[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [pbmserv-dev] tournament scripts
'Twas 09:49 AM -0700 on Sunday 9/7/2003 when Lyman Hurd wrote:
I am personally of the opinion that Cathedral is less
asymmetric than folk wisdom would have one believe.
I think it's a difference of about two squares. But more important, among
great players I think it's something like 80% victory for the second player.
I
would be interested in seeing on the server what the
point spread is between black and white players
(although in some cases one would have to interpolate
the final result).
We wouldn't have to interpolate the blowouts, because the interesting date
point is the median not the average. Yet analyzing the server history
might be misleading. Maybe stronger players tend to go first to minimize
their advantage. Restricting that analysis to games which are begun in
pairs would clear up that.
Making that change might affect the way people play,
since often the usual strategy for the endgame is to
guarantee that one of ones opponent's pieces cannot be
placed and then play conservatively to make sure that
one wins if only by a point or two.
Which is another problem in analyzing the history. It would change the way
it's played, and that's a good thing. According to Moore, the game should
be played in pairs, and that's the way it is played in most tournaments
(not that there have been a lot).
I am not convinced that Cathedral is not a draw with
perfect play!
I am.
I was anticipating setting things up as we did the
Zertz+11 tournament in which everyone plays two games
per opponent once per side. In this case one owuld
have to count wins as 1 and draws as 0.5.
I'll play in such a tournament, but I think point count in paired games is
better than simple won-lost.
PS The current scripts identify tournament games as
(say) a12 if I (the first player) am player 1 in pool
a and you (the second player) are player 2. Thus in
the situation described above we would be playing
games a12 and a21. This worked quite well (keeping
the pool size manageable in that case 4).
That worked just fine in the Quoridor tournament. Some games are played
for simple win or lose, some are played for point count. Since the first
time I played it, I have believed that Cathedral is better when played for
point count. Moore agrees, just look at the Cathedral page.
If the first round of your tournament has everyone tied at 3-3, don't say I
didn't warn you.
--
RB |\ Randall Bart
aa |/ admin@RandallBart.com Home: 1-212-965-1786 Barticus@att.net
nr |\ Cell: 1-917-715-0831 I LOVE YOU Work: 1-212-343-1234 x3016
dt ||\ http://RandallBart.com/ Ånåheim Ångels 2002 World Chåmps!
a |/ Multiple sclerosis: http://www.cbc.ca/webone/alison/
l |\ DOT-HS-808-065 The Church Of The Unauthorized Truth:
l |/ MS^7=6/28/107 http://yg.cotut.com mailto:s@cotut.com