Author: Robert Hyatt
Date: 06:26:43 10/20/00
Go up one level in this thread
On October 20, 2000 at 01:00:07, Ratko V Tomic wrote: >> IE if my program plays Rc6 and I can prove it is correct, I am happy. >> If I can prove it is bad, even though it won the game, I am not happy. >> If I can't prove it either way, I am concerned. That was the point >> here. I want my fate in my hands, not resting on whether my >> opponent overlooks something or not. > >You are idealizing ability of risk-averse programs. If it were tic-tac-toe >you can prove move is correct. But in chess, just because some hand-put >tangle of evaluation terms gives, say, 0.3 pawns more for move A than >for other moves B, C,... you haven't proven move A is correct. It is >only "correct" within the model game (-tree) your program substitutes >for the full chess tree (where every leaf is win, draw, loss). You are making the assumption that "heuristics" cannot be "accurate". I can give you lots of examples where this is a false assumption. IE try to play a simple k and p vs k ending against Crafty. With no tablebases. It only takes a few heuristics to play this perfectly, as any good endgame book whill explain. > >And only the final outcomes (and lots of them) can tell you which toy >model of the game simulates the real game best. That is the criteria >not only for some complex positional terms, but for every term, as much so >for Knight = 3 Pawns as for "this particular king attack" = 3 pawns. > >There is no rule of the game which lets you "cash in" at will your >Knight for 3 pawns, or the other way around, just as there is no >rule letting you "cash in" some king-attack poise for 3 pawns. Both >figures 3 are pure constructs of the respective models, they're little >wheels in a toy which is trying to simulate the real thing. > >So, Crafty is only "correct" or "accurate" in following its model game, while >Gambit Tiger is as "correct" or "accurate" in following its own model game. The >two are two different model games (somewhat similar, well, yes), and neither >model game is the full chess tree (not even close). And whichever one beats the >other more that one has better model of the game, the model overall closer to >the object it models. > >From this more abstract perspective your objections to GT's "risk taking" is of >this kind: I see that odd wheel in that toy model, and if I were to put it into >my toy model (or any model I understand or can imagine) it would wobble and slip >so much that my whole toy model would fall apart. Therefore, that is a bad >little wheel, and the whole model which has it can't be very good or solid. The >only thing that really follows is that it's a "bad little wheel" if it were >transplanted into your model game, not necessarily bad for Gambit Tiger's model >game, much less for all other possible model games simulating chess. I don't know whether the speculativeness of GT is good or bad. I simply brought up the possibility that if a program 'gambits' away positional advantages for other positional advantages, it is perhaps not a game-winning/game-losing decision. But if you gambit away a piece, a mistake will change the game outcome significantly...
This page took 0.01 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.