The beginnings?
Date: Tue, 28 May 2002 00:14:19 -0700
Message-ID: <acvaof$dod$1@hplms2.hpl.hp.com>

Quoting Marcelo Weinberger in rec.sport.soccer:

Most WC predictions in this group are based on clear-cut, deterministic >specification of which teams will advance in each group, with no indication >of how close the outcome will be. It seems much more reasonable to assign to >each team in each group a probability of advancing, a number between 0 and
1. The sum of the probabilities of the four teams in the group should of >course be 2. The best predictor is the one for which the product of the >probabilities assigned to the teams that actually advance is maximum.

So, if you know for sure the 16 advancers and assign a probability of 1 to >each one of them, your score will be 1, best possible. The worst case is
when you gave a zero chance to some advancer, and your score is zero. (There >is a catch here, because a zero score should also be given when you assign a >probability of 1 to a team that doesn't advance; this problem could be
solved with a different system, but let's keep it simple.) No knowlwdge
means that you assign 0.5 to everyone, and your score is 1/64536, regardless >of who advances.

In this probabilistic game, I'm willing to make a prediction, which I never >do when you are asked to anticipate a clear-cut winner. Here it is:

A: France 0.95, Denmark 0.40, Uruguay 0.35, Senegal 0.30
B: Spain 0.70, Paraguay 0.65, South Africa 0.40, Slovenia 0.25
C: Brasil 0.99, Turkey 0.42, Costa Rica 0.39, China 0.20
D: Portugal 0.80, South Korea 0.55, USA 0.35, Poland 0.30
E: Germany 0.75, Cameroon 0.55, Ireland 0.45, Saudi Arabia 0.25
F: Argentina 0.63, England 0.53, Nigeria 0.46, Sweden 0.38
G: Italy 0.97, Croatia 0.40, Mexico 0.37, Ecuador 0.26
H: Japan 0.60, Russia 0.58, Belgium 0.57, Tunisia 0.25

-- Marcelo

--
Manchester United FC - CHAMPIONS
Latest: England '13 (20th) Europa '17, UEFA '08, World '08
--- SBBSecho 3.06-Win32
* Origin: SportNet Gateway Site (24:150/2)