xboard-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [XBoard-devel] Desired behavior: POV of scores


From: Tim Mann
Subject: Re: [XBoard-devel] Desired behavior: POV of scores
Date: Wed, 25 Nov 2009 10:41:41 -0800

I agree with all of that.  The -*ScoreIsAbs thing is new since 4.2.7; I
don't know when it was added.  I don't have an opinion about whether
it's worthwhile having a user option about which POV to use in
display.  We have a zillion options already, so if some people want
this, I don't see any harm in providing it -- we already lost the
"keep it simple" battle years ago.  :-)

On Wed, 25 Nov 2009 09:51:09 +0100, "h.g. muller" <address@hidden> wrote:
> There was some discussion on TalkChess on the Point Of View that should
> be used for score reporting. Although it is not mentioned in the specs,
> de-facto standard is that engines always report scores from the POV of the
> side to move.
> 
> This leads to a bit of a chaotic situation (which I do not want to address at
> this moment in case the engine is pondering), as some engines ponder the
> position after a speculative expected opponent move (so that their ponder
> scores are from the POV of the side they are actually playing), while others
> ponder the position as it is, and report scores from the POV of the opponent.
> 
> What I do want to address is score reporting while the engine is thinking /
> analyzing. There are engines which systematically report scores from the
> white POV. In my opinion, this is a non-compliancy, but to be able to 
> accomodate
> such engines in cases where the score is used for adjudication, WinBoard
> provides the option "-first(second)ScoreIsAbs true|false". In 4.4.1, 
> setting this
> option causes a score flip in MachineBlack mode, or for the engine playing 
> black
> in TwoMachines mode.
> 
> Now it seems to me that it should be counted as a bug that this is not 
> extended
> to other modes where the engine is evaluating a black move: IcsPlayingBlack
> in zippy mode, and the various modes where an engine is analyzing 
> (AnalyzeMode,
> AnalyzeFile or IcsObserving in zippy mode when the engine is analyzing the
> observed game), and black has the move. Also there the standard is that 
> engines
> should report from the POV of the side to move, so that engines that report 
> from
> the white POV need their score flipped to correct their non-compliancy.
> 
> Do we agree about that?
> 
> (In these new mode the score is not really used for anything like 
> adjudication,
> so it might not seem to matter that much, but the score is presented to the 
> user,
> and one of the major reasons why users would use AnalyzeMode at all, so I 
> guess
> this should count as a pretty important use.)
> 
> Now a second, and independent question is this:
> 
> Once all engines are compliant w.r.t. score reporting, or can be made 
> compliant with
> the ScoreIsAbs option, what POV do we want to use for presenting this score 
> to the
> user. It seems that the POV of the side to move, which is used now, is 
> compatible
> with the EPD standard. There seem to be many users that hate it, though. 
> Should
> this be a user-settable option? We could, for instance, have a check box in 
> the
> Engine-Output window that would enable the user to select white POV or stm 
> POV.
> 
> Is this worth having?
> 
> 


-- 
Tim Mann  address@hidden  http://tim-mann.org/




reply via email to

[Prev in Thread] Current Thread [Next in Thread]