17 March 2021
The Probabilistic AFL Football Tipping competition dates back to 1995 and allows tipsters to score points based on the probability given for a team to win the desired match - the greater the risk, there's an increasing reward but a rapidly increasing penalty if you're wrong.
The competition, designed by Associate Professor David Dowe from the Faculty of Information Technology at Monash University, replicates real-life scenarios when an outcome isn't 100 per cent guaranteed.
"This competition gives each tipper the opportunity to express their uncertainty or confidence level in the outcome, and have a bit of fun while doing so. The scoring system is designed to reward tipsters who are good at estimating probabilities.
“In some instances, people don't give probabilities that reflect their true beliefs. We see this occur when a tipster thinks their team will lose but they instead choose a probability just over 0.5, like 0.55, as a way of showing team loyalty but not risking losing much.
“As we saw in Round 1 of the 2019 Tipping Competition, people often get burnt by giving probabilities that are overly confident and far too close to 1 and far away from 0.5. There can sometimes be success with such strategies in the short term, but they tend to come unstuck before too long.
“In contrast, in season 2018 of the competition the leader after Round 1 was on 5.251 and there were very few entrants with negative scores at that stage. But, after Round 2, there was only one person with a total positive score and the leader after Round 1 had dropped to an overall negative score of -0.68.
“In any event, one of the fun things about this competition is how people might agree about which team they think will win but they might ascribe vastly different probabilities.
“There is also a Gaussian version of the competition which involves the tipper nominating a winning margin and standard deviation. The standard deviation allows the tipper to express their confidence in the choice of margin.
"In this competition, the accuracy of predicting the winning margin is important, but having a standard deviation that is too small can potentially cause major setbacks. Experience has shown a standard deviation of around 40 seems to be a good place to start," said Associate Professor David Dowe.
The Probabilistic AFL Football Tipping competition is free to enter. To register, please visit here.