From: Larry Kaufman <lkaufman <at> COMCAST <dot> NET>
Date: 10 dec 2002
Subject: Re: TPR [was: Report on the Terzo Campionato Italiano Open di Shogi]

----- Original Message -----
From: "SECELLE JOHANNES F L A" <johannes.secelle <at> MINFIN <dot> FED <dot> BE>
To: <SHOGI-L <at> TECHUNIX <dot> TECHNION <dot> AC <dot> IL>
Sent: Tuesday, December 10, 2002 9:19 AM
Subject: Re: TPR [was: Report on the Terzo Campionato Italiano Open di
Shogi]


> This has to be looked up in the Arpad's Elo Elo-table.
>
> I remember that, for a difference of 100 rating points, the
winn-loss-percentage
> for the stronger player is 68-32 (if my memory serves well :-).
>
     I'm afraid your memory does not serve well.  Since a 2-1 ratio is 120
points, 100 points must be somewhat less, around 64-36.
> a win-loss percentage of  70 pct (or scoring 7 out of  10) would
correspond
> to +/- 112 elo points (but i would have to look it up in my library at
home to be
> sure)
     And so 70% would be somewhat more than 120, around 145 I believe.  I
didn't check the actual table, but I'm sure this is quite close to the
correct figure.  I actually independently "invented" the use of the logistic
function for chess ratings about the same time as Elo did, and used it for
club ratings before it was published, but since Elo published the idea first
he deserves the credit (unlike the case of Fischer and increment, my other
discovery).
>
> Cheers,
>
> Hans
>
     Larry Kaufman