Dear All,

The Monash University School of Computer Science and Software Engineering's
probabilistic football prediction competition dates back to 1995 and is free
for all to enter, with entrants entering not just which team they think will
win a given Australian Football League (AFL) game, but also a probability
that they assign to that event.

This competition received some quite a lot of
http://www.csse.monash.edu.au/~footy/media.shtml
newspaper, radio and television media coverage.
This WWW-based probabilistic footy-tipping competition is at
www.csse.monash.edu.au/~footy .   It is free for all to enter, and there is
at least Aus$512 in prizes to be shared amongst the top 10 Australian
secondary student tipsters in the probabilistic competition.

The scoring system for the probabilistic competition is
1 + log_2(p)     if you're right
1 + log_2(1-p)   if you're wrong.

So, if you just go 50% all year, then you would end up with a score of
0.00 for each game and a total of 0.  The optimal (expected) long-term
strategy is to tip the true probability of the outcome of the game.
If you're inexperienced at probabilistic prediction, then it's probably
a very good idea to practice by tipping conservatively in the first few rounds.

People who join the competition late join with a score of 0.00, as though
they had tipped 50%-50% on every game up to that point.  With usually more
than half the field having a negative score, this is not such a great
disadvantage.

More details on the scoring system for the probabilistic competition and
information is given at
http://www.cs.monash.edu.au/~footy/about.shtml .
Details on two other competitions (Normal and Gaussian) are also given there.

Tips can be entered in advance.  As long as the AFL has supplied the
schedule for the game, we will put a template up on the WWW so that you
can enter your probabilistic tips in advance.  The cut-off for entering
tips is approximately 1/2 an hour before the scheduled bounce of the ball -
but see the WWW page for any clarification.

Teachers are not only encouraged to encourage students to enter, teachers
are encouraged to enter themselves.  For less advanced students, the
WWW site permits graphical plots of the progress of various tipsters.
For more advanced students - of calculus, it is an interesting result to
derive that, given our logarithmic scoring system, if one knew the _true_
probability, then the optimal long-term strategy would be to tip with the
_true_ probability.

If you want to practise your tips, two useful things to do are to look at
the scoring system: http://www.csse.monash.edu.au/~footy/about.shtml#info
and to monitor the progress of tippers from previous years:
http://www.csse.monash.edu.au/~footy/past/2001
and current tippers.

Please, all, feel free to enter, particularly primary and secondary students.
Where you deem it appropriate, please also feel at liberty to tell keen friends.


Yours faithfully.              - David Dowe.

Dr David Dowe, Competition organiser, School of Computer Science and
Software Eng., Monash University, Clayton, Victoria 3168, Australia
Fax:+61 3 9905-5146  http://www.csse.monash.edu.au/~dld/
http://www.csse.monash.edu.au/~footy/

-------------------------------------------------------------------------

Reply via email to