I've been making one EPD file per ebook. But I wonder if I should be using one EPD and save all my ebooks into it.
If I add new positions to an ebook and then save it to an existing EPD, appending positions to the end of the EPD, does that save any time in the "overnight analysis"? In other words, does the analysis skip over the positions already analyzed and just do the new positions?
If the answer to that last question is "yes", then should I really be saving always to the same EPD file, regardless of which ebook I'm saving from?
Confused about EPD files -- do I need just one?
-
jenesuispasdave
- Posts: 6
- Joined: Sun Dec 19, 2021 6:38 pm
-
jenesuispasdave
- Posts: 6
- Joined: Sun Dec 19, 2021 6:38 pm
Re: Confused about EPD files -- do I need just one?
Forgot to mention that I'm on a Mac, using COW Professional build 161.
-
FreeRepublic
- Posts: 179
- Joined: Wed Jan 23, 2019 12:31 am
Re: Confused about EPD files -- do I need just one?
Like you, I am not Dave and I too am confused about EPD files. While PGN seems like old hat, EPD is very new to me. My use has been limited to:
A) Evaluating final (leaf node) positions of 400 games. I did this because some games did not have results. A second reason is that some games may have been decided on time, rather than the final position.
B) Evaluating all positions of one game. I did this to speed-up the search for novelties and errors. It does the job, but not exactly as I had hoped. I see the Stockfish analysis of each position. Often SF chooses the move played. Sometimes it chooses a different move. Looking it over, move by move, I can spot an improvement.
What I was shooting for in B) was to have Stockfish find novelties automatically. So I decided to use the command, "Find Novelties by Assessment."
First you create an EPD file that contains all positions in the game. Then you have COW analyze all positions in the EPD file. Then issue the command to Find Novelties. COW does this and names every position containing a novelty. You can then jump to each position. It sounds promising, but there is a hiccup.
Some of the named positions are mirrored left to right! For example, White castles queen side. His king should be on the left side of the board. Instead, it is on the right side of the board. And so it goes for every piece and pawn in that position.
Some of the named positions are reversed! The White pieces are black and vice-versa.
A) Evaluating final (leaf node) positions of 400 games. I did this because some games did not have results. A second reason is that some games may have been decided on time, rather than the final position.
B) Evaluating all positions of one game. I did this to speed-up the search for novelties and errors. It does the job, but not exactly as I had hoped. I see the Stockfish analysis of each position. Often SF chooses the move played. Sometimes it chooses a different move. Looking it over, move by move, I can spot an improvement.
What I was shooting for in B) was to have Stockfish find novelties automatically. So I decided to use the command, "Find Novelties by Assessment."
First you create an EPD file that contains all positions in the game. Then you have COW analyze all positions in the EPD file. Then issue the command to Find Novelties. COW does this and names every position containing a novelty. You can then jump to each position. It sounds promising, but there is a hiccup.
Some of the named positions are mirrored left to right! For example, White castles queen side. His king should be on the left side of the board. Instead, it is on the right side of the board. And so it goes for every piece and pawn in that position.
Some of the named positions are reversed! The White pieces are black and vice-versa.