PGN Annotator 1.0 released

Chess Artist is a very nice program to annotate PGN games. Annotate follows a different path, it stores and keeps the analysis results in a multi functional database, the Polyglot book format. The current main Polyglot book database contains over 51 million positions with the scores and depths from top engines and with that we can annotate (your) PGN games in a blink of an eye.

...

Find Position - also direct reachable via the [F3] function key is a powerful feature to 1) instantly find a position (as EPD taken from the clipboard) in the 51 million Book Database, 2) analyze this position with your favorite UCI chess engine (default is Stockfish 12) and 3) post the outcome on a (chess) forum with a simple copy&paste. Example. To configure (change) engines press [F1]. Note, if the position is not in the Book Database the analysis and forum post functions still work.

A few examples to highlight the Find Position | Analyze | Post option.
2r2rk1/pp3pp1/4bb1p/q2p1P1Q/3P4/2N5/PP4PP/1K1R1B1R b - - bm Rxc3; c1 "Pillsbury - Lasker 1896";  [ view ]
b2r3r/k4p1p/p2q1np1/NppP4/3p1Q2/P4PPB/1PP4P/1K1RR3 w - - bm Rxd4; c1 "Kasparov - Topalov 1999"; [ view ]
2rq2k1/4bppp/p1rp4/1p1NpP2/4P3/2PQ4/PP4PP/3R1R1K w - - bm Ra1!; c1 "Fischer - Olicio Gadia 1960"; [ view ]

http://rebel13.nl/rebel13/pgn-annotator.html

Version 1.1

Added two new features -

1. Create selected opening books - from a large PGN create a new PGN with the opening moves of your choice. For instance, create (or open) the textfile opening.txt and type: "e2e4 e7e5 g1f3 d7d6 ", save and press [F6] select the PGN and all the games of the Pirc-Defence are stored in opening.pgn

2. Create an analyzed Polyglot opening book from analyzed EPD files - as long as the EPD's are valid and contain the bm | ce  and acd tags this will work. Press [F7] select the EPD and a corresponding Polyglot *.bin book is created, i.e. test.epd becomes test.bin.

http://rebel13.nl/rebel13/pgn-annotator.html

I like the idea to blunder check pgn games with an analyzed bin book with Annotator. That way an engine doesn't has to go over the same opening lines again and again when blunder checking pgn games.
Is it possible somehow to change the threshold, annotator uses to blunder check the games?
For how bad a move has to be to get the annotation ??, ? or !?
And also for how good a move has to be to get !

Jonathan003 wrote:

I like the idea to blunder check pgn games with an analyzed bin book with Annotator. That way an engine doesn't has to go over the same opening lines again and again when blunder checking pgn games.
Is it possible somehow to change the threshold, annotator uses to blunder check the games?
For how bad a move has to be to get the annotation ??, ? or !?
And also for how good a move has to be to get !

From the page:

Code:

Remarks
1. There are 5 annotation symbols.
?! - when score drops with 0.50
?  - when score drops with 1.00
?? - when score drops with 2.00 or more
!  - when score increases with 1.00
!! - when score increases with 2.00 or more

What would be the best settings for analyzing the bin books with poly 1.7b to accurate blundercheck the pgn games where the bin book is created from, with annotator 1.2?
I used the option W to only analyze the white moves.
And I used these options:


A get a lot of false blunder alerts (?), when blunder checking the pgn, with annotator 1.2
The moves with the annotation (??), are correct and are indeed very bad moves. But most moves annotated with (?), are not blunders at all, most of the times it are the best moves to play in the position. Also the most moves annotated with ?! are not dubious moves at all.
So, something goes wrong.

Here are some examples

Code:

[Event "?"]
[Site "?"]
[Date "????.??.??"]
[Round "1"]
[White "?"]
[Black "?"]
[Result "1/2-1/2"]
[PlyCount "23"]

1.d4 {[%eval 21,27] Unkown Best e4 0.22/26} 1...Nf6 2.c4 e6 3.Nf3 d5 4.Nc3
c6 5.e3 Nbd7 6.Bd3 {[%eval 10,26] Unkown Best Qc2 0.27/25} 6...dxc4 7.Bxc4
b5 8.Bd3 a6 9.e4 c5 10.d5 Bb7 11.dxe6? {[%eval -172,21] Unkown Best O-O
0.12/24} 11...fxe6 12.Bc2 1/2-1/2

The move 11.dxe6 is evaluated as a blunder and has a score of -1.72 if I do the analysis in Chessbase 11.dxe6 is the best moves with an evaluation of +0.27

Code:

[Event "?"]
[Site "?"]
[Date "????.??.??"]
[Round "1"]
[White "?"]
[Black "?"]
[Result "1/2-1/2"]
[PlyCount "34"]

1.d4 {[%eval 21,27] Unkown Best e4 0.22/26} 1...Nf6 2.c4 e6 3.Nf3 d5 4.Nc3
c6 5.e3 Nbd7 6.Bd3 {[%eval 10,26] Unkown Best Qc2 0.27/25} 6...dxc4 7.Bxc4
b5 8.Bd3 a6 9.e4 c5 10.d5 Qc7 11.dxe6? {[%eval -102,22] Unkown Best O-O
-0.05/ 24} 11...fxe6 12.O-O! {[%eval 13,23] Unkown} 12...Bb7 13.Bc2 Bd6
14.Ng5 Nf8 15.f4 O-O-O 16.Qe2 {[%eval -8,25] Unkown Best Qe1 0.08/23}
16...h6 17.Nf3 Bxf4 1/2-1/2

The move 11.dxe6 is evaluated as a blunder and has a score of -1.22. If I do the analysis in Chessbase with Stockfish 13 I see it is one of the best moves with a score of 0.00

1. It's unclear to me which book and engine you were using, it says unknown?
2. Comparing the result between 2 engines is (the %eval suggest a chessbase engine) is like comparing apples with oranges.

Anyway I ran your first example with the default settings and there are no good or bad annotation remarks.

Code:

[Event ""]
[Site ""]
[Date "????.00.49"]
[Round "1"]
[White ""]
[Black ""]
[Result "1/2-1/2"]
[BlackElo ""]
[WhiteElo ""]

1.d4 {0.21/24 SF12 Best e4 0.34/23} Nf6 {-0.31/24 SF12 Best c6 -0.23/24} 2.c4 {0.30/23 SF12 Best Nf3 0.35/23} e6 3.Nf3 {0.13/23 SF12 Best Bg5 0.32/16} d5 {-0.21/24 SF12 Best Be7 -0.16/23} 4.Nc3 {0.17/23 SF12 Best Bg5 0.33/23} c6 {-0.22/24 SF12 Best Bb4 -0.16/23} 5.e3 Nbd7 {-0.35/24 SF12 Best c5 -0.28/23} 6.Bd3 dxc4 {-0.25/24 SF12 Best b6 -0.12/28} 7.Bxc4 b5 {-0.37/24 SF12 Best Be7 0.00/1} 8.Bd3 {0.25/23 SF12 Best Be2 0.30/24} a6 {-0.61/24 SF12 Best Bd6 -0.25/22} 9.e4 c5 10.d5 Bb7 {-0.13/23 SF12 Best exd5 0.28/27} 11.dxe6 {0.53/23 SF12 Best O-O 0.54/23} fxe6 12.Bc2 1/2-1/2

I accidentally deleted the files, so I can't check anymore. I used stockfish 13 and a bin book with only analyzed moves for white.
The engine I used had a long name 'stockfish_13_win_x64_bmi2'.
Maybe this is why it says unknow? Maybe I better change the name of the engine to something short like SF13.
I will try it again to see what could be the problem.
Maybe it is a problem with my pc?
I use a HP all in one and let it run overnight. The fans make loud noises.
Maybe my hp all in one pc just can't handle such analysing tasks.

This is my hardware

Windows-versie
Windows 10 Home
2018 Microsoft Corporation. Alle rechten voorbehouden.
Systeem
Processor: Intel(R) Core(TM) i7-4770T CPU 2.50GHz 2.50 G Hz
Installed memory 16,0 GB (15,9 GB available)
(RAM):
Type systeem:
64-bits besturingssysteem, x64-processor

Your PC is fine. Let's solve one mystery. You have a book analyzed with SF13 but did not code the book as SF13 (option Engine Coding), hence you got the "Unknown Engine" message. Open the file "engines.txt", you get -

Code:

0 = Unkown = Unknown-Engine
1 = SF12 = Stockfish-12
2 = KD = Komodo-Dragon
3 = H6 = Houdini-6
4 = K14 = Komodo-14
5 = Lc0 = Lc0
6 = CCRL = CCRL-ELO-3200+
7 = Other = Strong-Engine
8 = TCEC = TCEC-Engine
9 = CCC = CCC-Engine

Easy fix - change 0 = Unkown = Unknown-Engine to 0 = SF13 = Stockfish-13

Are these good settings to use for making an analyzed bin book for only the white moves with Poly 1.7b

Code:

Divide EPD over threads : 4
Hash Table Mb : 512
Maximum Time in ms : 10000
Depth (0=default) : 29

(S)tart (E)ngine (T)hreads (H)ash (L)evel (D)epth (Q)uit

Or is it better to use these

Code:

Divide EPD over threads : 4
Hash Table Mb : 512
Maximum Time in ms : 20000
Depth (0=default) : 35

(S)tart (E)ngine (T)hreads (H)ash (L)evel (D)epth (Q)uit

I want to be sure that's why I ask. Because I don't want to waist time if I do the overnight analysis again, by using the wrong settings.

Both are fine.