@7381
"you do realize that such evaluations can be wrong?"
++ All evaluations are wrong to some extent:
the only right evaluation is draw / win / loss from the 7-men endgame table base.
That is also how Checkers has been weakly solved:
calculate until the exact evaluation draw / win / loss of the endgame table base.
the sources i have posted explicitly support my position. what do you define "evaluation" as anyways? a measure of who is winning, or by how much? you do realize that such evaluations can be wrong?
Whether they are wrong (and they are *all* presumed to be wrong until each game is solved) or right is immaterial here. Under state of the art conditions for Go and Chess playing software and using the best available methodologies available for each currently, an evaluation of a discrete Chess position should take more CPU power to complete than a discrete Go position when using a NNUE hybrid of machine learning and brute force calculation for chess versus straight machine learning for Go.
lmao what '"whether they are wrong or right is immaterial here"
you do realize that that is the most important part???
lmao what '"whether they are wrong or right is immaterial here"
you do realize that that is the most important part???
Not remotely. All "evaluations" derived for as-yet unsolved games are assumed to be wrong. I'm just winning the point you tried to argue so you will think harder next time. This discussion is meaningless for any other purpose. It adds nothing to the topic, and I have no interest in your conclusions. This is just a newspaper to the snout.
the sources i have posted explicitly support my position. what do you define "evaluation" as anyways? a measure of who is winning, or by how much? you do realize that such evaluations can be wrong?
Whether they are wrong (and they are *all* presumed to be wrong until each game is solved) or right is immaterial here. Under state of the art conditions for Go and Chess playing software and using the best available methodologies available for each currently, an evaluation of a discrete Chess position should take more CPU power to complete than a discrete Go position when using a NNUE hybrid of machine learning and brute force calculation for chess versus straight machine learning for Go.
and yet a go position ends up taking more computing power.
your logic must be flawed somewhere.
@7386
"Stockfish can not even solve simple 8 man TB positions."
++ If given enough time it can.
Besides you first have to prove your 8-men position is relevant, i.e. can result from optimal play from both sides. Your example cannot result from the initial position by optimal play from both sides and thus is not relevant to weakly solving Chess.
and yet a go position ends up taking more computing power.
your logic must be flawed somewhere.
Show your work. Your articles say no such thing. If you think that they do, then you do not understand them. You keep ducking the question...do you have any software or hardware background? Roblox does not count.
ok what does "The game of Go has long been viewed as the most challenging of classic games for artificial intelligence owing to its enormous search space and the difficulty of evaluating board positions and moves" mean?
are you seriously trying to argue that strength of evaluations doesnt matter?
ok what does "The game of Go has long been viewed as the most challenging of classic games for artificial intelligence owing to its enormous search space and the difficulty of evaluating board positions and moves" mean?
It means that overall, due to Go's much larger move tree, it's the most challenging "mainstream" game to solve. But we're not discussing that...and that imprecision is why you lost this argument.
did you miss the "difficulty of evaluating board positions and moves"?
I can evaluate a billion positions of chess a second with just a pocket calculator, and I will evaluate 3 tic tac toe positions a second with the same pocket calculator, ergo tic tac toe would require the more power to solve
did you miss the "difficulty of evaluating board positions and moves"?
Your logic is failing here.
Read the quote:
"The game of Go has long been viewed as the most challenging of classic games for artificial intelligence owing to its enormous search space and the difficulty of evaluating board positions and moves"
The word "and" here denotes a combination. There are games with more difficult to evaluate singular board positions (like Chess...). There are games with larger search spaces. There is no (popular/well known) game that has both to such a degree. An NBA team is #1 ranked. Do you assume that that this must ergo mean that they have the #1 offense *and* the #1 defense (as you are doing with this line of reasoning for Go)? Because they might be ranked #3 in one and #7 in the other, and still be ranked #1 overall. It's not a tough concept.
did you miss the "difficulty of evaluating board positions and moves"?
Your logic is failing here.
Read the quote:
"The game of Go has long been viewed as the most challenging of classic games for artificial intelligence owing to its enormous search space and the difficulty of evaluating board positions and moves"
The word "and" here denotes a combination. There are games with more difficult to evaluate singular board positions (like Chess...). There are games with larger search spaces. There is no (popular/well known) game that has both to such a degree. An NBA team is #1 ranked. Do you assume that that this must ergo mean that they have the #1 offense *and* the #1 defense (as you are doing with this line of reasoning for Go)? Because they might be ranked #3 in one and #7 in the other, and still be ranked #1 overall. It's not a tough concept.
whats funny is that i literally predicted that you would say that.
where u fail is that the "difficulty in evaluating board positions/moves" is mathematically equivalent to the overall difficulty.
how hard is it to see that the sheer amount of moves is what makes evaluating positions hard, among other things
its also funny how you still refuse to concede your original claim, that Go as a whole is less complex than chess.
how hard is it to see that the sheer amount of moves is what makes evaluating positions hard, among other things
Lol. You added "among other things" by choice, which confirms the point I just got done making. Why do I need to keep arguing when you tacitly admit you are incorrect all by yourself?
Good night. I'll eviscerate any new stuff that emerges later.
its also funny how you still refuse to concede your original claim, that Go as a whole is less complex than chess.
Link it. No such claim was made.
2 things.
first, Go is much more complex than chess. it took 20 years after deep blue to get the same level for AlphaGo.
second, a massive number of permutations doesnt necessarily mean that something cant be solved. checkers had 10 ^20 and was still solved. of course, chess is much much more complex, but the big number alone should dissuade us.
Your premise assumes that the efforts put into beating the world champs for Chess and Go were the same. This is not the case. Solving Chess was much better PR for IBM than solving Go would have been, so a lot more resources were brought to bear.
In terms of actually solving, IIRC Go has more positions...but evaluating Go positions should take less CPU horsepower than evaluating Chess positions.
btw, deep mind cost 400 million, while deep blue cost 100 million, although inflation might put deep blue over
@7376
"Chess positions should take more CPU to evaluate than Go positions, one for one."
++ Solving Chess does not depend on some evaluation, but on the 7-men endgame table base.
Stockfish is designed to play, i.e. find one good move in some time limit e.g. 3 min / move.
To do that, it depends on some evaluation as it cannot calculate all the way in that time limit.
Stockfish can be used to analyse, using more time and taking e.g. 2 moves instead of 1 move into account. Then it will in part depend on evaluation, as some lines will not reach the 7-men endgame table base.
Stockfish can be used to weakly solve Chess, using much more time:
5 years on 3 cloud engines of 10^9 nodes/s, or 15000 years on a desktop,
calculating all the way to the 7-men endgame table base.
Shoo. I've already refuted all your stuff, and this point is meaningless for the subtopic being discussed.