One could also filter based upon pawn moves, double-moves, and captures; and by game phase; truly many things can be explored.
Patzer + Python "Analysis" + Uninspired Youtuber = Perfect Recipe...
But for what?
But for what?
@ThePurpleBenOni said in #3:
> Python "Analysis"
Data visualization is legitimate, and is accompanied by a proper analysis. Or is there any fault with it?
Yes, I'm the first to tear into anyone for lazily relying upon Stockfish evaluations, but these methods seem robust enough.
> Python "Analysis"
Data visualization is legitimate, and is accompanied by a proper analysis. Or is there any fault with it?
Yes, I'm the first to tear into anyone for lazily relying upon Stockfish evaluations, but these methods seem robust enough.
specifically moving the f pawn is something I only do if I know based on previous analysis that it works
(for example, in the exchange ruy lopez as black f6 is a theoretical move, or in some scotch lines you trap a piece with f5 and it's completely sound objectively)
but if there's any doubt or there's a cloud of "maybe" above the move - I do agree it's fine to follow the finegold and bortholemew principle of triple checking before moving that pawn and avoiding it in general
As for the data thing... I wouldn't want to have this data taken in to my considerations about which pawn to move... that's hard enough already :)
(for example, in the exchange ruy lopez as black f6 is a theoretical move, or in some scotch lines you trap a piece with f5 and it's completely sound objectively)
but if there's any doubt or there's a cloud of "maybe" above the move - I do agree it's fine to follow the finegold and bortholemew principle of triple checking before moving that pawn and avoiding it in general
As for the data thing... I wouldn't want to have this data taken in to my considerations about which pawn to move... that's hard enough already :)
In crazyhouse, moving anything but the d-and-e pawns is often downright losing, with some exceptions. The c pawn should not be moved prior to castling, if at all. This is because any pawn move leaves behind weaknesses.
It would be interesting to see this analysis repeated for crazyhouse. I expect the results to be even more drastic than standard chess. Pawn/piece drops would also be interesting. I might give it a try some day.
It would be interesting to see this analysis repeated for crazyhouse. I expect the results to be even more drastic than standard chess. Pawn/piece drops would also be interesting. I might give it a try some day.
So, the data is based on the inaccuracy of pawn moves?
Perhaps it would be helpful to include inaccuracy for failing to move the pawn--especially the f-pawn.
We might find that failure to move the f-pawn--when it is the best (or a far better) move--is even worse than inaccurate moves of the f-pawn.
(That is, compare inaccurate moves of the f-pawn vs. inaccurate failures-to-move the f-pawn.)
Perhaps it would be helpful to include inaccuracy for failing to move the pawn--especially the f-pawn.
We might find that failure to move the f-pawn--when it is the best (or a far better) move--is even worse than inaccurate moves of the f-pawn.
(That is, compare inaccurate moves of the f-pawn vs. inaccurate failures-to-move the f-pawn.)
i read fast.. but does this look at all such moves at any depth? I assume so. but just checking.
As it might affect the meaning of accuracy by engine analysis. or just be a chess variable that might have some influence.
the accuracy given the depth. Otherwise the averaging is over positions and depth, and players (and their ratings).
above post might be answering my other question.. of comparing with same precursor position doing or not doing the move.. look at that difference in score from the resulting positions.. relative scoring. Maybe that is what you did, but I could not figure it out..
I also think there is no need to use the word accuracy. Just use the actual measure of the SF score might be a better defined unit of measure (but that might just be me, don't mind it if so).
Otherwise interesting question, and use of lichess data. Normal use of engine for position scoring, though. (but do we have any choice... given engine tournament optimization objectives).
As it might affect the meaning of accuracy by engine analysis. or just be a chess variable that might have some influence.
the accuracy given the depth. Otherwise the averaging is over positions and depth, and players (and their ratings).
above post might be answering my other question.. of comparing with same precursor position doing or not doing the move.. look at that difference in score from the resulting positions.. relative scoring. Maybe that is what you did, but I could not figure it out..
I also think there is no need to use the word accuracy. Just use the actual measure of the SF score might be a better defined unit of measure (but that might just be me, don't mind it if so).
Otherwise interesting question, and use of lichess data. Normal use of engine for position scoring, though. (but do we have any choice... given engine tournament optimization objectives).
Привет классно очень это ужас круто пожалуйста можете ставить супер больше комментарий мне нравится ваши комментарии пожалуйста сейчас вы самый лучший игра носвети
Umm. This doesn’t look like it normalized the data? Pawns on 7th rank are much much rarer than pawns on the second rank.
So what is this analysis showing?
So what is this analysis showing?