For a fifth consecutive year, it’s time to look back in order to look forward.
This is our annual Pythagorean wins prediction piece, which examines the teams who over- and underperformed in the previous season to predict who’ll bounce back or fall flat in the coming year.
It’s a stat with a strong track record – it tipped Brisbane and Geelong’s rises in 2019 and Hawthorn’s fall after 2016, among others – but just like the world, COVID has complicated matters.
Watch every match of the 2021 AAMI Community Series LIVE on Kayo. New to Kayo? Get your free trial now & start streaming instantly >
A shorter season with shorter games means a smaller sample size to analyse. It also throws out our ability to compare year-by-year results. Combine that with the hubs (and some teams being impacted by them more than others) and injuries having more of an impact on a shorter season, and it makes 2021 pre-season analysis more difficult.
But the basis of the stat is sound. Let’s recap how it works, how well it works, and what it’s tipping for 2021.
Get all the latest AFL news, highlights and analysis delivered straight to your inbox with Fox Sports Sportmail. Sign up now!
HOW IT WORKS
Pythagorean expectation is a formula developed by baseball stats guru Bill James. It estimates how many games a team ‘should have’ won based on its scoring. In baseball, this means runs scored and allowed.
This works analytically because looking at a team’s attacking and defensive performances overall, rather than purely whether they won or lost matches, gives us a larger sample size and more information.
Imagine two AFL clubs that each played four games against average opposition. The first team won three games by a point and lost the fourth by 100 points; the second game lost two games by a point and won the other two by 100 points.
Which team is better? The first team won one extra game but its percentage, and arguably its performance, is and was worse than the second team. We’d argue the second team was more impressive; Pythagorean expectation uses the same principle.
Using research by footy statistician Tony Corke, we can adapt the formula used to analyse baseball teams to study the AFL. To put it simply, we’re looking at which teams over and underperformed in the win/loss column based on their season-long performance.
HOW WELL IT WORKS
Year after year, Pythagorean expectation continues to predict risers and fallers successfully; in particular at the extremes.
Since 2010, there have been 17 teams that won or lost 2.5 more games than they ‘should have’. In the next season, 14 of those teams either rose or fell as correctly predicted by the stat.
We saw exactly what can happen in 2019 when Brisbane and Geelong both made leaps up the ladder, as successfully predicted by Pythagorean wins.
In 2018, the Lions won five games but had the percentage of an 8.5-win team; the Cats won 13 games but had the percentage of a 16.3 win team. Both teams underperformed.
Geelong ended up actually winning 16 games in 2019, while Brisbane joined them at that mark. The pair finished first and second, and while neither team made the Grand Final, they inarguably had successful seasons.
In terms of predicting collapses, the best recent example also comes from Geelong.
In 2013, the Cats won 17 games despite a percentage worthy of 13.7 wins. In short, they over-performed. The next season, Chris Scott’s side won just 11 games (with a draw) and missed the finals.
A SHORT REVIEW OF LAST YEAR
Like the rest of us, Pythagoras didn’t have a great 2020.
The stat’s strongest tip last year was for Sydney to rise. They didn’t, going from eight wins in 2019 down to five wins in 2020 (which is roughly like winning 6.5 games in a full-length season).
We’d put a caveat on this saying the Swans were incredibly unlucky with injuries, being down to just three healthy talls on their entire list at one stage.
The stat also predicted Richmond to fall after winning 16 games in 2019 off a percentage of a 13.6-win team. They won 12.5 games, which is roughly like winning 16 games in a full season.
So no decline there, though we’d point out the Tigers certainly weren’t as dominant in the home and away season as most expected.
SO WHAT’S GOING TO HAPPEN IN 2021?
For the stat to be truly be predictive, a team needs to have under or over-performed by a certain amount. In a normal year we use 1.5 games as the marker.
To analyse 2020, we’re going to use 1.1 games, because that cuts out roughly the proportion of the season that we lost.
Watch every match of the 2021 NAB AFLW Competition LIVE on Kayo. New to Kayo? Get your free trial now & start streaming instantly >
TEAMS THAT UNDERPERFORMED IN 2020 (Should rise in 2021)
Gold Coast Suns
Points for/against in 2020: 996/1099 (percentage of 90.6)
Actual wins: 5.5
Pythagorean wins: 6.9
The Suns’ 2020 season was similar to previous ones in a lot of ways. They had the now-traditional fast start – they were 3-1 in 2016, 2019 and 2020, and 2-2 in 2017 and 2018 – and then the fade-out.
But the fade-out wasn’t as bad last year. Perhaps staying home all season, with nine games at Metricon Stadium and another three at the Gabba, helped that. It wasn’t until Round 12, when the Suns looked visibly tired in the first quarter against Carlton, when the tempo of the season really seemed to get to them.
But before losing their last three games by a combined 118 points, the Suns were competitive, even without young gun Matt Rowell. They were in the top eight after Round 7, and still realistically in the finals mix until that Carlton loss.
Part of their record was bad luck in close games. The Suns went winless in games decided by a goal or less, with a five-point loss to the Bulldogs, a four-point loss to St Kilda and the draw against Essendon.
It doesn’t take a genius to predict a very young team will keep getting better, and perhaps break into the finals faster than expected, a la Brisbane of recent vintage. But this only adds to the growing Gold Coast bandwagon.
Points for/against in 2020: 1233/901 (percentage of 136.8)
Actual wins: 12
Pythagorean wins: 13.1
Chris Scott’s men are building a yearly tradition of winning the percentage premiership; they had the AFL’s best in 2020 and 2019, and the second-best in 2018 despite finishing eighth.
Usually most people explain that away by pointing to all of their home games at the Cattery, where they rarely lose and almost always demolish bad teams (like at the end of 2018, when they beat Fremantle by 133 and Gold Coast by 102 in consecutive weeks).
But you can’t blame GMHBA Stadium for ‘inflating’ their percentage this time. Geelong thumped teams no matter where they played; and good teams, too – 59 points over St Kilda, 60 points over Port Adelaide, then in the finals a 68-point thumping of Collingwood and 40-point beatdown of Brisbane.
The Cats were a properly good team who played like it; their Pythagorean expectation suggests they should’ve won even more than they did.
Oh, and then they added Jeremy Cameron, Shaun Higgins and Isaac Smith in the trade period, which seems like a good thing. Everything points to Geelong being better in 2021.
TEAMS THAT OVERPERFORMED IN 2020 (Should fall in 2021)
Points for/against in 2020: 1184/948 (percentage of 124.9)
Actual wins: 14
Pythagorean wins: 11.9
Our strongest prediction for 2021, using Pythagoras, is that Brisbane will decline.
That makes sense even before you look at the underlying stats. The Lions were fortunate to have the league migrate to their home state, handing them nine games at the Gabba and another four down the road at Metricon Stadium. They played just three games outside of Queensland, going 1-2.
Then you need to look at close games. Everyone knows one bad bounce can cost a team a game – or even a premiership – and so any game decided by a kick had some luck involved.
Brisbane was the only team to go unbeaten in games decided by a goal or less, with a one-point win over North Melbourne, a two-point win over St Kilda and a four-point win over Melbourne. In fact they went unbeaten in all games decided by two goals or less, also beating Collingwood by eight points and Fremantle by 12.
Throw in the fact Brisbane has been one of the AFL’s luckiest teams with injuries for two consecutive seasons, and all signs are pointing to a slide. They’re still a very strong team, and adding Joe Daniher could be a gamechanger, but history suggests some things won’t go their way in 2021 – and that could be the difference between another top-two finish and being stuck in an elimination final.
Points for/against in 2020: 938/1185 (percentage of 79.2)
Actual wins: 6.5
Pythagorean wins: 4.9
It was pretty clear by the end of last season that Essendon was playing at a bottom-four level, and the underlying numbers reflect that.
The Bombers’ six wins all came by 15 points or less, including three by a goal or less. They did lose a couple of close ones too, but they were overall quite fortunate to win as much as they did.
We’d suggest off-field issues were a factor here. There was “serious unrest” within Bomberland as they wrapped up the season, which contributed to Adam Saad’s surprise exit, along with confusion over the coaching handover from John Worsfold to Ben Rutten. It would not be a surprise if this impacted their poor on-field performance in the last month of the season, where they went 1-6.
The trade period, which saw the Bombers get younger, will help them in the long run but not in the immediate future. Most analysts are tipping Rutten’s side to slide in 2021 and the numbers support that.
WHAT ABOUT THE OTHER TEAMS?
The gaps between reality and expectation for the other 14 teams aren’t so notable as to be predictive.
Adelaide Crows: 3 actual wins, 2.61 Pythagorean wins
Carlton: 7 actual wins, 7.54 Pythagorean wins
Collingwood: 9.5 actual wins, 9.98 Pythagorean wins
Fremantle: 7 actual wins, 7.43 Pythagorean wins
GWS Giants: 8 actual wins, 7.76 Pythagorean wins
Hawthorn: 5 actual wins, 5.75 Pythagorean wins
Melbourne: 9 actual wins, 9.72 Pythagorean wins
North Melbourne: 3 actual wins, 3.59 Pythagorean wins
Port Adelaide: 14 actual wins, 13.06 Pythagorean wins
Richmond: 12.5 actual wins, 12.46 Pythagorean wins
St Kilda: 10 actual wins, 10.9 Pythagorean wins
Sydney Swans: 5 actual wins, 5.49 Pythagorean wins
West Coast Eagles: 12 actual wins, 11 Pythagorean wins
Western Bulldogs: 10 actual wins, 9.55 Pythagorean wins