gnugo-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[gnugo-devel] rating inflation on NNGS


From: Gunnar Farneback
Subject: [gnugo-devel] rating inflation on NNGS
Date: Wed, 16 Jul 2003 22:40:17 +0200
User-agent: EMH/1.14.1 SEMI/1.14.3 (Ushinoya) FLIM/1.14.2 (Yagi-Nishiguchi) APEL/10.3 Emacs/20.7 (sparc-sun-solaris2.7) (with unibyte mode)

GNU Go is now 8k* on NNGS. This is not due to a sudden improvement of
GNU Go but is caused by a general rating shift of 2 stones on NNGS.

The explanation for this is given by the command "help new-ratings",
as follows:

|   Ratings raised 2 stones on July 16, 2003
| 
| During the spring a poll was made among some of the NNGS players about
| their view of the level of the NNGS rating system. As you know, the
| level is "low" compared to most national ranks and other go servers.
| 
| The result of the two relevant questions in the poll can be seen below
| (there were also reference questions asked about other known ranks and
|  nationality). As seen, a clear majority thinks that the ratings were
| too low and should be raised by 1-2 stones in most cases.
| 
| Based on this, we have now raised the ratings 2 whole stones.
| This is of course mostly a "cosmetic" change, a matter of psychology
| (as some people recent having a rating of, say 1k, when they are 2d
| in "real life"), but not only that. We also think it's impractical to
| have a rating that's differs too much from the rest of the world.
| Having a rating level in tune with the rest of the world makes it
| easier for people to get an accurate rating from the start.
| 
| Note: This does not affect the relative ratings or the handicap you should
|        use. It does however change things for players who have set a rank
|        with the "rank" command, but have not yet got a rating. They should
|        adjust the rank accordingly.
| 
| --
| The poll result:
| 
| 1) I belive the current NNGS ratings are
|      A) ok              18%
|      B) too low         56%
|      C) too high         0%
|      D) not important   25%
|      no answer           1%
| 
| 2) (If B or C above)  I think the ratings should be
|      raised by  0.5 stone     16%
|      raised by  1   stone     29%
|      raised by  1.5 stone      4%
|      raised by  2   stones    35%
|      raised by  3   stones    13%
|      raised by  4   stones     3%

/Gunnar




reply via email to

[Prev in Thread] Current Thread [Next in Thread]