Tinkering with Winboard

[Naum]

Thanks Alessandro.
You are doing a great job.

Is some kind of integrated tournament manager in your plans?

[George Lyapko]

NPS for Atlanchess for some reasons are always 100 times bigger than they are actually. Time used is wrong. I think this problem will be with all engines which outputs time in seconds(and not in cs). The easiest solution is to get the time elapsed not from engine's thinking line but from OS...

[Alessandro Scotti]

Is some kind of integrated tournament manager in your plans?

Hi Alex,
a tournament manager is currently not in short-term plans. One of the (many) things I've still not tried to figure out in Winboard is how to change an engine on the fly. This would allow several interesting new features, and is also a necessary first step towards a tournament manager.

[Alessandro Scotti]

NPS for Atlanchess for some reasons are always 100 times bigger than they are actually. Time used is wrong. I think this problem will be with all engines which outputs time in seconds(and not in cs). The easiest solution is to get the time elapsed not from engine's thinking line but from OS...

Hi George,
much easier for me is if Atlanchess fixes this bug in the PV output!

[Dann Corbit]

Is some kind of integrated tournament manager in your plans?

Hi Alex,
a tournament manager is currently not in short-term plans. One of the (many) things I've still not tried to figure out in Winboard is how to change an engine on the fly. This would allow several interesting new features, and is also a necessary first step towards a tournament manager.

Here is a perl tournament manager for xboard. Ought to be trivial to use it in Winboard. Or there is galais, etc.
http://www.hoicher.de/hoichess/tourney_ ... index.html

In Winboard, you can at least do the /mg <n> command line option for two opponents.

I usually use Arena for tournaments, or G.L.'s utility.

[Daniel Shawul]

Hi Alessandro
In the move list window, the score of the engine when pondering is recodred and appended to the pgn. Some engines ,like scorpio, can do pondering even when not out of book. This can cause confusion as to when the engine got out of book. Note that there is no problem when ponder off.
Daniel

[carlospesce]

Hi, Alessandro!
As an old winboarder, i want to say to you my congratulations for your great job!
But i want to insist in to add some tournament mode. Perhaps not complex, but its the las important feature that Wb needs.
Maybe Dann?s data can be the base..but, all in all, this feature should be added.
Cheers
Carlos Pesce

[Alessandro Scotti]

Hi,
just a little update this time, with the ability of changing the move history font (*) and finally analysis mode in the engine output window. Download as usual from:

http://www.ascotti.org/programming/chess/winboard_x.htm

(*) Setting extra styles like bold, italic, etc. is not supported yet.

[Thomas Mayer]

Hi Alessandro,

I did mention the problem with the repeating thinking lines in analyze output every 2 seconds. Well, I could it finally reproduce, but first I thought I should give your new version I try. The problem seems still to exist.

Here is what I got analyzing with Quark from the root position:

Code: Select all

 13   +0.23   26.6M   0:44.09   d2d4 d7d5 c1f4 g8f6 e2e3 e7e6 g1h3 f8d6 f1d3 e8g8
 13   +0.23   25.4M   0:42.09   d2d4 d7d5 c1f4 g8f6 e2e3 e7e6 g1h3 f8d6 f1d3 e8g8
 13   +0.23   24.2M   0:40.09   d2d4 d7d5 c1f4 g8f6 e2e3 e7e6 g1h3 f8d6 f1d3 e8g8
 12   +0.23   23.5M   0:39.03   d2d4 d7d5 c1f4 g8f6 e2e3 e7e6 g1h3 f8d6 f1d3 e8g8
 12   +0.23   22.9M   0:38.09   d2d4 d7d5 c1f4 g8f6 e2e3 e7e6 g1h3 f8d6 f1d3 e8g8
 12   +0.23   21.7M   0:36.09   d2d4 d7d5 c1f4 g8f6 e2e3 e7e6 g1h3 f8d6 f1d3 e8g8
 12   +0.23   20.5M   0:34.09   d2d4 d7d5 c1f4 g8f6 e2e3 e7e6 g1h3 f8d6 f1d3 e8g8
 12   +0.23   19.6M   0:32.53   d2d4 d7d5 c1f4 g8f6 e2e3 e7e6 g1h3 f8d6 f1d3 e8g8
 12   +0.06   19.3M   0:32.09   d2d4!
 12   +0.06   18.1M   0:30.09   d2d4!
 12   +0.06   17.0M   0:28.09   d2d4!
 12   +0.06   15.8M   0:26.07   d2d4!
 12   +0.06   14.6M   0:24.07   d2d4!
 12   +0.06   13.4M   0:22.06   d2d4!
 12   +0.06   13.1M   0:21.54   d2d4!
 12   +0.05   12.2M   0:20.10   e2e4 e7e5 g1f3 g8f6 f3e5 d7d6 e5c4 f6e4 d2d3 e4c5 b2b4 b7b5
 12   +0.31   12.2M   0:20.06   e2e4 e7e6 d2d4 d7d5 b1c3 f8b4 d1g4 d5e4 g4g7 d8f6 c1h6
 12   +0.31   11.0M   0:18.06   e2e4 e7e6 d2d4 d7d5 b1c3 f8b4 d1g4 d5e4 g4g7 d8f6 c1h6
 12   +0.31   9.8M   0:16.06   e2e4 e7e6 d2d4 d7d5 b1c3 f8b4 d1g4 d5e4 g4g7 d8f6 c1h6
 12   +0.31   8.6M   0:14.06   e2e4 e7e6 d2d4 d7d5 b1c3 f8b4 d1g4 d5e4 g4g7 d8f6 c1h6
 12   +0.31   7.5M   0:12.04   e2e4 e7e6 d2d4 d7d5 b1c3 f8b4 d1g4 d5e4 g4g7 d8f6 c1h6
 12   +0.31   6.2M   0:10.04   e2e4 e7e6 d2d4 d7d5 b1c3 f8b4 d1g4 d5e4 g4g7 d8f6 c1h6
 12   +0.31   4.9M   0:08.03   e2e4 e7e6 d2d4 d7d5 b1c3 f8b4 d1g4 d5e4 g4g7 d8f6 c1h6
 11   +0.31   4.6M   0:07.56   e2e4 e7e6 d2d4 d7d5 b1c3 f8b4 d1g4 d5e4 g4g7 d8f6 c1h6
 11   +0.31   3.6M   0:06.03   e2e4 e7e6 d2d4 d7d5 b1c3 f8b4 d1g4 d5e4 g4g7 d8f6 c1h6
 11   +0.31   3.3M   0:05.51   e2e4 e7e6 d2d4 d7d5 b1c3 f8b4 d1g4 d5e4 g4g7 d8f6 c1h6
 11   +0.26   2.4M   0:04.03   e2e4 e7e6 c2c4 b8c6 b1c3 d7d6 g1f3 c8d7 d2d4 c6b4
 10   +0.26   1.6M   0:02.75   e2e4 e7e6 c2c4 b8c6 b1c3 d7d6 g1f3 c8d7 d2d4 c6b4
 10   +0.26   1.3M   0:02.32   e2e4 e7e6 c2c4 b8c6 b1c3 d7d6 g1f3 c8d7 d2d4 c6b4
 10   +0.45   1.1M   0:02.01   e2e4 e7e6 c2c4 b8c6 d2d4 d7d5 c4d5 f8b4 b1c3 e6d5
  9   +0.45   772744   0:01.46   e2e4 e7e6 c2c4 b8c6 d2d4 d7d5 c4d5 f8b4 b1c3 e6d5
  9   +0.45   684510   0:01.32   e2e4 e7e6 c2c4 b8c6 d2d4 d7d5 c4d5 f8b4 b1c3 e6d5
  8   +0.25   284968   0:00.65   e2e4 e7e5 g1f3 d7d6 f1c4 c7c6 e1g1 f8e7
  8   +0.25   260536   0:00.60   e2e4 e7e5 g1f3 d7d6 f1c4 c7c6 e1g1 f8e7
  8   +0.06   164614   0:00.45   e2e4!
  8   +0.05   115210   0:00.35   d2d4 d7d5 c1f4 c8f5 e2e3 g8f6 b1c3 c7c6
  7   +0.35   67407   0:00.26   d2d4 d7d5 c1f4 c8f5 e2e3 e7e6 b1c3
  6   +0.19   20598   0:00.09   e2e4 e7e5 d2d4 e5d4 c2c3 d4c3
  5   +0.42   6275   0:00.03   d2d4 d7d5 c2c3 e7e6 g1f3
  4   0.00   1344   0:00.01   e2e4 d7d5 d2d3 e7e6


I am still not sure what the reason is, it seems to me that the problem starts as soon Quark marks a move with "!" but that is just a guess...

here is the interesting part of the debug file for the output:

Code: Select all

1234 >first : accepted done
1250 >first : new
random
1250 >first : ics -
1250 >first : level 0 5 0
1250 >first : post
1250 >first : hard
1250 >first : ping 1
1437 <first : pong 1
3796 >first : force
3796 >first : analyze
3812 <first :  4      0       1       1344  e2e4 d7d5 d2d3 e7e6
3843 <first :  5     42       3       6275  d2d4 d7d5 c2c3 e7e6 g1f3
3906 <first :  6     19       9      20598  e2e4 e7e5 d2d4 e5d4 c2c3 d4c3
4078 <first :  7     35      26      67407  d2d4 d7d5 c1f4 c8f5 e2e3 e7e6 b1c3
4156 <first :  8      5      35     115210  d2d4 d7d5 c1f4 c8f5 e2e3 g8f6 b1c3 c7c6
4250 <first :  8      6      45     164614  e2e4!
4421 <first :  8     25      60     260536  e2e4 e7e5 g1f3 d7d6 f1c4 c7c6 e1g1 f8e7
4453 <first :  8     25      65     284968  e2e4 e7e5 g1f3 d7d6 f1c4 c7c6 e1g1 f8e7
5140 <first :  9     45     132     684510  e2e4 e7e6 c2c4 b8c6 d2d4 d7d5 c4d5 f8b4 b1c3 e6d5
5281 <first :  9     45     146     772744  e2e4 e7e6 c2c4 b8c6 d2d4 d7d5 c4d5 f8b4 b1c3 e6d5
5812 >first : .
5812 <first : stat01: 201 1132726 10 19 20 e2e4
6125 <first : 10     26     232    1315088  e2e4 e7e6 c2c4 b8c6 b1c3 d7d6 g1f3 c8d7 d2d4 c6b4
6546 <first : 10     26     275    1589467  e2e4 e7e6 c2c4 b8c6 b1c3 d7d6 g1f3 c8d7 d2d4 c6b4
7828 >first : .
7828 <first : stat01: 403 2385489 11 19 20 e2e4
9312 <first : 11     31     551    3296802  e2e4 e7e6 d2d4 d7d5 b1c3 f8b4 d1g4 d5e4 g4g7 d8f6 c1h6
9828 >first : .
9828 <first : stat01: 603 3608293 11 18 20 d2d4
11359 <first : 11     31     756    4572037  e2e4 e7e6 d2d4 d7d5 b1c3 f8b4 d1g4 d5e4 g4g7 d8f6 c1h6
11828 >first : .
11828 <first : stat01: 803 4860650 12 19 20 e2e4
13843 >first : .
13843 <first : stat01: 1004 6174063 12 19 20 e2e4
15843 >first : .
15843 <first : stat01: 1204 7457014 12 19 20 e2e4
17859 >first : .
17859 <first : stat01: 1406 8630880 12 19 20 e2e4
19859 >first : .
19859 <first : stat01: 1606 9843640 12 19 20 e2e4
21859 >first : .
21859 <first : stat01: 1806 11007265 12 19 20 e2e4
23859 >first : .
23859 <first : stat01: 2006 12221632 12 19 20 e2e4
23906 <first : 12      5    2010   12248000  e2e4 e7e5 g1f3 g8f6 f3e5 d7d6 e5c4 f6e4 d2d3 e4c5 b2b4 b7b5
25359 <first : 12      6    2154   13106870  d2d4!
25859 >first : .
25859 <first : stat01: 2206 13424831 12 18 20 d2d4
27875 >first : .
27875 <first : stat01: 2407 14628885 12 18 20 d2d4
29875 >first : .
29875 <first : stat01: 2607 15843319 12 18 20 d2d4
31890 >first : .
31890 <first : stat01: 2809 16956942 12 18 20 d2d4
33890 >first : .
33890 <first : stat01: 3009 18129689 12 18 20 d2d4
35890 >first : .
35890 <first : stat01: 3209 19311401 12 18 20 d2d4
36328 <first : 12     23    3253   19578613  d2d4 d7d5 c1f4 g8f6 e2e3 e7e6 g1h3 f8d6 f1d3 e8g8
37890 >first : .
37890 <first : stat01: 3409 20543894 12 16 20 c2c4
39890 >first : .
39890 <first : stat01: 3609 21695708 12 16 20 c2c4
41890 >first : .
41890 <first : stat01: 3809 22937262 12 8 20 b2b4
42828 <first : 12     23    3903   23529335  d2d4 d7d5 c1f4 g8f6 e2e3 e7e6 g1h3 f8d6 f1d3 e8g8
43890 >first : .
43890 <first : stat01: 4009 24170019 13 19 20 d2d4
45890 >first : .
45906 <first : stat01: 4209 25398020 13 19 20 d2d4
47890 >first : .
47890 <first : stat01: 4409 26620452 13 19 20 d2d4
49890 >first : .
49890 <first : stat01: 4609 27803617 13 19 20 d2d4
51890 >first : .
51890 <first : stat01: 4809 28976223 13 19 20 d2d4
53890 >first : .
53890 <first : stat01: 5009 30101395 13 19 20 d2d4
55890 >first : .
55890 <first : stat01: 5209 31254752 13 19 20 d2d4
57890 >first : .
57906 <first : stat01: 5410 32408203 13 19 20 d2d4
59890 >first : .
59890 <first : stat01: 5609 33544880 13 19 20 d2d4
61890 >first : .
61890 <first : stat01: 5809 34759862 13 19 20 d2d4


I wonder if that problem occurs only with Quark ? If you can not reproduce it with the public v2.35 I could also send you the v2.62 with which I have tested it.

Greetings, Thomas

P.S.: Great work your winboard - simply fantastic, can't be said often enough - and with all the enhancements, WinBoard still keeps it's easyness and low cpu-usage. Not to mention the compatibility to all available tools.

[Alessandro Scotti]

Hi Thomas!

I did mention the problem with the repeating thinking lines in analyze output every 2 seconds. Well, I could it finally reproduce, but first I thought I should give your new version I try. The problem seems still to exist.

From the output you post it seems the lines are not repeating "verbatim", because the node count changes. So the engine is still searching and sending output every two seconds, although the PV stays the same.

Usually I just take the engine output and display it, but in these cases this behavior is probably not entirely correct. What I should do is compare the data with the last PV, and if the only changes are in time and nodes just update those without adding a new line to the output. Ok, something for the "todo" list then!

P.S.: Great work your winboard - simply fantastic, can't be said often enough - and with all the enhancements, WinBoard still keeps it's easyness and low cpu-usage. Not to mention the compatibility to all available tools.

Thanks a lot Thomas, I'm happy you like it! Actually, I like Winboard for the same reasons too, so I always try to pay attention to the original design and "philosophy".

[Thomas Mayer]

Hi Alessandro,

From the output you post it seems the lines are not repeating "verbatim", because the node count changes. So the engine is still searching and sending output every two seconds, although the PV stays the same.

I believe the problem is different then what you describe... take again a look at the winboard.debug for ply 12:

Code: Select all

23906 <first : 12      5    2010   12248000  e2e4 e7e5 g1f3 g8f6 f3e5 d7d6 e5c4 f6e4 d2d3 e4c5 b2b4 b7b5
25359 <first : 12      6    2154   13106870  d2d4!
25859 >first : .
25859 <first : stat01: 2206 13424831 12 18 20 d2d4
27875 >first : .
27875 <first : stat01: 2407 14628885 12 18 20 d2d4
29875 >first : .
29875 <first : stat01: 2607 15843319 12 18 20 d2d4
31890 >first : .
31890 <first : stat01: 2809 16956942 12 18 20 d2d4
33890 >first : .
33890 <first : stat01: 3009 18129689 12 18 20 d2d4
35890 >first : .
35890 <first : stat01: 3209 19311401 12 18 20 d2d4
36328 <first : 12     23    3253   19578613  d2d4 d7d5 c1f4 g8f6 e2e3 e7e6 g1h3 f8d6 f1d3 e8g8
37890 >first : .
37890 <first : stat01: 3409 20543894 12 16 20 c2c4
39890 >first : .
39890 <first : stat01: 3609 21695708 12 16 20 c2c4
41890 >first : .
41890 <first : stat01: 3809 22937262 12 8 20 b2b4
42828 <first : 12     23    3903   23529335  d2d4 d7d5 c1f4 g8f6 e2e3 e7e6 g1h3 f8d6 f1d3 e8g8
43890 >first : .
43890 <first : stat01: 4009 24170019 13 19 20 d2d4
45890 >first : .
45906 <first : stat01: 4209 25398020 13 19 20 d2d4
47890 >first : .
47890 <first : stat01: 4409 26620452 13 19 20 d2d4


as you can see, Quark sends as line only once d2d4!... The rest are just reaction on the "." to provide some search information.

Now your analysis window shows the following for ply 12:

Code: Select all

 12   +0.23   19.6M   0:32.53   d2d4 d7d5 c1f4 g8f6 e2e3 e7e6 g1h3 f8d6 f1d3 e8g8
 12   +0.06   19.3M   0:32.09   d2d4!
 12   +0.06   18.1M   0:30.09   d2d4!
 12   +0.06   17.0M   0:28.09   d2d4!
 12   +0.06   15.8M   0:26.07   d2d4!
 12   +0.06   14.6M   0:24.07   d2d4!
 12   +0.06   13.4M   0:22.06   d2d4!
 12   +0.06   13.1M   0:21.54   d2d4!
 12   +0.05   12.2M   0:20.10   e2e4 e7e5 g1f3 g8f6 f3e5 d7d6 e5c4 f6e4 d2d3 e4c5 b2b4 b7b5


So that means for each "." it did repeat the given line. (This is repeated also in the next plys, when you take a look)

Greets, Thomas

[Alessandro Scotti]

Hi Alessandro,
I believe the problem is different then what you describe... take again a look at the winboard.debug for ply 12:

Hi Thomas,
sorry I was looking at "full lines" and didn't look enough at the PV part. I see the problem now... will try to fix it soon!

[Alessandro Scotti]

Hi all,
I've uploaded a bug-fix version of Winboard at the usual place. It seems I managed to introduce a serious bug a couple of versions ago, when I failed to notice that a variable I was modifying was global and thus having side-effects on all the program.
I've fixed that now, and also cleaned up the analysis output.

Ciao,

Alessandro

[David Weller]

Hi Allessandro,

I tried to compile xboard_4.2.6 with your mods on Linux [Ubuntu Breezy Badger]

and get this error concerning too few params to

selected_fen_position = (char *)PositionToFEN(currentMove);

Maybe you recognize a bug right-a-way ??

It's in xboard.c - am looking for the declaration now .....

-David

found another ...

backend.o: In function `SendProgramStatsToFrontend':
backend.c:(.text+0x783e): undefined reference to `SetProgramStats'
backend.o: In function `HandleMachineMove':
backend.c:(.text+0x8019): undefined reference to `DisplayNote'
backend.c:(.text+0x8122): undefined reference to `DisplayNote'
backend.c:(.text+0x8265): undefined reference to `DisplayNote'
collect2: ld returned 1 exit status
make: *** [xboard] Error 1

[mridul]

Hi,

Hope this is the right place to post this :

Was fooling around with Allessandro's version of winboard and noticed the following:
Sometimes , when playing at icc , on quitting the interface it does not send a 'quit' to the engine but kills it.
Is this an observed behaviour ?
I rely on the 'quit' being sent (I have exit handlers in place to cleanup stuff) to take care of some cleanup which fails 'cos of this.

Regards
Mridul

PS : I will save the logs next time , just wanted to know if this was known behaviour. The winboard.debug and the engine logs did not show any quit being sent - last command was 'force' if I recall correctly.

[mridul]

Hi,

Hope this is the right place to post this :

Was fooling around with Allessandro's version of winboard and noticed the following:
Sometimes , when playing at icc , on quitting the interface it does not send a 'quit' to the engine but kills it.
Is this an observed behaviour ?
I rely on the 'quit' being sent (I have exit handlers in place to cleanup stuff) to take care of some cleanup which fails 'cos of this.

Regards
Mridul

PS : I will save the logs next time , just wanted to know if this was known behaviour. The winboard.debug and the engine logs did not show any quit being sent - last command was 'force' if I recall correctly.

There was this last snippet I might have missed earlier in the logs :

Code: Select all

<ICS: You are not observing or playing in any games.\015\012aics%
GameEnds(0, (null), 2)
408078 >first : force
408078 >first : ping 4
408078 >first : quit
Process 1628 still alive after timeout, killing...


What is the timeout value by default ?
Can I configre this value ? I might need sometime to do the cleanup ...

Thanks,
Mridul[/code]

[Alessandro Scotti]

Hi Allessandro,

I tried to compile xboard_4.2.6 with your mods on Linux [Ubuntu Breezy Badger] and get this error concerning too few params to...

Hi David,
I think the amazing thing is that you get only one error!
I don't expect my Winboard code to compile "as is" on Linux, but probably a bit of diff and merge work should do the trick (with respect to backend.c and related stuff).

[Alessandro Scotti]

What is the timeout value by default ?
Can I configre this value ? I might need sometime to do the cleanup...

Hi Mridul,
I don't remembers the default now but the parameter should be called "delayAfterQuit" and be expressed in milliseconds. You can change it in your winboard.ini file.

[Niyaz Khasanov]

Hi,
the version of Winboard used by Leo Dijksman has an error in saving PGN files.
Sometimes the {score/depth} information is from the wrong side.
It seems that Winboard does not properly check which opponent sent a PV line. Maybe Winboard guesses based on the time when the PV was received. However, when ponder=on both opponents can send their PVs in any moment.

For example, you can see the following PGN from Leo's test tournament:
[Event "WBEC12 1st Division TEST"]
[Site "OPTERON-B"]
[Date "2006.01.19"]
[Round "1"]
[Number "113"]
[White "Delfi 4.6"]
[Black "Ufim 7.01"]
[Result "1-0"]
[TimeControl "40/180"]
[Annotator "6. -0.24 3... +0.05"]

1. c4 e5 2. Nc3 d6 3. d4 exd4 {+0.05/9} 4. Qxd4 Nf6 {+0.06/9} 5. Bg5
Be7 {+0.12/9} 6. e4 {-0.24/10} O-O {+0.27/9} 7. Nf3 {-0.08/11} h6 {+0.24/9}
8. Bf4 {-0.01/11} Nc6 {+0.03/11} 9. Qd2 {-0.08/12} Re8 {+0.14/10} 10.
Rd1 {+0.02/10} Nxe4 {+0.69/9} 11. Nxe4 {-0.32/11} Bf8 {+0.92/9} 12.
Bd3 {-0.73/11} d5 {+1.44/9} 13. Be2 {-0.94/10} dxe4 {-1.12/11} 14. Qxd8
Nxd8 {-1.05/11} 15. Nd2 {-1.08/11} Ne6 {-1.28/12} 16. Be3 f5 {+1.63/10} 17.
g3 {-1.39/11} Nc5 {-1.39/10} 18. O-O {-1.19/11} Nd3 {+1.60/10} 19.
Bxd3 {-0.95/12} exd3 {-1.12/12} 20. Nf3 f4 {+1.10/10} 21. Bxf4 {-0.60/11}
Bf5 {-0.76/10} 22. Rfe1 {-0.81/11} c5 {+1.27/9} 23. Be5 {-0.68/11}
Rad8 {+1.44/9} 24. Nd2 {-0.66/11} Re6 {-0.66/11} 25. Re3 {-0.72/12}
g5 {+1.56/9} 26. f4 {-0.58/11} h5 {-0.43/10} 27. Ne4 {-0.34/11}
gxf4 {+1.66/9} 28. gxf4 {-0.23/10} h4 {+1.24/8} 29. Kf2 {-0.23/9}
Rg6 {-0.39/11} 30. Rd2 Be7 {+1.05/9} 31. Kf3 {+0.00/11} Rg1 {+0.79/9} 32.
Nf2 {+0.14/10} Kf8 {+0.50/9} 33. Bc3 {+0.16/11} Ra1 {+0.20/10} 34.
a3 {+0.27/11} Rf1 {+0.07/9} 35. Kg2 {+0.29/12} Rb1 {+0.12/10} 36.
Re5 {+0.87/12} Bh7 {+0.61/11} 37. Rh5 {+0.98/10} Bg8 {-0.66/8} 38.
f5 {+1.14/11} Bxc4 {+1.34/10} 39. f6 Bd6 {+1.81/10} 40. Rh8+ {+2.04/11}
Bg8 {+2.14/10} 41. Ne4 Bc7 {-1.99/8} 42. Ng5 {+3.20/11} Rd5 {+3.20/12} 43.
Ne6+ {+3.51/13} Kf7 {-3.38/11} 44. Nxc7 Rd7 {+3.51/12} 45. Kf2 {+3.70/12}
Rh1 {-3.31/10} 46. Be5 {+3.77/12} c4 {+3.88/12} 47. Rxh4 Kg6 {-3.62/9} 48.
Nb5 {+3.88/10} Kf5 {-3.17/8} 49. Rh5+ {+4.04/11} Kg6 {+4.04/11} 50.
Rh8 {+4.11/12} Bf7 {+4.13/11} 51. Nd6 {+4.14/12} Bd5 {-4.42/9} 52.
b3 {+4.26/10} b5 {+4.41/11} 53. Nxb5 {+4.50/11} Kf5 {-4.55/9} 54.
Bb2 {+4.71/10} a5 {-5.02/8} 55. Nd4+ {+5.13/10} Kxf6 {-5.24/8} 56.
Nf3+ {+4.97/10} Kf5 {-5.02/9} 57. Ne5 Rd6 {+4.93/9} 58. bxc4 {+5.53/11}
Be4 {+5.78/10} 59. a4 {+5.89/11} Rb1 {-5.75/8} 60. Rh5+ {+6.12/11}
Kf4 {+6.11/11} 61. Rh4+ Kf5 {-6.22/9} 62. Ba3 {+6.18/11} Re6 {+6.27/11} 63.
Rh5+ {+6.38/10} Kf4 {-6.63/9} 64. Nxd3+ {+6.87/11} Kg4 {-6.86/10} 65.
Rxa5 {+6.87/10} Rf6+ {+7.16/10} 66. Ke2 {+7.19/11} Rb3 {-7.42/8} 67. Bc1
Rc3 {-7.18/8} 68. Ne5+ {+7.18/11} Kh3 {-7.66/10} 69. Bb2 {+7.23/11}
Rc2 {+7.28/10} 70. Rxc2 {+7.48/11}
{Black resigns} 1-0
For example, the 65th and the 69th Ufim's (Black) moves have the {score/depth} information from the wrong (Delfi's) side.

[Alessandro Scotti]

Hi,
the version of Winboard used by Leo Dijksman has an error in saving PGN files.
Sometimes the {score/depth} information is from the wrong side.

Hi Niyaz,
this problem is fixed in the latest version (see post above). BTW only some games were played by Leo with that version because he noticed immediately and I was able to send him a fix several days ago.