If the present-day best computer starts calculating from move one, does it draw vs the theoretical perfect strategy when its white? If yes what about black?

Might a present-day computer be more likely to beat a 2700 than perfect play? (It is logically possible because perfect play might eschew a move that gives the opponent a fifty-move forced checkmate while the computer increases its win chances because it doesn't see it and makes the move. Obviously, such scenarios do occur in poker.

Why you assume perfectly play equal a win and not a draw ?

Don't get what you are saying. Perfect play means that if the opponent moves into a position where there are now a forced series of moves that will checkmate him, the perfect player will always see it and do it no matter how many moves it takes. I'[m thinking that you were thinking about both players being perfect. They say that is almost certainly a draw. My question is whether a slightly imperfect player, ie modern computers, are good enough to avoid the forced checkmate the perfect player will always see.

I don’t know if this info would help but computers now are playing perfectly up to this :

https://en.m.wikipedia.org/wiki/Endgame_tablebase

« Tablebases are generated by retrograde analysis, working backward from a checkmated position. By 2005, all chess positions with up to six pieces, including the two kings, had been solved. By August 2012, tablebases had solved chess for almost every position with up to seven pieces, but the positions with a lone king versus a king and five pieces were omitted because they were considered to be "rather obvious."[1][2] These positions were included by August 2018.[3] As of 2022, work is still underway to solve all eight-piece positions.

The solutions have profoundly advanced the chess community's understanding of endgame theory. Some positions which humans had analyzed as draws were proven to be winnable; in some cases the tablebase analysis could find a mate in more than five hundred moves, far beyond the horizon of humans, and beyond the capability of a computer during play ».

By the way it is logically possible, though I assume very unlikely, that a computer could be beaten by a non-perfect player but not a perfect one. If the non-perfect opponent makes a move against the computer that is a 70 move forced checkmate against it, the computer might not see it and unwittingly play a counter move that is a 70 move checkmate against its own self which it also doesn't see. Opening the door for the non perfect player to win. That couldn't happen if the computer's opponent was perfect and saw all forced checkmates because he would never make that first erroneous move.

The bottom line is that if there is presently a computer that has lost one game in a thousand you can't deduce with absolute certainty that a perfect player could beat it as well.

I feel like you've asked this question multiple times in various forms which always boils down to "can you play exploitatively in chess and have better results than theoretically optimal", and just like poker the answer is generally no, but there can be exceptions if you can find certain situations where a "blind spot" exists and gamble - in the case of chess, something beyond the computer's calculation horizon.

Jim Plaskett's Puzzle
White looks lost but wins
FEN format:
8/3P3k/n2K3p/2p3n1/1b4N1/2p1p1P1/8/3B4 w


Its been a while since I played much chess so I'm a little rusty. I think this position is a candidate for such a position. Looks at first that White is lost. Its a famous chess puzzle that take around 50 moves ahead to see the correct move. Even at a highly strong level chess computer usually choose the wrong first move since grabbing material right away seems best. To computer chess software it looks like White is losing badly in all trees. A human may not see that far ahead either, but by chance or instinct he might pick the right first few moves more often than a computer.

There are two differences between chess and poker. Chess has draws. That means that perfect play may not beat a non perfect computer once that computer has reached a certain level. I was wondering whether we are there yet. And the fact that such a computer just might lose to a human does not automatically prove that we are not there. The other difference is that in chess a win is a win. No extra points or money is given based on how badly you beat an opponent.

Here is a strange experiment for computer people. How many moves on average will it take for various programs to beat someone who moves randomly? (except for moving into check). In other words, play a million games where one side is you super chess computer and the other side is a chimp. Obviously the answer will be a lot higher than for a good playing human.

Great players would take different routes to winning a match. Karpov was famous for grinding down opponents into losing positions, while other players like Fisher could win games convincingly in under 25 moves. Both players were well-established all time greats. I'm not sure that trait means all that much. It simply reflects how a player trains to play. Winning is important, not how quick it takes.

This I don't get.

First you say that a win is a win and no extra points are gained. In your next paragraph, you essentially equate winning faster to winning better.

Let me put it another way: "In other words, play a million games where one side is you super chess computer and the other side is a chimp. Obviously the answer will be a lot higher than for a good playing human"

The super chess computer will win 1 million times. The good playing human will not.