I agree assigning skill levels to backgammon players is a very complex
matter and what is difficult varies from player to player. As to your
position you must remember that gnubg counts close doubles and not
non-obvious cube actions. In the position the cube actions may be
non-obvious, but they
I didn't mean it was close with my example. Perhaps I've
misunderstood, but I was under the impression that these 'close calls'
really meant what range of equity loss it would use to consider a
double into the number of decisions it judged a human made for the
rating. In other words it would
Sure, but I think that 0.160 is to tight a window. Another example,
which came to mind as soon as I sent the last note, was on holding
games, or 3-pt anchors. Often the defending side is down by quite a
bit in the race (10-20 pips) and the other side has escaped the
checkers. It can be a No
The threshold of 0.25 seemed too large for common sense, far too many
doubles very counted as being close, but also the function was plain
stupid and hard to understand. For example all too-good positions were
counted in for some reason. The threshold could be set back to 0.25
for counting of the
On 9/4/06, Christian Anthon [EMAIL PROTECTED] wrote:
The threshold of 0.25 seemed too large for common sense, far too manydoubles very counted as being close, but also the function was plainstupid and hard to understand. For example all too-good positions werecounted in for some reason.
I see...
what is the threshold now?
0.16 mainly chosen for some internal consistency (0.16 is the standard
threshold of VERYBAD) and because 0.25 returned too many silly cubes,
which is important because the function is used for choosing which
positions to export as closed doubles too.
0.16 is perhaps
I think it is rather complicated. Admittedly, some doubles with such
an equity difference can seem absurd, but many will leave players very
much in doubt, especially if they are dependent on the score. Here is
an example:
GNU Backgammon Position ID: 7LYBAGPN3QYgIA
Match ID :
For good order I checked the code, and it is as I suspected, the close
cube decisions that is used for the calculation of the absolute fibs
rating (which is what I read from Kees post as well). That was the
function I changed and that implies that the absolute fibs ratings are
no longer valid.
On 9/3/06, Christian Anthon [EMAIL PROTECTED] wrote:
For good order I checked the code, and it is as I suspected, the closecube decisions that is used for the calculation of the absolute fibsrating (which is what I read from Kees post as well). That was thefunction I changed and that implies that
I don't think the formulas are, but 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
This is as it should be. Most (all?) players will observe that they
have lower error rates in short matches without so many cube
descisions and gammons not counting in many games. So since a players
error rate depends on the matchlength so does the translation of error
rates to fibs rating.
I do
This is as it should be. Most (all?) players will observe that they
have lower error rates in short matches without so many cube
descisions and gammons not counting in many games. So since a players
error rate depends on the matchlength so does the translation of error
rates to fibs rating.
-Original Message-
From: Christian Anthon [mailto:[EMAIL PROTECTED]
Sent: Thursday, August 17, 2006 5:04 PM
To: Albert Silver
Cc: bug-gnubg
Subject: Re: [Bug-gnubg] Removal of absolute fibs ratings
I think I understand the background for the translation. The problem
The absolute fibs ratings are based on error ratings for cubes and
moves combined. After I sanitized the isclosedouble count the absolute
fibs ratings doesn't make sense anymore and should therefore be
removed.
Personally I doubt that the absolute fibs ratings ever made sense, but
I know that
, August 17, 2006 12:59 PM
To: bug-gnubg
Subject: [Bug-gnubg] Removal of absolute fibs ratings
The absolute fibs ratings are based on error ratings for cubes and
moves combined. After I sanitized the isclosedouble count the absolute
fibs ratings doesn't make sense anymore and should therefore
15 matches
Mail list logo
