jefk wrote:Lyudmil Tsvetkov wrote: So forget all about theory: theory brings you nothing, it only handicaps you intellectually.
that's a rather bold statement; would you also say 'forget about
all endgame theory'?
Indeed humans cannot calculate an endgame as eg a syzygy tablebase
can do, but there are other ways of mastering endgame theory.
In a similar way there is still scope of finding new ways in opening theory,
which btw is also what i have been doing with the help of comp evaluation (although not perfect,i admit, as you often are pointing out here, but still gradualy improving in fact partly as result of people like you who suggest possible improvements) and playing games. For me that's the creative part even although its not on the board, but in the preparation stage.
If Carlsen is that good in your opinion, ask him to play some standard
games on ICC against a few stockfish accounts, including mine (currently nr 1 again in standard at 2880), i bet he will not come further than about
2750 or so, because those ratings are inaccurate; on playchess the top comp ratings also have decreased, and i bet Carlsen in the engine room would not get further than 2650 or so.
So summarizing i do *not* mean/advise memorizing 'current' theory, as eg indicated by these percentages in the typical chess base .ctg books, but more looking at eg top accounts in playchess, or bbuilders games on ICC, listening to advise by seconds who also are good in computer chess. Take eg these boring Berlin games, imho you first play Rd1+ to prevent the black king escaping to c8, and only later moves as h3! followed by a later Nc3.
When Carlsen played against Anand he did not know such things.
And there certainly is scope for improvement, provided such a human
indeed wants to exceed the 2900 barrier. If a bloke as MC wouldnt like it,
well then maybe younger guys as Wesley So, Anish Giri, or others will later do it anyway, and become a new generation of world champions.
Not creative ? Well if a person is very creative then maybe chess is not the right game, and becoming an artist in eg painting would be more appropriate and more fun. Eg after my first book, about chess , i now am thinking of becoming a writer about futuristic thrillers, a mix of Sf and crime.
Chess is not the only intellectual occupation out there, you know.
jef
Thanks for thinking I had some suggestions for computer chess.
No one can refute a statement like:
- opening accounts for 5% of performance
- late endgame accounts for 5% of performance
- and middlegame accounts for 90% of performance
If everything is decided into the middlegame, why study opening and endgame theory? Better just concentrate on playing/replaying/engine analysing thousands of middlegame positions.
But I agree that you should know endgame theory, you really can not do without that, as those are the basics, and besides in endgame theory everything is finite and pretty much solved.
What concerns opening theory, however, it is endless, you do not know where to start and where to finish. And you also do not know if that theory is true, as it is not clear at all, theory changes with time and understanding, the lines are very deep, etc., so really difficult to make sense of it all.
Engine room? What does Carlsen have to do with an engine room? Human-computer competition is difficult under standard conditions, primarily because of 3 reasons:
- humans get tired, computers do not, so at some point the human just can make a mistake as he is tired after a prolonged game
- humans get nervous sometimes, computers never get nervous; well, you have a very good position, even winning, and suddenly you get nervous, for whatever reasons, you can not think soundly, make a rash move, and you lose the game, that never happens with computers
- humans get distracted; well, you might be well concentrated for a long time, but then suddenly someone distracts you, even if briefly, you lose concentration, and make a blunder, meaning the game is over. Nothing similar happens to computers
The only advantage of humans is better knowledge, but that is already not sufficient to offset the 3 abovementioned computer advantages, as engines also advance in knowledge.
Therefore, the only reasonable way to match computers and humans would be playing under strictly Fischer conditions: in a back room, with no cameras, any annoying buzzing sounds and sensation-hungry crowds staring at you.
I bet Carlsen would have good chances to perform well, even win a match against current top engines under such conditions.
If Carlsen knows the Berlin or not, I do not know, but it is a fact he performs very well there. If engines say Carlsen is bad in a certain position, for example the last game of the recent final with Anand, and Carlsen himself believes he is having advantage, I am inclined to trust Carlsen rather than the engine, even if SF and Komodo.
As said, engines play weakest in the opening, they are much better in the middlegame and endgame, simply because those stages are less complex.
I really do not understand a person, who memorises 25 or 30 book moves, has the luck to play all of them in a certain OTB game, winning, and this makes him happy. What is there to be happy about, that you memorised something successfully, when you did not play even a single move of your own, you did not even think about it?
I myself, would prefer to play f3 on the first move, however suspicious that may seem, but think about the position throughout the game. Btw., do you know that the Barnes Defence, e4 f6, is actually not that bad, if that is true, there are good chances f3 is even better?
Futuristic thrillers? Sounds good.
Stockfish and crime? Even better.
(I imagine one of the main SF contributors assassinating another main SF contributor, simply because the former stole and implemented as his an idea that was actually the latter's; imagine the fuss on this forum)But seriously, no need to write about that, this is already part of history, the future has come to stay with us a long long time ago.
Gary Linscott managing the framework, pooling together the efforts of people all over the world, working for nothing more than the simple task to create the strongest engine on Earth. Who would have believed that to be possible some 20-30 years ago, when Clarke and Asimov were still active?
Or even better, a Chinese CPU contributor appearing yesterday for the first time on the SF framework. Is not this sensational? I think you can start your next novel like that: John Stevenson expected everything from the framework, but not ever seeing a Chinese red flad waving across the contributor machines bar, and you can choose for a title something like The Chinese have come, or Red Flag.
Good luck.
