pls can anybody tell
You don't. Because they don't make much sense from a logical point of view. You can only ever make the evaluation worse, not better.
If you look at what the green pawn does... it is rather to please the users and make them feel great. The algorithm behind it seems to be quite unsophisticated and rather meaningless.
It has been discussed a couple of times in the Feedback Section here, maybe you can find something using the search.
If you look at what the green pawn does... it is rather to please the users and make them feel great. The algorithm behind it seems to be quite unsophisticated and rather meaningless.
It has been discussed a couple of times in the Feedback Section here, maybe you can find something using the search.
Well, the point is this: a certain move is labelled the 'best move'...
until a smart guy, who could be anyone, including YOU by the way, finds a better move.
So... the lesson is, never call a move, the 'best move'.
until a smart guy, who could be anyone, including YOU by the way, finds a better move.
So... the lesson is, never call a move, the 'best move'.
@PawnSprint said in #3:
> Well, the point is this: a certain move is labelled the 'best move'...
> until a smart guy, who could be anyone, including YOU by the way, finds a better move.
> So... the lesson is, never call a move, the 'best move'.
No, this is not the problem. You can and would only assign "!" and "!!" to what you see and know, and the 3600+ Elo engine is probably good enough.
The problem is to come up with a sensible definition of a good or brilliant move. You simply cannot see that from the evaluation. The computer also doesn't know what moves are difficult to find and what are not. For the computer, a move is a move. The evaluation is based on the best line anyway, so you can only detect mistakes that way.
> Well, the point is this: a certain move is labelled the 'best move'...
> until a smart guy, who could be anyone, including YOU by the way, finds a better move.
> So... the lesson is, never call a move, the 'best move'.
No, this is not the problem. You can and would only assign "!" and "!!" to what you see and know, and the 3600+ Elo engine is probably good enough.
The problem is to come up with a sensible definition of a good or brilliant move. You simply cannot see that from the evaluation. The computer also doesn't know what moves are difficult to find and what are not. For the computer, a move is a move. The evaluation is based on the best line anyway, so you can only detect mistakes that way.
@nadjarostowa
yes, the computer can label a certain move the 'best move', even within a line which was praised by Tartakower and was considered as 'good' or 'winning' in 1920, but is now considered as bad or losing, and isn't played anymore, except by beginners.
So, I maintain that the label 'best move' is very relative, can be some sort of rubbish, and overall doesn't make much sense.
yes, the computer can label a certain move the 'best move', even within a line which was praised by Tartakower and was considered as 'good' or 'winning' in 1920, but is now considered as bad or losing, and isn't played anymore, except by beginners.
So, I maintain that the label 'best move' is very relative, can be some sort of rubbish, and overall doesn't make much sense.
@PawnSprint
With stronger engines, or simply deeper analysis, sometimes evaluations change. That's true, but completely besides the point here. Nobody wants to label "best moves".
If you annotate a game, you do it to the best your current abilities, and take that as truth. Nothing else makes sense, or you simply don't annotate at all. It simply leads to corrections in the future, like in all fields.
But: People want "good" and "brilliant" moves. And even if the evaluation of positions would never ever change, this still would not be possible, because for the engine an icredibly difficult to find move looks just exactly the same as on obvious only-move.
With stronger engines, or simply deeper analysis, sometimes evaluations change. That's true, but completely besides the point here. Nobody wants to label "best moves".
If you annotate a game, you do it to the best your current abilities, and take that as truth. Nothing else makes sense, or you simply don't annotate at all. It simply leads to corrections in the future, like in all fields.
But: People want "good" and "brilliant" moves. And even if the evaluation of positions would never ever change, this still would not be possible, because for the engine an icredibly difficult to find move looks just exactly the same as on obvious only-move.