Most of the time, the human element is the least enjoyable thing about this game. Even if the game is perfectly solved, if we can emulate human-like learning and fallibility without the rest of the unpleasantness of the interpersonal experience, we're golden.
@9497010838 - well said, agreed! 
Dodged!
Until Skynet comes on line, I think computers will be able to figure out a lot of theory and in fact, computer analysis has changed many openings theory over the years. What it cannot do is replicate the human artistry of a chess game. This is due to the varied styles in both players and the ability to come across c combinations c that while not as effective as a computer might give in a position, achieve the same end as what the computer suggests as best play. Also, many positions I have played here when placed under analysis give inaccuracies if the evaluation c is say .3 of a pawn difference or if the move allows for tactical play by your opponent. Worse yet is when you go from quick to deep to maximum, the same move may be better or worse c than the previous lower evaluation. So if three different analyses yield three different move notations, it seems unlikely that computer engines and AI Will Dodge l figure out all of chess. A good example is to look at the Blumenfeld Counter Gambit. That opening has a lot of room to expand on theory still.
@billy223 - thank you for mentioning Blumenfeld Counter Gambit
We haven't even solved the opening phase... For each opening combination there is an equally huge amount of middle game continuations.
You cannot prove that the basic position a Queen up is a win. Show me the TB!
The problem of demonstrating a win with queen odds at the start is definitely easier than solving chess, but is it enough easier to be feasible?
I was thinking that there could be a point where the ability to see so many moves ahead might create more problems than answers. Perhaps a computer that can only see 15 moves ahead will beat one seeing 20 moves ahead, if my theory turns out to be correct..
I was thinking that there could be a point where the ability to see so many moves ahead might create more problems than answers. Perhaps a computer that can only see 15 moves ahead will beat one seeing 20 moves ahead, if my theory turns out to be correct..
No.
hopefully not
Yes.
"the opposite of your image of him"
I don't have an "image" of Kramnik. He had the balls to be the last champion to play even odds with an engine. That's it. I just thought it was pointless to point out how poorly he played in said match when no other GM would even risk making the attempt ...
It’s easier to project courage by playing a machine, knowing that losing to it won’t endanger your title, than to offer Kasparov a rematch, when Kasparov was still the number one-rated player and was winning strong tournaments left and right. So I question that image of a ‘courageous Kramnik’.
"the opposite of your image of him"
I don't have an "image" of Kramnik. He had the balls to be the last champion to play even odds with an engine. That's it. I just thought it was pointless to point out how poorly he played in said match when no other GM would even risk making the attempt ...
It’s easier to project courage by playing a machine, knowing that losing to it won’t endanger your title, than to offer Kasparov a rematch, when Kasparov was still the number one-rated player and was winning strong tournaments left and right. So I question that image of a ‘courageous Kramnik’.
Read the first 3 sentences of your quote of me again. They still apply. The last sentence still holds true, as well.
There is miles of room between "craven" and "courageous" for Kramnik, or anyone else for that matter.
There should be a logical fallacy for arguing everything by pushing things to opposite extremes of black or white and ignoring other alternatives or any grey area at all 
...oh wait there (informally) is: https://en.wikipedia.org/wiki/False_dilemma
I was thinking that it might be interesting to program these computers so that a draw isn't an option. It would be interesting to see how differently the games are played when computers are going for a win not a draw.
They already have this setting/option, as is easy to discover if one looks
. Obviously, it makes engines play more aggressively. It also sometimes forces them to lose when in a drawn position because they sacrifice material or ruin their position trying to win.
Nakamura won a famous game against Rybka in 271 moves this way...Rybka's "contempt" setting was set to max, so Nakamura locked up the the game in a closed position, then waited...when the 50-move rule was about to force a draw, Rybka sacrificed material and lost badly because its contempt setting was set to win at all costs.