Author: Tord Romstad
Date: 03:57:10 11/06/05
Go up one level in this thread
On November 06, 2005 at 06:16:25, Reinhard Scharnagl wrote: >On November 06, 2005 at 03:56:49, Tord Romstad wrote: > >>... However, many programmers *want* to do things differently in FRC >>compared to normal chess. ... > >If there would be a need to, you simply could check, whether there are castling >rights without having involved King and Rook at traditional places or not. This works in some positions, but not in all. An example is the following position: [D]4k2r/5ppp/5n2/8/1p6/1P6/1P6/1P3PPP/B3K2R w Kk - 0 1 This position is clearly from an FRC game and not a normal chess game, but your rule above is not sufficient to determine this. It would of course be possible to design much more complicated rules which would handle a bigger percentage of all possible positions correctly, but what's the point? It is much easier if the GUI can simply tell the engine what game it is playing. The given position is also an example of a situation where I evaluate differently depending on the game type, by the way. In FRC games, I have a big penalty for the white bishop in the corner blocked by the friendly pawns on b2 and b3. This pattern is not used in normal chess, for the simple reason that such positions can never arise there. >It is a very bad behaviour to demand for unnecessary and redundant information. What information is redundant depends on the program. For me, the game type is a very important piece of information. >My opinion is, that an engine's basic behaviour should only depend on the input >situation and not on any artificial imposed on view. You are entitled to your opinion, but this is a very subjective point. Again an analogy to telling the program that it is analysing rather than playing a game is appropriate. Some people think that programs should always give the same evaluations and main lines when analysing and playing, and that there shouldn't be any need to tell the engine that it is analysing. You cannot expect to change the basic design decisions of all chess programmers who decide to add FRC support to their engines. As long as a protocol doesn't inform the engine about what kind of game it is playing, a lot of programmers will avoid using the protocol for that reason alone. Tord
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.