Game Theory - Perfect Strategy?

It is true that the history of a position (beyond indicating the legality of castling or en passant captures) is relevant to the evaluation of a position. From the point of view of a tablebase,  three-fold repetition can ignored for theoretical purposes (the 50 move rule will eventually cover all circumstances where it would be possible to force a threefold repetition). But three-fold repetition is affect the value of a real position in a curious way. If a player is ever able to reach a three-fold repetition that gives him/her a draw, even if the position they reach is a tablebase loss.

However, this shows how theoretically unimportant this issue is: in order to be significant, players will have have to reach a position which was won for one side twice, without that player making any progress. Clearly that player is not up to the task of winning a won game.

From the point of view of theory, it is reasonable to ignore such silliness, and assume that players don't get to winning positions and get nowhere. Tablebases are based entirely on best/best play from a position.

This is well said!

A "perfect" strategy for chess is at this time absolutely unfathomable, for obvious reasons, but ones I will explain to the OP anyway.

For every strategy you come up with, there are several reactions for your opponent and those may have certain flaws that require a strategy modified from the original to take advantage of. Within these strategies are tactics too, and how the tactics work out can often be the sole determination of what strategies in a specific position work and don't work.

So really the perfect strategy is like a preparation of thousands of strategies, combined with millions of tactics, to whatever your opponent does. Good luck professor.

Everything before your last paragraph is obviously correct, but also something every mammal who plays chess probably knew.

Of course, none of this matters.

Checkers was solved in 2007. Do you know how many human players have the ability to memorize the solution completely, and implement it perfectly in every game, even against imperfect play? Zero. The solution to chess will be even more complex and unlearnable to humans.

Since it won't affect our ability to play and enjoy the game, who cares if it gets solved?

Exactly, and well put.

There can be no unique solution to chess, only a large number of "solutions" covering all variations. For example: Suppose "best" play  by both sides, which is "solved," turned out to be a variation of the Ruy Lopez [Blasphemy! It has to be a Queen's pawn opening!], and this solution was a win for white in all cases.  Of what relevance is that solution to a Sicilian Defense, a French Defense, the Caro-Kann, etc.? Each would likely require its own solutions, for "best" play as well as imperfect play. Memorizing them all would likely be vastly more difficult than memorizing MCO.

A "solution" might be fascinating. It won't "kill" chess between humans, though.

rethinking the SPNE approach, perhaps starting from every possible mate from whites then reworking the game backwards to every opening permutation would be the best way to solve chess.  of course this would also apply to every possible mate for black, and if no solution is reached then one could imply a draw as equilibrium.  To calculate backwards (subgame perfect nash equilibruim) one would have to start at (for example) a smothered mate (persumably with a knight) and work back all possible permuations on the randomly placed initial mate.  just from this type of mate one would have to then add (legal) pieces to each square to exhuaste every permuation for that mate in that position.  this gets very large very fast.  Also, the SPNE approach would require that all forms of mate are known, while it is clearly defined one would have to run through every permuation that would be a checkmate on every square in every possible position that maintians that mate; only to rework it backwards and find the "dominant opening" which leads to mate in every permuation for white.  This search would extend to black as well. 

I think theres more of a chance in applying something like graph theory where the strategy would be to keep several paths to your oppenets king open, and if they block one move to open another (move like geometry).  This is of course hypothetical, but this type of approach rejects the game thoeriest analysis based on the (current) impossiblility of calculating the possible permuations.  

Unlike tic-tac-toe, it is impossible to master perfect play as a human being which makes it a game that will live on, in relation to our lifespans.

I would consider a full solution of chess to be a tablebase giving the correct result of any position and,  if it is 1-0 or 0-1, the number of moves required with best play. Of course this would allow you to evaluate every move in any position. It is highly likely that this solution would confirm that a wide range of openings are theoretically sound (very likely including the Lopez and Sicilian as well as some queen pawn openings!). In most positions there will be several moves that achieve the same result.

Professor Scott P. Stevens' quote at the top of this thread (now five years  old) is likely mistaken due to the perfection of chess machinery.

Recent advances in opening theory using Komodo 9.1 show White sustains a win (at ply 98, #23) with perfect play on both sides using D00 Queen's pawn Chigorin (4.Bf4).  (This is quicker and sooner than with D01 Veresov-Richter attack (4.Bg5).)  What follows now is that "1.d4 is best by test".  With subsequet replication, this hypothetical rule may stand the test to become a theorem.

You really haven't the slightest clue what you're talking about, in chess OR mathematics. 

I think there's a way to solve chess but everybody is looking at it all wrong.

Sometimes we forget to take the camera lens off.

Yeah, the main thing is to look at it from both sides of the board.

(That's actually almost serious).

I know this is 5 year old, but holy crap. This is like saying the sun is likely bigger than a tennis ball.

This angle hasn't been considered yet:

Let's assume, for the sake of argument, that these two HIGHLY UNLIKELY things happen:

1 - A solution to chess is found 

2 - It is something that can be grasped and memorized by the average player

How many of you would rush out and attempt to learn it?  Knowing that there really would be NO POINT to playing competitively any more, with things like the world championship obviously obsoleted and irrelevent, along with all of the "in between" tournaments to get there.

I know that I would personally go out of my way to NOT learn the solution... as I actually ENJOY playing chess!

Of course... I'm probably in the minority, as I have noticed that many people have a pathetic need to win at all costs... like unappreciative attention starved 5 year olds who only feel pleasure when they shame someone.

The second thing is not highly unlikely, it is simply impossible. (The first is infeasible without quantum computing).

A solution to chess at the very least needs to provide one move for every position that may be reached. Since the opponent has complete freedom, there are generally multiple plausible moves, meaning a solution needs to deal with a lot more positions (say around 10^40) than humans have neurons.

This was how draughts/checkers was solved: by dealing with every move by the opponent, but being selective on the moves by the player. Roughly speaking this reduces the size of the problem by a power of 0.5 (i.e. it is the square root of the size of the full problem, dealing with all possible games).