Scores from white's POV or from engine's POV?

[mhull]

Viewing the engine as a virtual player, it seems right to have it's score from it's own view, helps add personality.

For analysis you'd probably prefer the white black thing. So perhaps you could go on that, depending if it will be used most for play or analysis.

Stan

Now there's an idea. And to take it one step further, when pondering, reverse this again so that you display it from the opponent's point of view.

Then _nobody_ could understand the scores...

Indeed, I was very confused by this exact behavior when I modified GNU Chess to post it's thinking in an ICS terminal window. It alternated the POV score between thinking and pondering, so during fast games, glancing at the ICS window doesn't tell you if the score is rising over time or dropping. It also displays this confusion in the show-thinking box in xboard between thinking on its own time and the opponent's time.

If I had the time, i'd change GNU Chess to only show white perspective scores.

[Dirt]

Which reminds me, I'm a bit disappointed that only programmers respond in this thread. It is not our opinions which count. I hope not everyone who has voted is a programmer.

What's to say? I prefer scores from white's perspective, but I already said that by voting. Reasons are kind of pointless since almost everyone has a preference and isn't likely to change it.

[krazyken]

The most common notation in regular chess, in newspapers and in books where most people learn about chess, score + as good for white, - good for black. Opening databases also use this convention. Chess engines may be the last group in the world to come around to this.

[Tord Romstad]

Which reminds me, I'm a bit disappointed that only programmers respond in this thread. It is not our opinions which count. I hope not everyone who has voted is a programmer.

What's to say? I prefer scores from white's perspective, but I already said that by voting. Reasons are kind of pointless since almost everyone has a preference and isn't likely to change it.

I agree, there isn't much to say. I just hope that the majority of readers didn't interpret this as a poll meant for programmers, and that the people who have posted in this thread are not representative for the people who have voted.

So far, it looks like there is a big majority in favor of scores from white's point of view, so I'll probably change to that.

Thanks for voting, by the way.

Tord

[Tord Romstad]

The most common notation in regular chess, in newspapers and in books where most people learn about chess, score + as good for white, - good for black. Opening databases also use this convention. Chess engines may be the last group in the world to come around to this.

I think this is because it is usually the programmer who makes the choice, and programmers may have a different point of view from ordinary users. When we programmers watch our engines play, we want to see whether or not our engines think they are better. This is easier when the score is from the engine's point of view. Users, on the other hand, either play against chess engines, or use them for analysis. When they play, they usually don't look at the engine output, and the sign convention is therefore irrelevant. When they analyse, they want to know which side is better, and scores from white's point of view make most sense.

Tord

[hgm]

Indeed, I was very confused by this exact behavior when I modified GNU Chess to post it's thinking in an ICS terminal window. It alternated the POV score between thinking and pondering, so during fast games, glancing at the ICS window doesn't tell you if the score is rising over time or dropping. It also displays this confusion in the show-thinking box in xboard between thinking on its own time and the opponent's time.

If I had the time, i'd change GNU Chess to only show white perspective scores.

This is indeed very annoying. As unanymous as WinBoard engines are in reporting thee score from their own search from their own POV, as divided they are in the sign of the score during ponder. I really think this should be outlawed. How the user wants the score displayed is a matter of GUI options. But they can only be implemented if the engines stick to a standard.

[kranium]

what about offering a choice to the user via a UCI option (or INI file option)

-> score: absolute or relative...

from a programmers standpoint it should not be difficult,
i.e. adding a + or - (flipping) before the displayed value...

[Tord Romstad]

what about offering a choice to the user via a UCI option (or INI file option)

-> score: absolute or relative...

from a programmers standpoint it should not be difficult,
i.e. adding a + or - (flipping) before the displayed value...

There is no INI file, and if there were, there would be no way for the user to edit it. A UCI option would also be impossible, partly because it isn't a UCI engine (Apple don't allow programs to start any additional processes, so I can't do UCI), and partly because UCI does specify the sign of the score, as any sane protocol would do.

I could have a preference setting for the sign of the score (as I do in the desktop version of my program), but as I explained in the initial post, I don't want to clutter the preferences dialog in the handheld version with too many settings, because it quickly makes the user interface awkward to use.

Tord

[kranium]

oh i thought you asking about regular wb and uci engines...
if the question concerns handheld devices, please disregard.

[ernest]

I prefer scores from white's perspective,

So do I, this way the "evaluation profile" diagram of a game is clear!

But of course I am not against a possibility of choice...