RE: [Bug-gnubg] Re: Strange FIBS ratings

[Albert Silver]

> > Not really. The checker error rate was around 0.013 to 0.014, which
in
> > my book is very poor and deservedly described as only Intermediate.
> 
> I'd like to read that book :-)

I presume you're joking, but I was serious. I was inititally spoiled by
Snowie's very generous textual ratings, which was worsened by GNU's even
more generous textual ratings (before the revision), so my head was
about to blow up it got so big. The current textual ratings are much
better and do a much better job of reflecting relative strengths IMHO.
For myself, I consider an overall Advanced rating to be normal, Expert
to be good to very good, and World Class (over more than 4 games) a sign
my glasses are dirty. Below is mediocre to bad. And my rating as stated
on FIBS has been around 1700 or so. In contrast, I played against David
Levy (dlevy on both FIBS and GG) for the Gold Medallion tournament and
who is rated 1900 or so on FIBS and GG. We played 5 matches of 11
points, all close, and though I played 2 Advanced, 2 Intermediate, and
one Expert, he *averaged* Expert in all 5 matches doing a solid 0.004 to
0.005 better than me. This is actually what I would expect of a 1900
player. 

> For example  my "kvandoel" alias over  176 games gets a  GNUBG rating
of
> 1878 +-26 (90% conf. interval). My actual rating is 1910.  Another
alias
> I sometimes play  under over 63 games gets a GNUBG  rating of 1935
+-28,
> actual rating is  1900. Yet another alias got 1916 +-  41 over 25
games,
> actual rating 1860.
> 
> Some data for other players that I've played a lot (I won't mention
their
> alias here, for privacy reasons):
> 
> (Average of 6  ) 2033 +- 30  actual 2070
> (Average of 7  ) 1963 +- 67  actual 2000
> (Average of 8  ) 1795 +- 107 actual 1870
> (Average of 6  ) 1882 +- 134 actual 1960
> (Average of 11 ) 1805 +- 125 actual 1900
> (Average of 10 ) 1756 +- 83  actual 1900
> 
> (I set the rating offset to  2200, compared to 2050 default, so
subtract
> 150 to those number to convert to your convention).

I'm a little confused then. GNU gives you an adjusted rating of how much
then? 1750? The ratings I quoted are all using the 2050 parameter, thus
it is saying I am playing about 200 points worse than its 'perfect'
rating, and my genuine rating on FIBS is 1700 or so.

> As far as  I know the previous error based  classification was not
based
> on facts, the current one is. What  reason do you have to believe in
the
> current relation between level and error rate, which you describe by
the
> word "rightfully", besides you have gotten used to it?

By rightfully, I mean that a match with few mistakes is given a
proportionately high textual rating (Expert, etc.), and a poorly one is
given a proportionately low rating. It's true I *am* used to it, but it
is hard for me to reconcile an error rating of 1850+ with an error rate
calling my play Intermediate. Here is a match I just played against GNU,
over 400 analyzed plies, where I was Intermediate (barely, it's true),
and got a ~1900 rating (using the 2050 reference):

                                      GNU                  Albert

Overall Statistics:
Error rate (total)                    +0.000               -1.838
(-30.462%)   
Error rate (per decision)             +0.000               -12.09 (
-0.200%)   
Equiv. Snowie error rate               -0.00                -4.84

Overall rating                        Supernatural         Intermediate

Actual result                          +50.00%              -50.00%

Luck adjusted result                   +55.29%              -55.29%

Luck based FIBS rating diff.          n/a

Error based abs. FIBS rating          2050.0               1896.0

Chequerplay errors rating loss           0.0                137.4

Cube errors rating loss                  0.0                 16.6



> 
> > Don't get me wrong, I am very appreciative of your efforts, but
right
> > now there are problems that need resolving IMHO.
> 
> You'll  have  to  be more  specific  I  think  to create  a
meaningfull
> argument.  

How can I be more specific than saying my rating is one thing and my
attributed rating another?

If you  are right,  there must  be something  wrong  with the
> results      of      the       simulations      I      presented
on
> www.cs.ubc.ca/~kvdoel/tmp/ratings.
> 
> I'm open  to that possibility,  but if there  is nothing wrong  with
it,
> then your accustomed way of thinking is wrong.
> 
> Anyways, I think  the best thing to do now is  to try to
validate/refute
> the  current  rating estimator  by  estimating  ratings  of players
and
> compare with their actual rating.

That's why I gave you those examples. However, I checked and have my
FIBS matches for the last couple of months still on my HD, so I'll
analyze them and compare the results. I'll also send you the matches. I
don't have my opponents' ratings mind you, but I can say that none of my
opponents was rated below 1650 (IIRC) and most are rated 1700 or more.

BTW, although I have no idea whether this makes any difference, *all* my
matches on FIBS are 5 or 7-pointers, never against a bot, and never
(except once in a blue moon) shorter.

                                                Albert

> 
> 
> Kees