From: Larry Kaufman <lkaufman <at> COMCAST <dot> NET>
Date: 13 nov 2002
Subject: Re: byoyomi,  Bronstein, "Fischer"

----- Original Message -----
From: "David J Bush" <twixt <at> CSTONE <dot> NET>
To: <SHOGI-L <at> TECHUNIX <dot> TECHNION <dot> AC <dot> IL>
Sent: Wednesday, November 13, 2002 11:30 AM
Subject: byoyomi, Bronstein, "Fischer"


> |      The "solution" in my opinion is quite simple; replace byoyomi with
> | "time delay" (a.k.a. "Bronstein mode") which is available on most
digital
> | chess clocks.  This is actually much closer to what is used in
professional
> | shogi.  They don't do byoyomi as such; they simply don't count the first
59"
> | (or any other fractional minutes) of each move.
>
> I must plead ignorance here. Please, what is the difference between shogi
> byoyomi and not counting the fractional part of each minute? My only
> reference on this subject is the instruction manual to my Chronos clock,
> which describes shogi byoyomi (as opposed to byoyomi for go) as just that,
> rounding time up to the nearest minute (or whatever the byoyomi period is)
> after each move, throughout the game.
>
     Byoyomi in amateur shogi normally means playing with X minutes of
normal time, then Y seconds per move once time runs out.  But in Pro shogi
it works differently.  Thruout the game, seconds are just not counted, only
minutes, so it's like playing with only byoyomi from the start, but with a
huge number of violations (300 for a 5 hour game !) allowed.  Presumably the
shogi byoyomi mode described in your manual is the Pro method.

> But fo        amateur shogi games, the time control typically does not
round
> up your time until you reach the final byoyomi period, correct? So the
> Chronos "shogi" time controls would probably not be used, right? Chronos
> offers quite a few byoyomi related time controls other than "shogi."
>
     The actual method used in Pro shogi is not appropriate for use with
digital clocks.  It is perhaps left over from a time when clocks were
inaccurate, or was done to minimize recording work for the game recorder.
Time delay (a.k.a. Bronstein) is in the same spirit, but is much more
logical.  The Pro method works out to something like an average time delay
of 30" (half the 1 minute nominal figure) because in actual play, the player
would move at some random number of seconds past a multiple of one minute,
and so would save anywhere from 1" to 59" depending on when he moved.  It is
obviously more logical and fair to use a uniform delay rather than one that
is essentially random.  But once all the time except the last minute has
been used up, the two systems are the same as normal byoyomi.

> I also have a question about Bronstein timing. Suppose you have used up
> all but ten seconds of your main time. With an initial delay of 1 minute,
> when your turn comes the clock will show one minute counting down,
> correct? Then if you run the clock down to zero, the time will shift
> back to your final ten seconds? I think my Chronos has a mode that can
> do that, but it seems potentially confusing from a user's point of view.
>
> An increment time control, on the other hand, will show all the time
> you have remaining under all circumstances, which seems much simpler
> to deal with.
>
     This applies equally to increment and to delay.  Both are superior to
either amateur or pro byoyomi.

> | ... But playing pure increment shogi does not work well in
> | tournaments due to the unpredictable length of the game.  With delay
> | the initial time becomes important, and this method is probably more
> | suitable for actual tournaments as the game length is more predictable.
>
> With pure increment, the amount of time allotted would depend on
> the number of moves in the game, right? Perhaps a tournament director
> could make a "reasonable estimate" on the maximum number of moves the
> players are likely to make in each game, and schedule the rounds
> accordingly? Has this been tried?

     Of course this is done in games now, which use byoyomi, and it would
also be done with increment or delay events.  But the degree of
unpredictibility is much greater when more of the game is played with
increment or delay.  In a pure increment event, in order to accommodate the
occasional entering kings game, the time limit would have to be too fast.

With any of these modes, a game
> could theoretically last an indefinite period, although I certainly
> agree that a pure increment time would result in the greatest variation
> on a practical level.
>
      One solution would be to allow for a shortened byoyomi for entering
king games.

> By the way, Larry, thanks for discovering incremental time control!
>
     You're welcome, but I'm sure it was inevitable anyway once digital
clocks became common.  I just resent Fischer claiming the idea as his own.,
and everyone giving him credit, especially now that he has proven to be such
a jerk with his post- 9/11 comments praising the killers.

> David

    Larry