From: Larry Kaufman <lkaufman <at> WIZARD <dot> NET>
Date: 1 sep 2000
Subject: Re: MSO 2000 Shogi Festival

     Some comments on the MOS 2000 Shogi Festival:

     It was exceptionally well organized and run, and the format worked
quite well in most respects.  I feel, as I think most of the participants
do, that it demonstrated that the Swiss system is eminently suited for
shogi.  It is true that it did not produce a last round deciding game
situation; the key game between the top two finishers (in the Open)
occurred much earlier.  However the players were still treated to a live
professional analysis of a demonstration game (in this case between myself
and Pro Nakao 6 Dan).
     The handicap system proved its worth amazingly well.  In the team
event, 100 handicap (counting sente as a handicap in a few games) games
were played.  The rule was that the rating difference was rounded down to
the next lower handicap value, and that handicap was used.  Based on this
rule, I predicted before the event that the score should be around 60-40
for the handicap givers (as was the intent of the organizers), and the
final score was 59-41!  The handicap simuls used a similar rule but with a
40 point bias added for the handicap givers along with an estimated value
for the simul, and so the handicap givers should have scored somewhat more
than the 59% in the team event.  In fact they scored 62%, close to
expectations.  All in all the handicap system worked amazingly well, though
it seems we may want to make some minor adjustments to some of the
individual handicap values based on the results.
     How might things be improved in the future?  One suggestion I heard,
which I like very much, was to use a McMahon Swiss.  This avoids wasting
the early rounds (I believe there were no upsets at all in the first round
of the Open, though there were several huge ones later on) on huge mismatches.
     As for time limit, my previous experience had been almost completely
with events in which the byo-yomi (in seconds) was not less than the main
time (in minutes).  In such games, it is normal to end up in byoyomi, as it
is no faster than the main time spread over 60 moves.  In this event, the
main time was one hour, while the byoyomi was 30".  As is now clear to me,
with such a time limit it is foolish to enter byoyomi at all, as I did in
nearly every game (much to my regret).  But if the players avoid byoyomi,
then it doesn't have much point.  So my recommendation would be to use 40'
+ 40" or 50' + 50" (whichever fits the schedule better).
     Actually much better than byoyomi is time delay, where the clock
doesn't move for a set number of seconds each move.  This is really more
like the method used in pro play, in which only whole minutes count;
seconds are disregarded.  Almost all digital chess clocks on the market now
support time delay.  However the Japanese byoyomi clocks do not, so this
suggestion is probably not practical as long as most of the clocks
available to shogi players are Japanese.
     Anyway, it was a really great event, and we all look forward to next
year.

     Larry Kaufman, amateur 5 Dan