3 minute read

There's no sure way to pick a winner in a footy-tipping competition, but there is some tried-and-tested logic, an experienced professor says.

With the AFL and NRL back in full swing, a leading sports statistician has some advice for tipsters looking to lift their game this season - keep it simple.

For the first time in 39 years, Stephen Clarke will not operate Swinburne University's famed footy-tipping computer after benefactors pulled out.

He has run the program since 1981, consistently outperforming so-called AFL pundits.

The average AFL fan tips through a combination of head and heart, but the 75-year-old says two data points trump all - recent form and home ground advantage.

"The big thing is form," Prof Clarke told AAP.

"What happened more than eight or 10 weeks ago is irrelevant."

Under Prof Clarke's formula, interstate AFL clubs are up to four goals better off playing at home against some travelling sides.

"If you ignored everything else and just picked the home teams, you wouldn't be embarrassed," he said.

"I don't think you'd win a comp but you'd get 60 per cent correct."

While some footy fans routinely sweat over ins and outs as well as the impact of off-season recruits and departures, Prof Clarke suggests personnel changes are "overrated" in a game with up to 23 players per side.

It means tipsters may as well submit their picks on Monday, rather than waiting to consult team sheets later in the week.

"A lot of people overcompensate when the star player's out," he said.

"In the old days, most people would have thought 'if Tony Lockett's not in the team that would have a big bearing', and while I think it has some bearing it's not as important as people think."

The theory of one team simply having "the wood" over another is also spurious at best, with Prof Clarke saying maths doesn't back up the notion.

"When you've 18 teams and they play 17 other teams, the number of combinations is something like 180," he said.

"Now if the teams were all equal ... the team that won the previous 90 will win two in a row and 45 times they'll win three in a row and 20 times they'll win four in a row. So there will be 10 teams that win five in a row.

"It's pure randomness."

There is just no accounting for that randomness, Prof Clarke concedes, with even his supercomputer's predicted margins out by five or six goals on average.

Without a financial backer, the now-retired academic has folded the initiative - much to the sadness of loyal followers who have for years relied on his tips to one-up family, friends and work colleagues.

"We've had a pretty good run," Prof Clarke said.