PDA

View Full Version : Experimenting opening/game analysis with engine.



HAHAHAHA
08-08-2021, 09:12 PM
Hi All,
I am keen to play out my analysis and post mortem ideas from my long play games against a sparring partner.

Using an engine, can someone more experienced explain how best to do it.
Most chess website engines will only let me start the game from the initial position.

Analyzing the game in Chess base, game analysis style will suffice at this stage but lacks the "playing feel"

Thanks guys in advance for suggestions.


P.S (of course if a Melburnian 1800+ is wishing to analyse with me, that would be great too)

Patrick Byrom
09-08-2021, 08:21 PM
I don't know too much about either Chess.com or li chess.org, but they both seem to allow this. On Chess.com, you can play a daily game with a specific opening (https://www.chess.com/forum/view/chess-lessons/if-you-want-to-practice-a-specific-opening-how-can-you-do-it-like-play-games-50445472), and on lichess you can do this as long as you're playing a 'friend' (https://lichess.org/forum/lichess-feedback/playing-a-particular-opening).

blackbishop
19-09-2021, 03:29 PM
Perhaps using a program such as ChessX, with say Stockfish 14 installed, to house a chess opening repertoire is a common practice.
It just dawned on me recently that there are many advantages to doing this rather than using a Word Processor.

I started a new database in ChessX and a new game but imputed opening variations only of the Kings Indian. You can input multiple branching lines.
On revising a particular line which comes to a branching line, you make the required branch move and the computer will jump to the required branch.
All the while the computer will be analysing the positions. If an improvement is suggested you can input the new line and delete the old line using the right mouse button.

Don't forget to save the game and save the database before closing. Start a new game for a different opening or different variation within the same opening. There are a number of things you an do to improve the readability
of the "games". In Game => Edit Tags, I put a description and a few moves indicating the content of a "game" in the "Event" section the other sections being vacant. You can drag the Event section to the left so that only it is displayed.

If you need to, you can edit your file of openings by renaming the pgn database file to a text file and editing the text file. When editing is finished, rename the text file back to a pgn. Example KIFianchetto.pgn => KiFianchetto.txt => KIFianchetto.pgn
Edit a copy of your file and keep the original just in case. I used editing to separate different openings in the same database into several different databases. You will need to experiment to find the best solution for you.

1Nf3
22-09-2021, 03:52 AM
Perhaps using a program such as ChessX, with say Stockfish 14 installed, to house a chess opening repertoire is a common practice.
It just dawned on me recently that there are many advantages to doing this rather than using a Word Processor.

I started a new database in ChessX and a new game but imputed opening variations only of the Kings Indian. You can input multiple branching lines.
On revising a particular line which comes to a branching line, you make the required branch move and the computer will jump to the required branch.
All the while the computer will be analysing the positions. If an improvement is suggested you can input the new line and delete the old line using the right mouse button.

Don't forget to save the game and save the database before closing. Start a new game for a different opening or different variation within the same opening. There are a number of things you an do to improve the readability
of the "games". In Game => Edit Tags, I put a description and a few moves indicating the content of a "game" in the "Event" section the other sections being vacant. You can drag the Event section to the left so that only it is displayed.

If you need to, you can edit your file of openings by renaming the pgn database file to a text file and editing the text file. When editing is finished, rename the text file back to a pgn. Example KIFianchetto.pgn => KiFianchetto.txt => KIFianchetto.pgn
Edit a copy of your file and keep the original just in case. I used editing to separate different openings in the same database into several different databases. You will need to experiment to find the best solution for you.

Thanks for the tip on ChessX. Always on the lookout for something chess useful I haven't already got and have now downloaded it.

1Nf3
22-09-2021, 04:14 AM
I have started to notice some rather glaring differences in game analysis assessment between Lichess and Chess.com. I'm sure I'm not the first to have picked up on this.

Lichess uses Stockfish 14 and I understand Chess.com also uses Stockfish but perhaps an earlier version? Please correct me if I am wrong.

Here's a game I played recently and note the difference in assessment between Lichess and Chess.com :

1. Nf3 Nf6
2. c4 e6
3. Nc3 Bb4
4. Qc2 0-0
5. a3 Bxc3
6. Qxc3 d6
7. d4 b6
8. Bg5 h6
9. Bh4 g5
10. Bg3 Ne4
11. Qd3 Bb7
12. h4 Nxg3
13. fxg3 f5 - at this point Lichess has white +4.0 and Chess.com has white +0.75
14. hxg5 - at this point Lichess has white +3.7 and Chess.com has white +2.2
14. ...Be4
15. Qe3 Qd7
16. gxh6 Kh8 - at this point Lichess has white +3.0 and Chess.com has white +3.15

Can anyone shed any light on the discrepancies ???

Ian Rout
22-09-2021, 11:06 AM
Can anyone shed any light on the discrepancies ???There are only three correct evaluations of a position - White or Black has a forced mate, or it burns out to a draw. If you could calculate every permutation of moves from a position you would know which one applies.

Even a computer can't do that so it looks as far as it can and then bean-counts the position at the end - 100 points for an extra pawn, less 23 for a bad Bishop, plus 16 for an out post on c6, etc. It then reports the outcome if both players choose the best outcome at each turn. Different programs with different parameters or depths, or even the same program with more time or memory and so looking longer (= deeper) will give different evaluations.

1Nf3
27-09-2021, 12:11 PM
There are only three correct evaluations of a position - White or Black has a forced mate, or it burns out to a draw. If you could calculate every permutation of moves from a position you would know which one applies.

Even a computer can't do that so it looks as far as it can and then bean-counts the position at the end - 100 points for an extra pawn, less 23 for a bad Bishop, plus 16 for an out post on c6, etc. It then reports the outcome if both players choose the best outcome at each turn. Different programs with different parameters or depths, or even the same program with more time or memory and so looking longer (= deeper) will give different evaluations.

Cheers for that.

Seems computers have 'opinions' just as people do.