New Live Comp-Comp Website!

[hgm]

And how exactly you do that?
Do you set a line inside Winboard.ini that says:
/variant="gothic"

or you just do it for every engine for example:
TJchess10x8 /boardWidth=10 /variant="capablanca"
joker80 /boardWidth=10 /variant="capablanca"
etc....

Or with another way?

I put /variant=gothic in the winboard.ini (I think it doesn't matter if you put quotes around the "gothic" or not.) The variant name now implies a default board size; you don't have to give /boardWidth or /boardHeight anymore for playing Capablanca or Gothic on 10x8.

I have no idea if it would work if you gave this argument per engine. You could try it in the start-up dialog box, but when you do automated play under PSWBTM, you never get the start-up dialog. In general, it doesn't make sense to have engines play a different variant.

And how do you force Winboard_F to play from a PGN? Or PSWBTM does the PGN management?
I guess the latter right?

PSWBTM does the PGN management. When you start a new tournament you can select if you want to play from the default opening position, from a PGN or from an EPD/FEN file, and if you select a file you can enter the file name. PSWBTM does communicate this to WinBoard through the options /loadGameFile="..." and /loadPositionFile="...", and /loadPositionIndex=N to select which position from the file is used. So you could do it from a bare WinBoard as well (otherwise PSWBTM couldn't do it either), but as PSWBTM does it for you, you don't have to bother.

So does that mean that some engines(that don't support Gothic Chess) can't castle?

The castling rules for Gothic and Capablanca are the same, are they not? It is only FRC and CRC that have different castling rules, and they are upward compatible. Janus Chess would be a problem, that really has different castling rules.

[smrf]

Still I cannot find out, which SMIRF version is running there. That would be version MS-171e now. I also miss an ability to review selected games.

[hgm]

I finally merged Tony Hecker's suggested modifications into my own version of Smirfoglot. This Smirfoglot 1.1 now uses the WinBoard 'myname' feature to relay the name and version of the Smirf DLL (so that it appears in my viewer).

What exactly o you mean by "replaying selected games"?

In principle WinBoard should display a list of all games when you click a file with multiple PGNs in it, and you set WinBoard as the default application for the PGN extension. (I suppose that you could acieve the same thing by starting WinBoard, and opening the PGN file through the 'file' menu). You can then click the selected game, and WinBoard will re-play it. If this doesn't work, it should be considered a bug.

[smrf]

Thank you for answering. But another time I failed trying to communicate what I intended. May be there will be more successful moments one day.

[TonyJH]

Reinhard, in case you missed it, there is a link to a PGN file from this page:
http://home.hccnet.nl/h.g.muller/battlehome.html

However, it doesn't have the games from the current test tourney.

It's SMIRF version MS-171d playing now, if I remember correctly. (I can't check right now because other programs are playing).

[smrf]

Thank you!

[hgm]

Reinhard, in case you missed it, there is a link to a PGN file from this page:
http://home.hccnet.nl/h.g.muller/battlehome.html

However, it doesn't have the games from the current test tourney.

It's SMIRF version MS-171d playing now, if I remember correctly. (I can't check right now because other programs are playing).

Sorry, I forgot to update this link.

Currently I store all games in files with the name of the tourney with .pgn extension appended. So if the link still points to games of a previous tourney because I forgot to update, you can always read out the link (rightclicking it, and asking for 'properties'), paste that link in the address field of your browser, change the name to that of the tourney you want (e.g. the one that is currently running, TT4), and type enter.

I guess in the future I will delete old uninteresting tourneys, and zip the interesting ones and put them on my main website. (Currently they are coming directly from the PC where the games are played.)

Note that the current Smirfoglot sometimes hangs after a game (using full CPU), probably because of a faulty partial attempt at implemennting the 'move now' command. I had to kill two rogue processes this morning. I will try to fix the problem tonight.

[smrf]

Thank you, Harm! I luckily noticed, that also FEN and SetUp now are included in PGN , that is optimal! But I also noticed, that in a draw game between SMIRF and BigLion80 there has been a repetition draw claim even before the third repetition move had been possible. SMIRF sometimes is making some time by repeating moves but by avoiding a 3-time repetition. Thus there might be a point of review in the Winboard_F GUI.

[smrf]

Well, I also found a lost game between SMIRF and BigLion80 there, which is starting unbelievable for both sides ...

[hgm]

From looking at the moves it seems like a regular 3-fold repetition to me: the white C is moving back and forth between e5 and f5 2 times, and the black King between d6 and d7. So after 44. ... Kd6 the same position as after 40. ... Nd5 is created, making it the third occurrence. (The second was after 42. ... Kd6.)

I did not really change anything about the FEN/Setup treatment: The FEN is included here because the variant is capablanca, but I start from the gothic position as a kludge to make TSCP-G play capablanca. TSCP-G ignores the variant command, while some other engines do not support variant gothic. So to bring all engines in the same state I set variant to capablanca, and load them with a gothic FEN. Then I load moves from a PGN to turn the gothic position into the opening positions from which I want to start.

This is all needed because TSCP-G also doesn't support 'setboard'. This means that to play TSCP-G I can only use opening positions that can be derived from the gothic one by reversible moves. So I cannot even include the regular capablanca position. This makes it all very cumbersome. I wrote an e-mail to Michel Langeveld to ask if he could fix this, or send me sources so that I could fix it, but no reply so far.