible weakness should Black take with the pawn on c6 - the a-rook to b8 and the bishop to f6 is a natural and obvious plan and would likely induce the weakening move b3 after which there is a target for an a5, a4 pawn advance as well as dark square weaknesses for the f6 bishop and the Queen to prey on. I agree that the computer probably prefers the Queen recapture due to concerns with the greater number of pawn islands but contend that the dynamic attacking possibilities of the open b file are more important. I'd like to hear a GM's view on the matter, but this was also the conclusion arrived at during the game postmortem. 30.. Bc4 is of course a blunder, but my opponent had simply missed the rather pretty queen sacrifice in the game. Imo he is holding on after Rxg6. 22.. Rf7 is I assume the computer suggestion - I think White's attack is still faster than the Black one but it isn't totally clear if it is enough.

In any case whilst this isn't the best game I've played it remains one of my more memorable ones, because of the winning combination and also by dint of my opponent being rated about 2450 at the time of the game. I have wins against GMs in blitz but this is one of my more notable tournament scalps.

[georgey]

 

Thanks.. it's so long since I even thought about this game that I've never checked to see how well it holds up under computer analysis and I don't even have a program on my pc. I don't agree with the computer wanting to recapture on c6 with the queen and speculate that this is an example where human understanding trumps silicon calculation. The reason is that the defining strategic characteristic of the position is the opposite side castling means Black attacks on the queenside and White on the kingside. The way play went is I think indicative that White's attack with the resulting structure after Qxc6 is faster than Black's. This is partly because the open file Black has for attacking purposes is the c file and as things stand c2 is quite well protected. b2 is a more susceptible weakness should Black take with the pawn on c6 - the a-rook to b8 and the bishop to f6 is a natural and obvious plan and would likely induce the weakening move b3 after which there is a target for an a5, a4 pawn advance as well as dark square weaknesses for the f6 bishop and the Queen to prey on. I agree that the computer probably prefers the Queen recapture due to concerns with the greater number of pawn islands but contend that the dynamic attacking possibilities of the open b file are more important. I'd like to hear a GM's view on the matter, but this was also the conclusion arrived at during the game postmortem. 30.. Bc4 is of course a blunder, but my opponent had simply missed the rather pretty queen sacrifice in the game. Imo he is holding on after Rxg6. 22.. Rf7 is I assume the computer suggestion - I think White's attack is still faster than the Black one but it isn't totally clear if it is enough.

In any case whilst this isn't the best game I've played it remains one of my more memorable ones, because of the winning combination and also by dint of my opponent being rated about 2450 at the time of the game. I have wins against GMs in blitz but this is one of my more notable tournament scalps.

1) Good explanation on the recapture on c6 and this does make a lot of sense to me.  The computer actually preferred bxc6 in its first 30 seconds of thinking but then went to Qxc6 for the next 4 minutes it thought.  At the end of thinking it felt the game was exactly even with Qxc6 and black was behind 0.18 pawns with bxc6.

2) It was a rather pretty queen sacrifice you offered with 31. Qh7+!    I thought about this for quite a while before I could see the mate that followed if he took the queen.  That was a great move!  You played a very accurate game.  If your opponent did not blunder on 30. … Bc4 and instead played  30. … Rxg6, here is how the computer calculates the continuation:

31. Bxg6  Qf4+ (Qf4+ forcing the Queen exchange)
32. Qxf4  Rxf4
33. Rg5 to capture a pawn.

If queens were exchanged as shown above, it would likely come down to you having a 1 pawn advantage: white having 2 pawns (on b and c files), a white bishop and rook vs black having 1 pawn on the “a” file, a white bishop and rook.  But he did blunder (or missed seeing the great move you made next would be a better way to describe).  Congratulations on a great win!

note: I corrected the edit in my prior post: Be6 (not Be7).

[georgey]

Hello chess/piano enthusiasts! Just an update. Comments welcome from all you chess experts and masters.

I am close to finishing going through 15 books (first of several readings) and I feel I chose great books to learn tactics, strategy, general opening strategy and end games. I have not posted anything on any chess websites, just here.  I decided to pick an opening repertoire rather than wait even though I am a beginner.  I just ordered 6 opening books.  I don’t plan on memorizing, but rather I will try to come up with a move on my own for each position and test it against Komodo 10 software before seeing the recommended move in the my opening book.  Of course the books will give the reasons for making moves.

I am deciding upon the following opening repertoire after doing research: Either A or B:

A:
White: Colle-Zukertort system – an easy to learn system playing 1. d4 – book called “Zuke-em”
Black: for 1. e4: French defense
Black: For all else: queen gambit declined Tartakower variation

B:
White: 1. e4 repertoire that is basic in a book called 1. e4! By Neil McDonald in the Everyman starting out chess series. Nothing fancy.  He picks 3. d4 over Roy-lopez or Giuoco piano for example.  He appears to covers the open Sicilian defenses pretty well (2...e6 and 2...Nc6 and 2...d6).
Black: for 1. e4: Sicilian Scheveningen variation plus learn to defend closed Sicilian
Black: For all else: queen gambit declined Tartakower variation

I am currently planning to pick B.  Reason: to be familiar with the traditional white openings plus I like open games more.  I may get slaughtered playing the Sicilian defense as black though but it should complement my learning 1. e4 as white.

[pianoplunker]

Hi Pianostreet members.  Just wondering if there are any chess enthusiasts here.  I might be posting in the wrong site but there are chess grandmasters that were world class pianists (e.g. Taimanov who was USSR world champion chess player candidate and is listed as one of the great pianists of the 20th century).  French composer Philidor was the greatest chess player of the 1700's. Although my first love is music, I view chess to be an equal to classical piano as a pursuit.   I would say there are as many people passionate about chess as classical piano, maybe more.  Chess.com has I think 15 million members with 30k on line at any time.

I dont play chess all the time but enjoy it just as much as piano except sometimes in Chess I get mad. When my opponent destroys my whole plan I want to slap somebody. To me Chess is very much like piano in the sense if I dont do it often I lose something. Right now I havent played a Chess game in a year or so. It is possible I could be defeated.

[georgey]

Final update:  I went through 22 books one time in the last 9.5 months and I will be going thru 15 of these multiple times in the next 2-3 years.  Of course tons of games will be played against Komodo 10 – THE super Grandmaster!  I probably will start playing people about a year from now.  Goal – 2000 elo in the next 3-5 years.  Goals are needed!  Some parts of my brain work not so well learning as a beginner at age 58, but some parts work better than ever, including planning and organization skills.

My final choice of openings having gone through this one time is:

Playing White:

“1. e4!” book by Neil McDonald covering basics such as: Philidor Defense, Petrof Defense, Scotch opening, Alekhine Defense, Caro-Kann,  Scandinavian, Pirc, French defense Tarrasch variation.

For Sicilian with me playing white, I chose the Grand Prix attack.  The “1. e4!" book has the open Sicilian: 1. e4, 2. Nf3, 3. d4, but I found this to be too much. So instead I chose the Grand Prix attack as white: 1. e4, 2. Nc3, 3. f4.  This closed Sicilian is much less complicated than the open Sicilian and is covered well in Lev Alburt’s “Chess openings for White explained”.

Playing Black:

Against 1.e4 – Scheveningen Sicilian with Najdorf move order to avoid the dreaded Keres attack.

Against all other first moves for white: Queen Gambit Declined – Tartakower variation.

I am using 3 books to piece together a very modest Black repertory as a beginner.  Books by Matthew Sadler, Neil McDonald and John Emms.

[cuberdrift]

Hah. Rachmaninoff posts an "anyone here follow basketball" thread and gets a handful of replies.

Then THIS thread comes along and well...dozens of replies.

It's clear that pianists are nerds.

(And yes, I know how to play chess, and don't know how to play basketball).

[klavieronin]

I don't play chess but one of my favourite films of all time is an Indian film called "The Chess Players" directed by Satyajit Ray. I definitely recommend it for chess enthusiasts.