2008/11/17 Alexander Wagner <[EMAIL PROTECTED]>
> Pascal Georges wrote:
>
> Hi!
>
> > I don't stick by heart to it. I'd suggest to port over the
> > statistics, maybe also an evaluation at the end, say, the
> > line played with the annotations from the mask, all the ? an
> > ?! one moved and maybe the commentary if any. Just to give
> > some feedback where it might have went wrong and where to
> > have a deeper look into the positions.
> >
> >
> > So there are two things here :
> > 1. Statistics
> > 2. Feedback when playing in Tree's training mode.
> >
> > 1. For statistics : of what use is : played 56 moves, 32
> > in mask, 5 dubious, coverage of Mask is 14 % ?
>
> I agree that the pure numbers alone are only some sort of a
> start. They might only get a hint about the overal idea of
> where to go. Well, if I get 25 ?? positions out of 25
> reached, this is some significance, that I don't know the
> ideas.
I still don't think those stats are of any value, so I will not add them.
But if someone feels the need to do it, that's ok.
>
>
> Anyway, 5 doubious might have the problematic position
> "behind it", say Scid could store the position where the
> doubious moves occured (?!), the position where the blunders
> occured (?, ??), or where the player has chosen an
> interesting move (!?). I can imagine a sort of report of
> the kind:
>
> 5 ?? Pos 1, Pos 2
> 10 ? Pos 3, Pos 4, Pos 5, Pos 6, Pos 7
> 7 ?! Pos 8, Pos 9
> 15 !? Pos 10
>
> And here I imagine to click on the postions to show them for
> further study. I think this would be some valuable feedback
> as the mask should contain stronger moves etc.
>
> I could also imagine giving something like the key
> positions, ie. positions that are reached by different play
> again and again, something like
>
> Key positions:
>
> Pos 1: 37 times
> Pos 2: 20 times
> Pos 3: 5 times
>
> I think as the tree is based on position it might help to
> sort things a bit. I've transpositions in mind here.
>
> Also it might give hints if there are positions with
> somethign like:
>
> move 1: !!
> move 2: !
> move 3: !?
>
> and the player always choses move 2 only.
UI must be kept simple and I want to filter out things in order to keep only
stuff that has a real value in the learning process. For example I met a
chess teacher last days that said he has given up the use of chess engines
for 18 months (he is a 2200 - 2300 FIDE player) ... It seems weird and
extreme at first glance, but I think the idea behind this is to sharpen
one's own mind without artificial external tool even if the drawbacks are
obvious. I want something that emphasizes on understanding, not memorisation
and the user should get by himself where he has problems, and there is no
need of stats for that.
So when in training mode, my idea is just to add extra info in PGN window
(like the info you get when annotating a game, that is when the user
selected a move out of the mask or with a bad value).
>
>
> > 2. For feedback : many options should be available, and I
> > think it is better to let the user play 14 moves even if
> > move #5 with a ?? The moves are recorded in current's
> > game so the user can review it and understand what has
> > gone wrong instead of a pop up that is a bad thing I think
> > in such a learning phase.
>
> Right, but it might be worth to add the ?? at move #5
> already in the game record. What do you think about that?
Yes, I will do it.
>
> And Scid might well remember that the user chose the ?? at
> #5 already and point it out again in an overall "statistic".
No, because I don't feel it is needed to keep an info that may be obsolete
at any time. Suppose you discover an opening, make a wrong #5 move (??),
understand it and play again a couple of times this opening without making
any longer the error : what is the interest of the initial record ? How do
you reset those stats ? What to do if you update your mask and like me you
switch 7 ... Qc7 in the French Winawer from "!?" to "?!" ?
I prefer to know by myself that 7 ... O-O is better than 7 ... Qc7 (from my
point of view) and no stats will never give me this information.
Pascal
-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Scid-users mailing list
Scid-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/scid-users
