Fiddling with the isolani
What is the best place for white's light squared bishop?
 |
| Black to move. Green = positive Red = negative |
r1bq1rk1/pp2bppp/2n1p3/3n4/2BP4/2N2N2/PP3PPP/R1BQR1K1 b - - 2 10
This is a position from the game Vadim Zvjaginsev - Rodrigo Vasquez, World Rapid Ch., Khanty-Mansiysk 2013. White has just played 10.Re1. I let Stockfish think for 27 ply at average.
The best place of whites light squared bishop is apparently c2. From there it exerts influence on both the diagonal b1-h7 and d1-a4
The two highest values are b1 and c2, indicating that b1-h7 is the most important diagonal.
It is striking that the bishop on c4 isn't better placed than on f1
Furthermore, the bishop on h3 is worse than the bishop on h1.
h1 is some kind of benchmark for this position, where a bishop would exert hardly any influence on the position at all.
 |
| Black to move. Green = positive Red = negative |
Same position, but now for the dark squared bishop. It is remarkable that there is no better place than Bg3. It cannot be improved on the diagonal it sits on.
 |
| Same position Black to move. Numbers for Nc3 |
Practically, Nc3 is on its best position. The reference square h1 where the knight doesný exert any influence on the position whatsoever, indicates that Nc3 can be rated as +0.56
 |
| Black to move Same position. Nf3 |
It seems that Nf3 is on its best position too, from a practical point of view.
 |
| Black to move Same position All minor pieces removed scores +0.41 |
I read somewhere the advice for black to keep the heavy pieces on the board in order to put pressure on e4, thus binding the white pieces to the defense of e4.
 |
| Black to move Same position |
When black is to move, the position scores +0.24 for black. When it is white to move, it scores -0.04
 |
| Black to move Same position Numbers for the white queen |
I only entered the numbers for the squares where the white queen is better. Note the reference square h1.
Same scheme for the dark squared bishop added
ReplyDeletePost updated with Nc3
ReplyDeletePost updated with Nf3. If anyone feels the inclination to help, feel free to do the same for black.
ReplyDeleteI will be at least 18 hours further before I am ready to start with black
DeletePlease describe the process you ae using to get the scores for each square. I am not aware of this feature in Chess Tempo or lichess.
DeleteI use Arena with Stockfish. Let it run for 25-30 ply. Write the value down on a piece of paper. Then I choose "set up position". Then I remove the piece and put it on the next position. In the end I copy the diagram to Paint and add the numbers.
DeleteIt would be great if you can do black. That way I can focus on more positions from whites perspective.
DeleteWhat setting are you using for the "# Lines" to analyze?
DeleteWhat criteria do you use to determine which squares to consider putting the piece on? Is it just shifting one piece at a time to the empty squares in the initial position?
Have you considered using the Stockfish Evaluation Guide?
LINK: https://hxim.github.io/Stockfish-Evaluation-Guide/
It’s not just information about the Stockfish evaluation functions. You can enter a FEN position at the top-right of the screen and then drill down through the various evaluation factors for that specific position. It provides an instantaneous evaluation for those factors in the position and identifies the pieces used for that evaluation in aggregate without a search.
Since it appears that you are looking for “hints” regarding the best location for the pieces relative to the isolani, this might speed up the process considerably.
There is an overall (main) evaluation and additional macro-level evaluations:
Main evaluation
Middle game evaluation
End game evaluation
Scale factor
Phase
Tempo
Rule50
Drilling down, there are the detailed evaluations, with sub-valuations under them.
Attack
-Pinned direction
-Knight attack
-Bishop xray attack
-Rook xray attack
-Queen attack
-Pawn attack
-King attack
-Attack
-Queen attack diagonal
-Pinned
Helpers
-Rank
-File
-Bishop count
-Queen count
-Pawn count
-Knight count
-Rook count
-Opposite bishops
-King distance
-King ring
-Piece count
-Pawn attacks span
Imbalance
-Imbalance
-Bishop pair
-Imbalance total
King
-Pawnless flank
-Strength square
-Storm square
-Shelter strength
-Shelter storm
-King pawn distance
-Check
-Safe check
-King attackers count
-King attackers weight
-King attacks
-Weak bonus
-Weak squares
-Unsafe checks
-Knight defender
-Endgame shelter
-Blockers for king
-Flank attack
-Flank defense
-King danger
-King mg
-King eg
Material
-Non pawn material
-Piece value bonus
-Psqt bonus
-Piece value mg
-Piece value eg
-Psqt mg
-Psqt eg
Mobility
-Mobility
-Mobility area
-Mobility bonus
-Mobility mg
-Mobility eg
Passed pawns
-Candidate passed
-King proximity
-Passed block
-Passed file
-Passed rank
-Passed leverable
-Passed mg
-Passed eg
Pawns
-Isolated
-Opposed
-Phalanx
-Supported
-Backward
-Doubled
-Connected
-Connected bonus
-Weak unopposed pawn
-Weak lever
-Blocked
-Doubled isolated
-Pawns mg
-Pawns eg
Pieces
-Outpost
-Outpost square
-Reachable outpost
-Minor behind pawn
-Bishop pawns
-Rook on file
-Trapped rook
-Weak queen
-King protector
-Long diagonal bishop
-Outpost total
-Rook on queen file
-Bishop xray pawns
-Rook on king ring
-Bishop on king ring
-Queen infiltration
-Pieces mg
-Pieces eg
Space
-Space area
-Space
Threats
-Safe pawn
-Threat safe pawn
-Weak enemies
-Minor threat
-threat
-Hanging
-King threat
-Pawn push threat
-Slider on queen
-Knight on queen
-Restricted
-Weak queen protection
-Threats mg
-Threats eg
Winnable
-Winnable
-Winnable total mg
-Winnable total eg
For example, under Pieces→Pieces mg (middlegame), it gives the “Per square values” for various pieces, individually and as an aggregate. There is a different evaluation “Pieces→Pieces eg” for the endgame.
""What setting are you using for the "# Lines" to analyze?
DeleteWhat criteria do you use to determine which squares to consider putting the piece on? Is it just shifting one piece at a time to the empty squares in the initial position?
Have you considered using the Stockfish Evaluation Guide?
I use no setting. I stop when the value doesn't change much. With a minimum of 25 ply.
Yes, just shifting the piece of concern from square to square.
I didn't thought of the Stockfish Evaluation Guide. I don't know whether the method I try to invent is viable in the long run. For now, I accept handwork.
It has been a long time ago that I looked into the Stockfish Evaluation Guide.
Somehow I didn't get a notification for this comment. (again!)
I added two more positions.
ReplyDeleteThis is from the intro of the course:
ReplyDeleteI first met Mauricio ten years ago, as we both competed in Pan-American youth competitions. Later we both received a chess scholarship to attend the University of Texas at Brownsville. We were teammates for four years and spent countless hours studying chess together, though our training preferences were very different. Mauricio read books, analyzed his games and prepared openings. I did these things too, but in reality, the vast majority of my time was spent looking over current chess games and playing. I was surprised when Mauricio told me he had written a book partially inspired by my training methods, and I was certainly interested to see what was in it. The truth about my training method is that looking over a game for just a couple of minutes can actually be a wonderful investment, if done correctly. The key is searching for repeating patterns; this takes some practice but is feasible. In my career, I have seen close to 100,000 chess games, including most of the grandmaster-level games played over the past decade. The cumulative experience from spending a minute or two on each of these games has allowed me to gain an excellent positional understanding. Staring at a position for a few seconds is often enough for me to see who is better, which plans will work, which pieces should be traded, etc. Acquiring such a level of experience and positional knowledge requires many years. Going through thousands and thousands of games takes a very long time, even if you only spend a couple of minutes on each. Most importantly, being able to actually see the patterns does not come easily to everyone. Addressing these two difficulties is exactly the purpose of this book.
I added the numbers for the white queen.
ReplyDelete