From: Larry Kaufman <lkaufman <at> WIZARD <dot> NET>
Date: 27 feb 1997
Subject: rating of handicap games

     This posting is to answer some questions about the rating of handicap
games.  George Jost asked why we don't use the Japanese Amateur rating
system?  The answer is that we do, pretty much.  The handicap values that we
used were originally identical to those that they published in an
explanation of their rating system over a decade ago, although I don't know
if in practice they need to rate handicap games, as there are so many
players that even games are easy to arrange.  We simply extended the scale
beyond two piece, and recently modified some of the values slightly to
conform to our experience.  However, both the Japanese amateur Renmei and
Europe use 100 point classes while we use 200, because 100 does not agree
with the normal practice in most Japanese clubs (and Shogi Renmei) where
each rank is quite wide.  Using 100 will cause shodan to be quite difficult
to earn (compared to Japan) and will cause 4-5 dan to be too easy.  If we
assume that our rating scale and Europe's are comparable (data is lacking,
but they are at least not way out of line, I think), then the scales agree
on 2 Dan, but we are more lenient at 1 dan, more strict at 3 dan and above.
We start 1 Dan players at 1700, 2 dan at 1900, 3 dan at 2100, 4 dan at 2300,
5 dan at 2500.  Even after hundreds of games, most experienced players
remain closer to the rank at which they play in Japan than to any other
rank.  If a player has no rating and no known rank on which to base an
initial rating, we have to guess based on his experience, then revise the
guess after the event based on his performance.
     Matt Casters says it is certainly not the case in Europe that an
average 4 dan can give rook & bishop to an average 1 dan.  No surprise here,
since Europe uses 100 point ranks while we use 200.  He says that the study
of handicap openings might give a player an inflated rating if he learns the
Joseki well.  To this I reply that handicaps are a traditional part of
shogi, that even the pros in Shoreikai play official Lance handicap games
for promotion, that the player will need to learn all the handicaps since he
does not choose his opponent or handicap in a tournament, and that there are
innumerable possibilities in shogi for the stronger player to confuse and
avoid the published Joseki.  He says that handicap play is just not fair.
We feel that if the opponents are rated far apart it is just a waste of time
to play an even game in a tournament, whereas a handicap game is exciting
for both sides.  We feel we must rate handicap games here, because we have
such a wide range of playing strengths in our tournament that pairings with
wide rating disparities are unavoidable.  We have seven players rated over
2300 (four Japanese, three American) in D.C. alone, all of whom generally
score over 50% against Japanese 4 dans, and none of our other players has
any real chance of winning an even game from any of these seven.  Even with
our generous handicaps, the 4-5 Dans usually win the handicap tourneys, but
not always--our last handicap event was won by the lowest ranked player, an
8 kyu (who happens to be my 14 year old son, an experienced chess player but
still a novice to shogi).


Larry Kaufman, 301-309-0904, 9213 Wooden Bridge Road, Potomac, Md. 20854