Author: Christophe Theron
Date: 10:42:50 03/18/04
Go up one level in this thread
On March 18, 2004 at 00:17:26, martin fierz wrote:
>On March 17, 2004 at 15:08:55, Christophe Theron wrote:
>
>[snip]
>
>>I think my formula is more accurate. And actually it's not _my_ formula. I have
>>found it somewhere, but I do not remember where.
>>
>>The real elo formula has the interesting property to be close to linear in the
>>20%-80% winning percentage range, hence the 80% validity limit.
>
>it's interesting that you find this interesting :-)
>your linear formula is simply a taylor approximation to the first order; it is
>always possible to do this, for any function in the vicinity of some point
>(except of course if the function is singular at that point).
It is always possible, but it does not happen so often that the linear
approximation works so well on a wide range of useful values.
So let's celebrate this! :)
> both you and uri
>get the coefficient wrong, because it's more or less right in the middle between
>7 and 8; so some people use 7, some use 8, because it seems nobody (except
>dieter) can remember the real value of the first taylor coefficient, which
>doesn't happen to be a whole number - of course...
>
>cheers
> martin
All I need to remember is that 7 does the trick very well in the 30%-70% range.
I will only have to grab my Palm and use the real formula when it's outside this
range, which will not happen often.
Christophe
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.