The Baron 1.7.0 FRC bug

[Matthias Gemuh]

The Baron 1.7.0 FRC does not understand that O-O is here valid

[D]1r3rk1/3b1ppp/8/4nP2/3NP3/q1P3P1/2Q2N1P/2R3KR w K -


Matthias.

[pijl]

The Baron 1.7.0 FRC does not understand that O-O is here valid

[D]1r3rk1/3b1ppp/8/4nP2/3NP3/q1P3P1/2Q2N1P/2R3KR w K -

Not exactly. When you feed this position to the Baron 1.7.0 FRC it knows it is valid. The problem was one move earlier where the rook moved from e1 to c1. Due to an error in the move parser it was interpreted as O-O-O (which would be a legal move as well), so from its own perspective it was correct in not allowing O-O.
I've already promised Ray to fix this and release a Baron 2.23 FRC this week, probably tomorrow.
Richard.

[Spock]

I'm really grateful to engine authors that continue to support FRC. It has a very small fan base, and understandably many authors either don't think it is worth the time to do it, or don't like FRC and don't do it. So thanks to Richard and all the other authors who enable me to pursue my interest in it.

A big thanks as well to Matthias. His ChessGUI enables me to play winboard and UCI FRC engines natively against each other, which I was not able to do before. Now I have quite a few engines to add to the FRC ratings list. The Baron 1.7.0 has been the first, and I have 500 games for that already. The above game was the only problem I encountered, both it and the GUI performed perfectly. It is often the case with small bugs like this that you may never hit them. But with the hundreds and often thousands of games that each engine plays over time, I usually strike them. The FRC castling rules are difficult for authors (heck, I don't understand them fully myself).

[Matthias Gemuh]

The Baron 1.7.0 FRC does not understand that O-O is here valid

1r3rk1/3b1ppp/8/4nP2/3NP3/q1P3P1/2Q2N1P/2R3KR w K -

Not exactly. When you feed this position to the Baron 1.7.0 FRC it knows it is valid. The problem was one move earlier where the rook moved from e1 to c1. Due to an error in the move parser it was interpreted as O-O-O (which would be a legal move as well), so from its own perspective it was correct in not allowing O-O.
I've already promised Ray to fix this and release a Baron 2.23 FRC this week, probably tomorrow.
Richard.

Sorry for my wrong guess

All my FRC engines have a yet unfixed castling bug : if castling King is between its To-Square and nearer board side, and the To-Square of Rook is under enemy attack, castling is skipped
Shall be fixed.

Matthias.

[pijl]

I've added the FRC version of the Baron 2.23 on the Baron's Lair. It is very similar to the regular version of the Baron, except that a few FRC related bugs (like the one mentioned in this thread) that were discovered in the Baron 1.7.0 are solved.
Have fun,
Richard.

[slobo]

I've added the FRC version of the Baron 2.23 on the Baron's Lair. It is very similar to the regular version of the Baron, except that a few FRC related bugs (like the one mentioned in this thread) that were discovered in the Baron 1.7.0 are solved.
Have fun,
Richard.

Hi Richard,

I tried two games with FRCbaron within HM winboard, but I foud some problems.

1. First game

[Event "Computer Chess Game"]
[Date "2009.01.31"]
[Round "-"]
[White "Glaurung 2.2 JA"]
[Black "The Baron 2.23 (FRC)"]
[Result "1-0"]
[TimeControl "40/300"]
[Variant "fischerandom"]
[FEN "rnqkbbnr/pppppppp/8/8/8/8/PPPPPPPP/RNQKBBNR w HAha - 0 1"]
[SetUp "1"]

{--------------
r n q k b b n r
p p p p p p p p
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
P P P P P P P P
R N Q K B B N R
white to play
--------------}
1. d4 d5 2. Qf4 Nc6 3. Nc3 Nf6
{False illegal-move claim} 1-0

Glaurund wanted to play O-O-O.

2. Second game

Reset(1, 1) from gameMode 0
recognized 'fischerandom' (-1) as variant fischerandom
GameEnds(0, (null), 2)
85313 >first : force
85313 >first : ping 2
shuffleOpenings = 1
TC string = '+40/5'
mps=40 tc=300000 inc=0
TC string = '+40/5'
mps=40 tc=300000 inc=0
85313 >first : new
random
85313 >first : variant fischerandom
85313 >first : force
write FEN 50-move: 0 0 0
e0. p=-4
85313 >first : setboard rbnnkrbq/pppppppp/8/8/8/8/PPPPPPPP/RBNNKRBQ w FAfa - 0 1
85313 >first : ics -
85313 >first : level 40 5 0
85313 >first : post
85313 >first : hard
85313 >first : easy
85313 >first : ping 3
85344 <first : pong 2
85344 <first : pong 3
90266 >first : force
StartChildProcess (dir="C:\ChessEngines_FRC\TheBaron\v2_23") FRCbaron
90766 >second: xboard
protover 2
90782 <second:
90782 <second: The Baron
90782 <second: version 2.23 (FRC), January 29th 2009
90782 <second: Copyright (C) 2007 Richard Pijl
90782 <second:
90782 <second: Pthreads-win32 Library
90782 <second: Copyright (C) 2001 Ross P. Johnson
90782 <second:
90782 <second: Late Move Reductions enabled
90782 <second: Size of hash element: 24 bytes
90782 <second: Transposition Hashtables set to 24576 kbytes
90782 <second: Pawn Evaluation Hashtables set to 3072 kbytes
90782 <second: Initializing Tablebases ... No Tablebases found
90782 <second: Not able to open the openingbook: baronW.obp
90782 <second: Position file belongs to different version: Baron.wlp
90782 <second: Not able to open the position file: oldbaron.wlp
90782 <second: Not able to open the position file: oldbaron.blp
90782 <second: Size of white learnfile: 0 positions
90782 <second: Size of black learnfile: 3 positions
90782 <second: W<1>->
90782 <second: feature done=0
90782 >second: accepted done
90782 <second:
90782 <second: feature analyze=1
90782 >second: accepted analyze
90782 <second: feature colors=0
90782 >second: accepted colors
90782 <second: feature draw=1
90782 >second: accepted draw
90782 <second: feature ics=1
90782 >second: accepted ics
90782 <second: feature myname="The Baron 2.23 (FRC)"
90782 >second: accepted myname
90782 <second: feature ping=1
90782 >second: accepted ping
90782 <second: feature playother=1
90782 >second: accepted playother
90782 <second: feature setboard=1
90782 >second: accepted setboard
90782 <second: feature time=1
90782 >second: accepted time
90782 <second: feature usermove=0
90782 >second: accepted usermove
90782 <second: feature variants="normal,nocastle,fischerandom"
90782 >second: accepted variants
90782 <second: feature done=1
90782 >second: accepted done
90797 >second: new
random
90797 >second: variant fischerandom
90797 >second: ics -
90797 >second: level 40 5 0
90797 >second: post
90797 >second: hard
90797 >second: easy
90797 >second: ping 1
90797 >second: force
write FEN 50-move: 0 0 0
e0. p=-4
90797 >second: setboard rbnnkrbq/pppppppp/8/8/8/8/PPPPPPPP/RBNNKRBQ w FAfa - 0 1
Two Machines
90797 >first : computer
90797 >first : name The Baron 2.23 (FRC)
90797 >second: computer
TC string = '+40/5'
mps=40 tc=300000 inc=0
TC string = '+40/5'
mps=40 tc=300000 inc=0
time odds: 1 1
90797 >first : time 30000
90797 >first : otim 30000
90797 >first : go
nps: w=-1, b=-1
90875 <first : 2 +11 0 46 Nd3 Nd6
90875 <first : 3 +50 0 216 Nd3 Nd6 Ne3
90875 <first : 4 +11 0 533 Nd3 Nd6 Ne3 Ne6
90875 <first : 5 +31 0 2647 Nd3 Ne6 f4 Nd6 Nc3
90875 <first : 5 +37 0 3770 f3 f5 Nd3 Nd6 Bd4
90875 <first : 6 +9 0 6039 f3 f5 Bf2 Nd6 O-O Bd5
90875 <first : 6 +11 0 7646 Nd3 f5 f4 Nd6 Bd4 Bd5
90875 <second: pong 1
90875 <second: Error {Error on input}: syntax error, unexpected STRING, expecting FEN
90875 <second: Fen string has illegal format
90891 <first : 7 +52 5 20633 Nd3 f5 f3 c6 Be3 f4 Bd4
90907 <first : 8 +29 5 27140 Nd3 f5 f3 c6 Bc5 Bd5 Nc3 O-O Nxd5 cxd5
90953 <first : 9 +37 11 60899 Nd3 f6 f3 Nd6 Be3 Bc4 O-O O-O Nc3
91000 <first : 9 +39 14 89728 f3 Nd6 Be3 f6 O-O Be6 Nc3 O-O Nd3
91094 <first : 10 +13 20 128663 f3 Nd6 Nd3 f6 Bc5 Be6 O-O O-O Ne3 Nc6
91250 <first : 11 +19 41 262750 f3 Nd6 Nd3 f5 Bc5 Be6 O-O O-O Ne3 Nc6 c3
91563 <first : 12 +21 72 459154 f3 Nd6 Nd3 f5 Bc5 Ne6 Bb4 f4 O-O Nd4 Nc3 Rf5
92313 <first : 13 +23 147 957613 f3 Nd6 Nd3 f5 Bc5 Be6 O-O O-O Nf4 Bc4 Ne3 Ne6 Nxc4 Nxc5
size-move
moving
moving
moving
moving
moving
moving
moving
moving
moving
moving
moving
moving
moving
moving
moving
moving
moving
moving
moving
moving
moving
moving
moving
93719 <first : 14 +15 288 1862191 f3 Nd6 Nd3 f5 Bc5 Be6 O-O O-O Nf4 b6 Bxd6 cxd6 Ne3 Nc6 c4
moving
moving
moving
moving
moving
moving
moving
moving
moving
moving
moving
moving
moving
moving
moving
exit size-move, size = 54
square size = 54
99360 <first : 15 +23 852 5473530 f3 Nd6 Nd3 f5 Be3 Nc6 O-O Bf7 a3 O-O Nc3 Nc4 Ba2 e5 Bxc4 Bxc4
101172 <first : 15 +23 1033 6660479 f3 Nd6 Nd3 f5 Be3 Nc6 O-O Bf7 a3 O-O Nc3 Nc4 Ba2 e5 Bxc4 Bxc4
101172 <first : move f2f3
machine move 0, castling = 5 0 4 5 0 4
move to parse: f2f3
5 0 4 5 0 4 Legality test? f2f3
FRC castling, 5 0 4 5 0 4
5 0 4 5 0 4 Legality test? f2f3
FRC castling, 5 0 4 5 0 4
(5,0) (0,0) (4,0) (5,7) (0,7) (4,7) castling rights
TC string = '+40/5'
mps=40 tc=300000 inc=0
CoordsToAlgebraic, piece=0 (5,1)-(5,2) -
5 0 4 5 0 4 Legality test? f2f3
FRC castling, 5 0 4 5 0 4
movetype=33, promochar=0=-
move: f2f3
, parse: f3 (
)
repeat test fmm=1 bmm=0 ep=-4, reps=6
1 ep=-3
0 ep=-4
time odds: 1 1
101188 >second: time 30000
101188 >second: otim 28960
101188 >second: f2f3
101188 >second: go
101282 <second: 1& 116 0 2 e7-e5
101282 <second: 1. 116 0 21 e7-e5
101282 <second: 2- 24 0 42 e7-e5 (Nb1-c3)
101282 <second: 2& 24 0 42 e7-e5 Nb1-c3
101282 <second: 2. 24 0 80 e7-e5 Nb1-c3
101282 <second: 3+ 106 0 128 e7-e5 (Nb1-c3) (Nb8-c6)
101282 <second: 3& 120 0 251 e7-e5 e2-e4 d7-d5
101282 <second: 3. 120 0 404 e7-e5 e2-e4 d7-d5
101282 <second: 4- 23 0 775 e7-e5 (e2-e4) (Nb8-c6) (Nb1-c3)
101282 <second: 4& 23 0 775 e7-e5 e2-e4 Nb8-c6 Nb1-c3
101282 <second: 4. 23 0 1342 e7-e5 e2-e4 Nb8-c6 Nb1-c3
101282 <second: 5+ 58 0 1826 e7-e5 (e2-e4) (Nb8-c6) (Nb1-c3) (Ng8-f6)
101282 <second: 5& 82 0 2462 e7-e5 e2-e4 Nb8-c6 Nb1-c3 Ng8-f6
101282 <second: 5& 88 0 3859 d7-d5 e2-e4 Nb8-c6 Nb1-c3 Ng8-f6
101282 <second: 5. 88 0 4353 d7-d5 e2-e4 Nb8-c6 Nb1-c3 Ng8-f6
101282 <second: 6- 18 0 7656 d7-d5 (e2-e4) (e7-e5) (Nb1-c3) (Ng8-f6) (d2-d4) (d5xe4) (d4xe5) (Qd8xd1) (Ke1xd1)
101282 <second: 6& 18 0 7656 d7-d5 e2-e4 e7-e5 Nb1-c3 Ng8-f6 d2-d4 (d5xe4) (d4xe5) (Qd8xd1) (Ke1xd1)
101282 <second: 6& 47 0 9534 e7-e5 e2-e4 Nb8-c6 Nb1-c3 Ng8-f6 Ng1-h3
101282 <second: 6& 60 0 11180 Nb8-c6 e2-e4 Ng8-f6 Nb1-c3 d7-d5 (d2-d3)
101282 <second: 6. 60 0 12967 Nb8-c6 e2-e4 Ng8-f6 Nb1-c3 d7-d5 (d2-d3)
101282 <second: 7& 75 0 23700 Nb8-c6 d2-d4 e7-e5 d4xe5 Nc6xe5 e2-e4 Ng8-f6
101328 <second: 7+ 95 0 34013 e7-e5 (e2-e4) (Nb8-c6) (Nb1-c3) (Ng8-f6) (Ng1-h3) (d7-d5)
101360 <second: 7& 104 0 49205 e7-e5 d2-d4 Qd8-h4+ g2-g3 Qh4xd4 Qd1xd4 e5xd4
101375 <second: 7. 104 0 50613 e7-e5 d2-d4 Qd8-h4+ g2-g3 Qh4xd4 Qd1xd4 e5xd4
101453 <second: 8- 62 0 81471 e7-e5 (d2-d4) (e5xd4) (Qd1xd4) (Ng8-f6) (Qd4-e5) (Bf8-e7) (e2-e4) (O-O)
101453 <second: 8& 62 0 81471 e7-e5 d2-d4 e5xd4 Qd1xd4 Ng8-f6 Qd4-e5+ Bf8-e7 e2-e4 (O-O)
101578 <second: 8. 62 0 122063 e7-e5 d2-d4 e5xd4 Qd1xd4 Ng8-f6 Qd4-e5+ Bf8-e7 e2-e4 (O-O)
101828 <second: 9& 81 0 197434 e7-e5 e2-e4 Nb8-c6 d2-d4 Nc6xd4 Nb1-c3 d7-d5 Nc3xd5 Ng8-f6
101922 <second: 9. 81 0 223224 e7-e5 e2-e4 Nb8-c6 d2-d4 Nc6xd4 Nb1-c3 d7-d5 Nc3xd5 Ng8-f6
102969 <second: 10& 54 1 552212 e7-e5 d2-d4 Qd8-h4+ g2-g3 Qh4xd4 Qd1xd4 e5xd4 Bc1-f4 Bf8-b4+ Nb1-d2
103438 <second: 10. 54 2 704413 e7-e5 d2-d4 Qd8-h4+ g2-g3 Qh4xd4 Qd1xd4 e5xd4 Bc1-f4 Bf8-b4+ Nb1-d2
104907 <second: 11& 71 3 1154405 e7-e5 e2-e4 Nb8-c6 d2-d4 Nc6xd4 Nb1-c3 Bf8-b4 Bc1-d2 Ng8-f6 Ng1-e2 d7-d5
size-move
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
106157 <second: 11. 71 4 1481845 e7-e5 e2-e4 Nb8-c6 d2-d4 Nc6xd4 Nb1-c3 Bf8-b4 Bc1-d2 Ng8-f6 Ng1-e2 d7-d5
sizing
exit size-move, size = 54
square size = 54
108828 <second: 12& 74 7 2241736 e7-e5 e2-e4 Nb8-c6 Nb1-c3 Ng8-f6 Bf1-c4 Nc6-a5 d2-d3 Na5xc4 d3xc4 Bf8-c5 Ng1-e2 (O-O)
size-move
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
sizing
exit size-move, size = 54
square size = 45
111219 <second: 12. 74 10 2856554 e7-e5 e2-e4 Nb8-c6 Nb1-c3 Ng8-f6 Bf1-c4 Nc6-a5 d2-d3 Na5xc4 d3xc4 Bf8-c5 Ng1-e2 (O-O)
111219 <second: move e7e5
machine move 1, castling = 5 0 4 5 0 4
move to parse: e7e5
5 0 4 5 0 4 Legality test? e7e5
5 0 4 5 0 4 Legality test? e7e5
(5,0) (0,0) (4,0) (5,7) (0,7) (4,7) castling rights
TC string = '+40/5'
mps=40 tc=300000 inc=0
CoordsToAlgebraic, piece=22 (4,6)-(4,4) -
5 0 4 5 0 4 Legality test? e7e5
movetype=33, promochar=0=-
FRC castling, 5 0 4 5 0 4
move: e7e5
, parse: e5 (
)
FRC castling, 5 0 4 5 0 4
repeat test fmm=2 bmm=0 ep=-4, reps=6
2 ep=-3
1 ep=-3
0 ep=-4
time odds: 1 1
111219 >first : time 28960
111219 >first : otim 28996
111219 >first : usermove 111219 >first : e7e5
111313 <first : 2 +31 0 66 c4 f5
111313 <first : 2 +41 0 120 Nd3 f6
111313 <first : 3 +58 0 287 Nd3 f6 Nc3
111313 <first : 3 +60 0 1025 Be3 f5 O-O
111313 <first : 4 +33 0 1221 Be3 f5 O-O Nd6
111313 <first : 5 +60 0 2572 Be3 f5 O-O f4 Bc5
111313 <first : 5 +72 0 3947 Bf2 f5 Nd3 Nc6 O-O
111313 <first : 6 +45 2 5712 Bf2 f5 Nd3 Nc6 O-O Nd6
111313 <first : 6 +50 2 7196 Be3 f5 Nd3 f4 Bc5 d6
111313 <first : 7 +47 3 16058 Be3 f6 O-O Be6 Nc3 O-O Nd3
111313 <first : 7 +70 3 19708 Nd3 f6 Bc5 d6 Be3 Ke7 O-O
111313 <first : 8 +37 6 31787 Nd3 f6 f4 d6 Nc3 Nc6 fxe5 fxe5
111422 <first : 9 +37 19 99383 Nd3 f6 e4 Nd6 Be3 Bc4 O-O O-O Nc3
111532 <first : 10 +23 30 174014 Nd3 f6 Nc3 Nc6 Be3 Be6 O-O O-O Nc5 Nd6 Nxe6 dxe6
112016 <first : 11 +27 77 438584 Nd3 f6 Nc3 Nc6 Be3 Be6 f4 O-O fxe5 fxe5 Rxf8+ Kxf8 Kf1
112516 <first : 12 +29 128 764018 Nd3 f6 Nc3 Nc6 Be3 Be6 f4 O-O fxe5 Nxe5 Nxe5 fxe5 Rxf8+ Kxf8 Ne4
114578 <first : 13 +41 333 2053248 Nd3 f6 f4 e4 Nc5 f5 g4 g6 gxf5 gxf5 Nc3 d6 Nb3
118266 <first : 14 +29 703 4413033 Nd3 f6 a3 Nd6 Nc3 Bc4 f4 N8f7 Nc5 O-O Nxd7 Rd8 fxe5 Rxd7 exd6 Rxd6
121563 <first : 14 +29 1033 6529364 Nd3 f6 a3 Nd6 Nc3 Bc4 f4 N8f7 Nc5 O-O Nxd7 Rd8 fxe5 Rxd7 exd6 Rxd6
121563 <first : move c1d3
machine move 2, castling = 5 0 4 5 0 4
move to parse: c1d3
5 0 4 5 0 4 Legality test? c1d3
FRC castling, 5 0 4 5 0 4
5 0 4 5 0 4 Legality test? c1d3
FRC castling, 5 0 4 5 0 4
(5,0) (0,0) (4,0) (5,7) (0,7) (4,7) castling rights
TC string = '+40/5'
mps=40 tc=300000 inc=0
CoordsToAlgebraic, piece=1 (2,0)-(3,2) -
FRC castling, 5 0 4 5 0 4
move: c1d3
, parse: Nd3 (
)
repeat test fmm=3 bmm=0 ep=-4, reps=6
3 ep=-4
2 ep=-3
1 ep=-3
0 ep=-4
time odds: 1 1
Interrupting second
121563 >second: time 28996
121563 >second: otim 27926
121563 >second: c1d3
121672 <second: Illegal move (Move not found): c1d3
TC string = '+40/5'
mps=40 tc=300000 inc=0
GameEnds(37, False illegal-move claim, 4)
GE(37, False illegal-move claim, 4) bare king k=16 color=0
121688 >first : result 1-0 {False illegal-move claim}
121688 >second: result 1-0 {False illegal-move claim}
121688 >first : force
121688 >first : ping 4
121688 >second: force
121688 >second: ping 2
121688 <first : pong 4
121688 <second: Book is not open
121688 <second: pong 2
GameEnds(37, False illegal-move claim, 2)
146985 >first : quit
147500 >second: quit

[pedrox]

Baron is a winboard engine.

There are 3 GUI that allow winboard engines play Chess960, are:
Arena, Winboard F and chessGUI.

Arena and ChessGUI play a game that allow Chess960 from the start position, the GUI sends to the engine the position as a FEN, thus engines that play well in Arena will make also in chessGUI or vice versa.

Winboard F sends to engines the format shredder FEN, which has an advantage, not only can play a game from the starting position, but you can set any position without ambiguity. But for that a winboard engine can play in Winboard F has to adapt and understand shredder FEN. I think DanaSah is one of the few winboard engines that have updated for this.

Perhaps Baron play well chess960 in the Arena and ChessGUI.

Pedro

[Spock]

I've played around 150 games so far with The Baron 2.23 FRC under ChessGUI, with no problems at all.

[hgm]

It seems that most WinBoard FRC engines are not capable of understanding FRC FENs. A bit surprising, as such FENs are standard in UCI.

I guess the next WinBoard version will come with a command-line option that allows the user to force the castling field in the setboard FEN to any character string (such as "KQkq"). This should allow them to play these buggy engines under WinBoard without problems.

[pijl]

90797 >second: setboard rbnnkrbq/pppppppp/8/8/8/8/PPPPPPPP/RBNNKRBQ w FAfa - 0 1
90875 <second: Error {Error on input}: syntax error, unexpected STRING, expecting FEN
90875 <second: Fen string has illegal format

Please use a GUI that sends standard FENs in the winboard protocol instead of the UCI variant of it.
Richard.