From: Ismail Sloan <ishipres <at> BEST <dot> COM>
Date: 27 feb 1997
Subject: Re: shogi ratings/rankings

At 03:28 PM 2/25/97 -0500, Larry Kaufman wrote:
>     Currently there is discussion in the U.S. about converting the
>Washington D.C. shogi rating system into a national one.
....
  The
>question I ask is: if a player who typically scores 50% against other
>players of his rank (2 dan or above) plays even games against average
>players of the next lower rank, what percentage will he score?
>     Thanks in advance for your opinions.
>
>

I am glad to see that Larry Kaufman has finally joined the 20th century and
has gotten e-mail.

As a shodan player in Japan, my score against 1-kyu players was something
like 99%, give or take a percentage point or two. It was just about
impossible for me to lose to them, no matter how badly I played.

I came back 8 years later without playing even a single game of shogi in the
meantime and was pronounced a 2-dan (rating inflation?). I hardly ever lost
to a shodan although I occasionally did, perhaps 10 or 15% of the time. I
won 100% of my games against 1-kyu players even while giving them kyo
handicap. On the other hand, I almost never defeated a three dan player.

The real problems are several. I do not think anybody believes these
American ratings. I certainly don't, especially since they are based on what
people from Japan say that they are and it is known that Japanese are
egomaniacs who wildly overestimate their own stength.

Another problem is that in Japan ratings never go down, only up, plus you
can buy a ranking and so on.

If you win 8 games in a row, you get promoted in Japan. While this is a
tough standard to meet, anybody who plays 256 games should win 8 in a row by
just the random flip of a coin.

So the problems are real. I hope that Larry can solve them.

Sam Sloan