gt; Michael Petch
> GNU Backgammon Maintainer / Developer
> OpenPGP FingerPrint=D81C 6A0D 987E 7DA5 3219 6715 466A 2ACE 5CAE 3304
>
>
> ___________
> Bug-gnubg mailing list
> Bug-gnubg@gnu.org
> https://lists.gnu.org/mailman/listinfo/bug-gnubg
>
--
Robert-Jan Veldhuizen
___
Bug-gnubg mailing list
Bug-gnubg@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-gnubg
help page
concerning the hyper databases?
Regards,
--
Robert-Jan Veldhuizen(Zorba on FIBS, Tiger and DailyGammon)
___
Bug-gnubg mailing list
Bug-gnubg@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-gnubg
s what's being
used for the luck-adjusted result, BTW).
For the rest, I agree with Massimiliano that the luck rates expressed in
EMG/move make very little sense, as I wrote in a different post already.
Greetings,
--
Robert-Jan Veldhuizen
___
Bug-gnubg mailing list
Bug-gnubg@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-gnubg
-- Forwarded message --
From: Robert-Jan Veldhuizen
Date: Sat, Jul 25, 2009 at 4:19 PM
Subject: Re: [Bug-gnubg] Base of luck
To: Frank Berger
I agree with luck being represented in MWC.
Also, I think the "luck per move" should be either removed, or at least, not
be
uses the cubeless values from the
database FOR THAT PLAY.
Conclusion: from 1-ply on, GnuBG is making some different play(s) in this
bearoff at -3,-3 than for money; at least it seems to dothis with teh cube
still owned on 2.
--
Robert-Jan Veldhuizen
___
Bug-gnubg mailing list
Bug-gnubg@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-gnubg
% MWC through skill:
>>
>> Luck adjusted result-17.57%
>>+17.57%
>>
>> or that he wins 32.43:67.57% of the matches. Whether these numbers are
>> good or meaningful I don't know.
>>
>
> Are sure about this?
be usage (after a miss).
Greetings,
--
Robert-Jan Veldhuizen
___
Bug-gnubg mailing list
Bug-gnubg@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-gnubg
(Errors) in points-8.329 (-7.290) -
> 3.549 (-3.441)
> EMG Error (Error) rate in mpoints/unf.decision -39.0 (-34.5) -
> 16.0 (-15.9)
> EMG error rate in mpoints/decision (Snowie) -19.4 -
> 9.2
>
> Luck Statistics:
> Total
t don't take too
long, the advice is simple: use even-ply both for cube and checker play;
preferably 2-ply for both (assuming you can't afford 4-ply or higher...).
Regards,
--
Robert-Jan Veldhuizen
___
Bug-gnubg mailing list
Bug-gnubg@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-gnubg
w/g/bg are still basically the same (as they would be in a cubeless
("dead cube") rollout), except that any (re)double/passes during the rollout
will use the evaluation breakdown at that point (i.e. truncated). The "live
cube" is reflected in the cubeful equity.
> Another re
ssed."
So, gammons might change these numbers I guess.
--
Robert-Jan Veldhuizen
(Zorba on FIBS)
___
Bug-gnubg mailing list
Bug-gnubg@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-gnubg
real life backgammon is concerned. This subject was discussed on GOL
once, and from what I remember there was consensus about this relation
beween early/late and pass/take ratios.
If anyone knows of a more theoretical approach to this problem or comes up
with
> hint offer to take?
Because when you drop, you also lose: 1 point right away. Without
considering gammons, the basic takepoint is therefore at 25% game winning
chances. Because after taking, you have a chance to redouble while opponent
cannot double you anymore, the gammonless takepoi
verall error statistics, but
I've been wondering sometimes if they could distort a luck analysis?
--
Robert-Jan Veldhuizen
___
Bug-gnubg mailing list
Bug-gnubg@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-gnubg
rtain to run into crashing trouble at some point. So I've
made it a habit to NEVER try and cancel/edit opening moves, as a
work-around.
Hope this helps the GNUbg team,
--
Robert-Jan Veldhuizen
___
Bug-gnubg mailing list
Bug-gnubg@gnu.org
http://lis
just forget about 3-ply at all since
it takes a lot longer than 2-ply and it's not clear that it's any
better. For a real improvement over 2-ply you need 4-ply it seems (but
even then the improvement is very small, and the time cost is huge).
--
Robert-Jan Veldhuizen
___
ally seem on the low side, especially for long races. This is
consistent with the earlier rollout I posted.
Hope this at least sparks some interest.
--
Robert-Jan Veldhuizen
___
Bug-gnubg mailing list
Bug-gnubg@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-gnubg
position type characteristic, which is
essentially what 0-ply tries to do and which is what I'm looking at.
So as an example, gnubg 0-ply queries the neural net and gets 68% wins, 16%
gammon wins, 32% losses, 2% gammon losses (cubeless equity: 0.50). Does it
double a center cube for mone
would be a great start as well.
That sounds interesting. What is the difference between your
algorithm/formula and GNUBG's present algorithm/formula?
--
Robert-Jan Veldhuizen
___
Bug-gnubg mailing list
Bug-gnubg@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-gnubg
dom
dice
Play: 0-ply cubeful prune [expert]
Cube: 2-ply cubeful prune [world class]
--
Robert-Jan Veldhuizen
___
Bug-gnubg mailing list
Bug-gnubg@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-gnubg
by quite a bit. Maybe this could actually be
improved quite easily?
Of course the question is if you change 0-ply cube behaviour, how will it
affect 2-ply cube behaviour?
--
Robert-Jan Veldhuizen
___
Bug-gnubg mailing list
Bug-gnubg@gnu.org
http://
ssions, EMG (normalised) and real points; in matches in EMG and in %MWC. The overall statistics show luck calculated with real points and %MWC.
On the luck tab you can see both values (plus a rather useless luck per move figure... perhaps this should simply be removed?).
-- Robert-Jan Veld
wer's interpolation for the Thorp count.
Here's a link to the summary of the Keith Count, lots of other info are on the rest of the same page:http://www.bkgm.com/articles/CubeHandlingInRaces/#summary_of_the_keith_count
Cheers,-- Robert-Jan Veldhuizen
__
r without introducing all sorts of other problems.-- Robert-Jan Veldhuizen
___
Bug-gnubg mailing list
Bug-gnubg@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-gnubg
I see... only counting close" too goods sounds like a good improvement. So what is the threshold now?-- Robert-Jan Veldhuizen
___
Bug-gnubg mailing list
Bug-gnubg@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-gnubg
ould* start to see an equalizing of the
statistics.Not really, the variance with backgammon games is quite large. A result like 200-130 (+70) after 150 games (with cube) or something is not extremely rare.Greetings,
-- Robert-Jan Veldhuizen
___
Bug-gnubg m
rate results for the absolute rating again.
-- Robert-Jan Veldhuizen
___
Bug-gnubg mailing list
Bug-gnubg@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-gnubg
t I'll leave things as they are from now.Christian.On 9/2/06, Robert-Jan Veldhuizen <[EMAIL PROTECTED]> wrote:> Hello Christian,
>> You wrote:>> >The absolute fibs ratings are based on error ratings for cubes and> >moves combined. After I sanitized the isclosedouble count the a
ormula is based on the sanitized "isclosedouble" count so there should be no problem with that.
In my experience, these FIBS ratings are actually one of the most reliable indicators of one's true rating in the long run so I definitely don't want t
Hi GNUBG team,See subject line... Windows
XP.. build 24 May 2006.. I cannot paste text from the clipboard into
the commentary field: nothing happens.Greetings,-- Robert-Jan Veldhuizen
___
Bug-gnubg mailing list
Bug-gnubg@gnu.org
http://lists.gnu.org
Hi GNUBG team,See subject line... Windows XP.. build 24 May 2006.. I cannot paste text from the clipboard into the commentary field: nothing happens.Greetings,-- Robert-Jan Veldhuizen
___
Bug-gnubg mailing list
Bug-gnubg@gnu.org
http://lists.gnu.org
31 matches
Mail list logo
