A player who accepts a midgame disadvantage in return for an endgame advantage has, at the present time, an inferior position. As we have seen, the currently-inferior position requires precise play in order to avoid loss; but this is different from the ordinary case of positional advantage, because in this case the player with the inferior position will eventually have the better game as long as he makes no mistakes.
It follows that it may be perfectly sound for one 1800 player to offer a gambit to another 1800 player, but it would not be sound for a grandmaster to offer the same gambit to another grandmaster.
In the game between the weaker players, the chance of the defender making a mistake and losing would be quite high, and so the gambit would be justified; but in the game between grandmasters, the chance of the defender playing perfectly and winning would be high enough to make offering the gambit a very bad wager. Let me phrase this more strongly: the same position may be a win for fish and a loss for GMs. That is, there exist positions such that in a fish-versus-fish game, both players of equal strength, one side will win most of the time; but in a GM-versus-GM game the other side will win most of the time.
From this we can draw a further conclusion that the many books of FIDE-chess opening theory which are based on grandmaster-level evaluations of positions are in many instances not correct for lower levels of play. For the 2100 player, opening monographs by minor masters may be more profitable than ECO.
In FIDE chess, the situation of present disadvantage being opposed to endgame advantage usually occurs only in the case of gambits or other speculative sacrifices. In games of chess with different armies, the starting position of the game may already present such a situation.
I've ranged a bit beyond the values of chesspieces here in order to investigate some relationships between material values, positional values, and the ELO ratings of chessplayers.
I have shown that all three of these factors are convertible into expected winning percentages, although each of these factors affects the winning percentage by a different mechanism.
The mechanisms by which they work have probably never before been examined so closely, and sometimes the knowledge of these mechanisms and how they work has allowed me to draw conclusions that can affect the way you play -- practical hints coming from such abstract ravings!
Therefore it is not unreasonable for me to hope that you have found this both interesting and profitable.
By all means, click on the mailme below and let me know you got this far, especially if you think you can add to the discussion, but even if you don't think so.
If you don't click the mailme, then writing this is like singing live on the radio: millions may hear, but it feels like shouting into the void.