Author: Odd Gunnar Malin
Date: 01:16:54 10/17/05
Go up one level in this thread
On October 15, 2005 at 18:44:59, Jan K. wrote: >Hi, I have the following problem under fritz 8 gui+wb2uci, most probably because >I lack any knowledge about uci. > >White:me;),Black:engine,pondering on, lets play some moves: > >1. e4 Nf6 2. e5 Nd5 3. h3 >...engine is now out of book, after some thinking it plays d6 and starts to >ponder Nf3, printing some output like >... >8 -2 0 5006 g1f3 d6e5 f3e5 b8d7 d2d4 d7e5 >9 -2 0 9592 g1f3 d6e5 f3e5 b8d7 d2d4 d7e5 d4e5 >10 -2 0 17050 g1f3 d6e5 f3e5 b8d7 d2d4 d7e5 d4e5 c8f5 >11 -2 100 259657 g1f3 d6e5 f3e5 b8d7 d2d4 d7e5 d4e5 c8f5 g2g4 >.... >(to be precise:g1f3 is the expected move and possible replies are being >evaluated, the output pv is always <....> <expected move from the opponent> ><best reply> <the rest...>;wb2uci gets the output and converts it to uci) > >Now:when I play a move different from Nf3, everything seems to be ok. If I play >Nf3, engine plays dxe5 and starts to ponder Nf3xe5(correctly,as far as I can >tell). But no output can be seen in gui output window. Gui obviously receives >output from engine, because it changes information about depth and nps, but it >doesn't print principal variations for those depths. And if I play Nf3xe5, it >plays the correct answer(based on pondering) with score 0.01/depth 1, continuing >to ponder next move still with no output. If I play move that's different from >pondered move, output is ok again. > >Thanks for any comments. The problem here is that it have to know the pondermove when it gives the best move. What I do is to take this from the PV and send it to the gui as pondermove. If this later is wrong, either seen by a hint move or in the next pv, I stop the PV line and only send the other data (nps etc.) In your sample it send a move (ponderhit) without sending a PV first, then I don't have any move to send as pondermove and thereby stop the PV. The 0.01/depth 1 is to stop any drawoffer made by the gui. I could probably collect more from previous PV lines to extend this problem a move deeper, but I didn't think of this as a big problem and therby save some cpu-cycles for the engine. Odd Gunnar
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.