In the ChessTalk thread titled "Time Increments", I asked Kevin Pacey: where is the interest in the most elite form of chess possible, namely chess played between top computer engines such as Houdini versus Stockfish? In other words, where is the book on such games? (those two engines did play each other I believe in 2013, in a match of many dozens of games, and this match went for the most part unnoticed and as far as I know, there is nowhere to be found an in-depth analysis of the games).
Kevin posted in that thread this response:
"A book on a computer vs. computer match would likely sell far less than a book on a match between elite humans that wasn't a blowout, because we can't calculate nor evaluate positions as well (or in the same way) as computers, and their 'styles' would be harder to explain in depth, or enjoy."
What Kevin appears to be saying is that not only he personally, but most chess players who follow the game's latest developments, would not enjoy a book analyzing Houdini versus Stockfish games because:
(a) they wouldn't understand the analysis (presumably even if the analysis was written by humans, which begs the question... who else would write it?), and
(b) they wouldn't be able to discern any 'style' in either computer engine's play.
I think Kevin's points are important enough to warrant a separate thread on the topic of just what IS a chess player's 'style', and asking whether Kevin's points are valid. Why is this important? I think because everyone who can call themself a serious chess player seems to want to see or to develop or see developed in others 'quality chess'. I keep seeing this mentioned again and again. Yet nobody seems to pay any attention to the greatest quality chess, computer engine versus computer engine. Regardless of style considerations, these engines when matched against each other give us the pinnacle of chess quality. You must believe in that if you believe in the ELO rating system. So what that means is: we should all be able to learn from analysing games between top engines, more so than analyzing ANY human versus human games or even human versus engine games.
On the question of 'style', only a few words come to my mind. There is 'positional' versus 'tactical', there is 'conventional' versus 'unorthodox' (especially in openings), and there is what I would call 'broad' versus 'narrow'. Those last terms refer to whether a player is adept and expert in all phases of the game (opening, middlegame, endgame) or only one or two of those phases, and may actually be weak in one or two phases.
Additionally, there could be the concepts of 'continuity' versus 'discontinuity'. This refers to whether a player shows a tendency to embark on a strategic plan and stick to it come what may, or instead shows a tendency to jump from strategic plan to strategic plan based on what the opponent is doing to change the nature of the game.
If anyone else has more definitions of style, by all means add them in this thread so we can all get a more complete picture of exactly what this entails.
So I suppose the first question many of you could answer is: for you to analyze a set of games set forth in a book, do you need those games to be a collection that contrasts two different styles? For example, Kasparov versus Karpov or Fischer versus Spassky?
Alternatively, would you ever be interested in a book that analyzes games purely for the sake of learning 'quality chess', i.e. a book that showcases games between many different players, in which the games have been selected for their instructive value alone?
If a prominent chess author, let's say a John Nunn or a Jeremy Silman, were to author a book analyzing Houdini versus Stockfish, would you buy it and if so, purely for that author's reputation?
What if the book were authored by some run-of-the-mill GM, and it seemed to have good annotations and diagrams, would you buy the book in order to learn about 'quality chess'?
And finally: would you agree that the lack of discernable 'style' in the play of computer engines on the whole (which we can agree for now is entirely possible) means that there is little to nothing you could learn from analyzing games between such engines, either on your own or with the aid of such a book authored by a GM?
That last question gets to the crux of Kevin's points: is chess 'objective' enough that players can learn something from the highest quality play in terms of ELO rating, even if that play exhibits no discernable 'style' over many games?
Kevin posted in that thread this response:
"A book on a computer vs. computer match would likely sell far less than a book on a match between elite humans that wasn't a blowout, because we can't calculate nor evaluate positions as well (or in the same way) as computers, and their 'styles' would be harder to explain in depth, or enjoy."
What Kevin appears to be saying is that not only he personally, but most chess players who follow the game's latest developments, would not enjoy a book analyzing Houdini versus Stockfish games because:
(a) they wouldn't understand the analysis (presumably even if the analysis was written by humans, which begs the question... who else would write it?), and
(b) they wouldn't be able to discern any 'style' in either computer engine's play.
I think Kevin's points are important enough to warrant a separate thread on the topic of just what IS a chess player's 'style', and asking whether Kevin's points are valid. Why is this important? I think because everyone who can call themself a serious chess player seems to want to see or to develop or see developed in others 'quality chess'. I keep seeing this mentioned again and again. Yet nobody seems to pay any attention to the greatest quality chess, computer engine versus computer engine. Regardless of style considerations, these engines when matched against each other give us the pinnacle of chess quality. You must believe in that if you believe in the ELO rating system. So what that means is: we should all be able to learn from analysing games between top engines, more so than analyzing ANY human versus human games or even human versus engine games.
On the question of 'style', only a few words come to my mind. There is 'positional' versus 'tactical', there is 'conventional' versus 'unorthodox' (especially in openings), and there is what I would call 'broad' versus 'narrow'. Those last terms refer to whether a player is adept and expert in all phases of the game (opening, middlegame, endgame) or only one or two of those phases, and may actually be weak in one or two phases.
Additionally, there could be the concepts of 'continuity' versus 'discontinuity'. This refers to whether a player shows a tendency to embark on a strategic plan and stick to it come what may, or instead shows a tendency to jump from strategic plan to strategic plan based on what the opponent is doing to change the nature of the game.
If anyone else has more definitions of style, by all means add them in this thread so we can all get a more complete picture of exactly what this entails.
So I suppose the first question many of you could answer is: for you to analyze a set of games set forth in a book, do you need those games to be a collection that contrasts two different styles? For example, Kasparov versus Karpov or Fischer versus Spassky?
Alternatively, would you ever be interested in a book that analyzes games purely for the sake of learning 'quality chess', i.e. a book that showcases games between many different players, in which the games have been selected for their instructive value alone?
If a prominent chess author, let's say a John Nunn or a Jeremy Silman, were to author a book analyzing Houdini versus Stockfish, would you buy it and if so, purely for that author's reputation?
What if the book were authored by some run-of-the-mill GM, and it seemed to have good annotations and diagrams, would you buy the book in order to learn about 'quality chess'?
And finally: would you agree that the lack of discernable 'style' in the play of computer engines on the whole (which we can agree for now is entirely possible) means that there is little to nothing you could learn from analyzing games between such engines, either on your own or with the aid of such a book authored by a GM?
That last question gets to the crux of Kevin's points: is chess 'objective' enough that players can learn something from the highest quality play in terms of ELO rating, even if that play exhibits no discernable 'style' over many games?
Comment