Precision algorithm

Sort:
Sofidellavit

Hi folks! Anybody knows if it is possible to have absolutely equal precision of the game if both gamers didn't  use any program/engine? And if it is possible, then how works the algorithm? Because if it just counts the percentage of the best moves, it is rather impossible to arrive at the same precision at the end, considering there were 33 moves, so not a short game.

Thanks in advance

Vocaloid39

That doesn't really seem possible. Equal positions could happen mid-game, but to remain equal the entire game seems impossible. First of all, odd-numbered moves like you pointed out, or ply (one side moving before the other side moves) giving one color an extra move than the other. These details would sway the advantage even if just temporarily. Chess is also currently unsolved; meaning that no computer has worked out all of the possibilities, so it wouldn't even be feasible to know how to even continue an equal position, or how that would even look.

A similar question is with optimal play from both sides, is chess a draw? This is highly debated, but it's all theoretical.

Sofidellavit

sorry, there are 34 moves,not 33

Sofidellavit

So, when you tell chess is unsolved, you mean that neither using some engine is improbable to arrive at the same precision? The precision during my game was different (opening, mid-game,ending), but at the end I received absolutely equal total precision. I am very interested how it is possible. I hypothise, as the algorithm takes into account not only the percentage of the best moves, but also the opening and the complexity of the position then the probability is much higher. Because if we take into account only the best moves, the probability to have the same precision is like 1/(n^34), right? I would like to here from somebody who knows this algorithm, or maybe somebody have had the same situation?

blueemu

The "precision" or "accuracy" calculation used by this chess site has only a tangential connection to the actual value of the moves.

A game evaluated at 80 accuracy for a 700 rated player might only be assigned 50 accuracy for a 1600 rated player.

Sofidellavit

So what does it mean? Which probability is to have absolutely equal precision for the game for both gamers with different rating?

blueemu
Sofidellavit wrote:

So what does it mean? Which probability is to have absolutely equal precision for the game for both gamers with different rating?

If you are talking about the percentage accuracy stat, then none.

That isn't its purpose. It's basically a feel-good stat.

I'm not sure that I even understand your question.

How do you define "precision"? Move matches to the computer's own top choice? But that changes with every new version of the computer software... does that mean that the definition of precision changes, and a move that was "accurate" is no longer accurate?

Sofidellavit
blueemu wrote:

I'm not sure that I even understand your question.

I've done the game review and received equal total accuracy for the game for both opponents, and equal game rating. The ratings of the opponents is different. So I am interesting to know is it very rare case?

birkiya

Hi

blueemu

I wouldn't take either of those numbers very seriously.

Sofidellavit
blueemu ha scritto:

I wouldn't take either of those numbers very seriously.

it's interesting for me if it possible to have the same accuracy by chance, or my partner did use some "trick". I don't take this numbers seriously