Moderator: Ras
I can email you the debug files for each game once I get home from work later if you wish.Stan Arts wrote:Nice picture.
However the problem remains the same.
Could you paste the "level" command that is sent to the engines as that (a horribly mangled level command, a level command each move, who knows) is the other explanation. But a lot of programs will have problems then.
Four debug files for you. Hope that they give you the information you're after.Graham Banks wrote:I can email you the debug files for each game once I get home from work later if you wish.Stan Arts wrote:Nice picture.
However the problem remains the same.
Could you paste the "level" command that is sent to the engines as that (a horribly mangled level command, a level command each move, who knows) is the other explanation. But a lot of programs will have problems then.
Thanks Graham,Graham Banks wrote: Four debug files for you. Hope that they give you the information you're after.
http://kirill-kryukov.com/chess/discuss ... p?id=33380
http://kirill-kryukov.com/chess/discuss ... p?id=33379
http://kirill-kryukov.com/chess/discuss ... p?id=33378
http://kirill-kryukov.com/chess/discuss ... p?id=33377
I'm hoping to start a Nemeton gauntlet soon. How far away is your next version, as I'll wait if it's coming soon?
Cheers,
Graham.
Stan - I'll point out what you've said to Matthias so that he can respond.Stan Arts wrote:.....The level command looks fine. level 40 25 0 as expected.
So it took a good look to figure out what is going on then.
It looks like ChessGUI plays out openingbook moves internally, then passes a setboard command with the resulting position, without even passing the moves played, or sending a new level command afterwards.
That's the problem, because Nemeton regards a setboard position as a new game.
That's a creative way to break it. Good game Matthias.
Surely StansChess and Neurosis never got that right either.
New version probably around christmas but not much use to testing in this setup till this is fixed. Nemeton not strong enough to hand out time odds.
PS. Self fulfilling titles
I don't think most other engines reset time control after setboard, since the documentation doesn't say setboard is supposed to indicate a new game at all. It only says to reset the clocks after new.Stan Arts wrote:Thanks Graham,Graham Banks wrote: Four debug files for you. Hope that they give you the information you're after.
http://kirill-kryukov.com/chess/discuss ... p?id=33380
http://kirill-kryukov.com/chess/discuss ... p?id=33379
http://kirill-kryukov.com/chess/discuss ... p?id=33378
http://kirill-kryukov.com/chess/discuss ... p?id=33377
I'm hoping to start a Nemeton gauntlet soon. How far away is your next version, as I'll wait if it's coming soon?
Cheers,
Graham.
By the way hash clearing/aging between moves changed for the updated version I can see by the search characteristics that you were indeed using the updated version. My apologies!
The level command looks fine. level 40 25 0 as expected.
So it took a good look to figure out what is going on then.
It looks like ChessGUI plays out openingbook moves internally, then passes a setboard command with the resulting position, without even passing the moves played, or sending a new level command afterwards.
That's the problem, because Nemeton regards a setboard position as a new game.
That's a creative way to break it. Good game Matthias.
Surely StansChess and Neurosis never got that right either.
New version probably around christmas but not much use to testing in this setup till this is fixed. Nemeton not strong enough to hand out time odds.
PS. Self fulfilling titles
Code: Select all
new
level x y z
setboard ...