ESPN

Footy Forensics: Who plays in September?

Has there ever been a more unpredictable season of AFL? Don't panic, the Elo-ratings are back to make sense of it all and predict which sides are finals bound.

Carlton beat GWS, North Melbourne beat Adelaide and the Brisbane Lions managed to down Fremantle by nearly 10 goals. Nothing in the AFL world is making much sense at the moment. But don't despair, Elo ratings are back and will attempt to make some sense of this madness.

With Elo ratings, we can work out which teams are genuinely in with a finals chance, setting aside the week-to-week freak results that litter the fixture. After each game, a team's Elo rating rises or falls depending on whether they performed better or worse than expected. Win narrowly at home against a team you're expected to beat comfortably and your club's rating will fall; post a big win on the road against strong opposition and the rating will rise.

After Round 12, Adelaide sit on top of both the AFL ladder and the Elo ratings ladder. They're followed by Geelong and the GWS Giants. Rounding out the top four is (perhaps surprisingly) Sydney. The Swans started the season as the highest rated team, but plummeted down the table after six poor weeks. But they've clawed their way solidly back into positive territory since and you'd be a fool to write them off.

Using the Elo ratings, I simulate the remainder of the season 20,000 times to calculate each team's chance of making the finals. The Lions, Blues, Dockers and Hawks all have less than a ten percent chance of ending up in the top eight, but every other team still has a double-digit shot.

In keeping with the even, topsy-turvy nature of the season so far, there's still a fair bit of uncertainty about how the ladder will end up. About the only thing we can say for sure is that the Lions won't make the eight, the Hawks won't make the top four, and the Crows won't finish near the bottom. Other than that, there are still a range of possibilities for this home-and-away season, as you'd expect from a league in which only two games separate 5th from 17th on the ladder. There are seven teams that have a better than 20% shot at making the top four, including the recently-mediocre Bulldogs and Eagles.

BASED ON ELO RATINGS AND 20 000 SIMULATIONS OF THE 2017 SEASON.

Each team's season so far can be pretty handily summed up by looking at how their estimated chance of making the finals has evolved. The Crows started off fancied, with their 76 percent pre-season finals shot rising to 99 percent after Round 12. The Lions' chances started remote and are now non-existent, while the Blues' victory against GWS hasn't been enough to give them more than a one-in-fifty shot at the finals.

Richmond's season has been a rollercoaster - the Elo-based simulations thought they were 28 percent likely to make the finals before the season started, a figure which then shot up to 80 percent with their run of early form before plummeting back down to 50, and then back up to 80. Very Richmond.

BASED ON ELO RATINGS AND 20 000 SIMULATIONS OF THE 2017 SEASON.

Not convinced by all this Elo stuff? Well, the finals chances from our Elo-based simulations are pretty close to the probabilities implied by the betting markets. If you think Elo has it wrong, you're also quibbling with punters.

Along with the Elo simulations and betting markets, a third set of probability estimates comes from Champion Data, which has calculated each team's chance of making the finals by using its Player Ratings to simulate the remainder of the season. Again, the results are mostly pretty similar to the Elo and betting market figures, except in the case of the Bombers.

ESTIMATED FINALS CHANCES USING AFL PLAYER RATINGS CALCULATED BY CHAMPION DATA. UBET ODDS CONVERTED TO IMPLIED PROBABILITIES, REMOVING OVERROUND.

So, now you're suitably convinced that the Elo-based estimates of each team's shot at September footy are credible, let's have a look at how each team's chances will evolve in the coming weeks depending on whether they win or lose.


.

Of course, the question for the Crows isn't really whether they'll make the finals as Elo gives them a one-in-100 chance of not being in the top eight. Even if they lose to both the Hawks and the Blues, Adelaide will still be more than 90% likely to play in September.

Return to top


.

The Lions' finals chances also won't be affected much by what happens in the coming weeks, but for the opposite reason than the Crows. Even a highly improbable string of victories against the Power, Giants and Bombers will still leave Brisbane's shot at September at close to zero.

Return to top


The Blues are very unlikely to book a berth in September. Still, it's not completely outside the realm of possibility (as it is for the Lions). If Carlton manages to knock off the Suns, the Tigers and the dominant Crows, they'll boost their finals chances from one-in-fifty to about one-in-five.

Return to top


The next few games after the bye will be very important for Collingwood. Will they stay in touch with that big pack of teams in the middle of the ladder, chasing a bottom-of-the-eight spot? Or will they fall well off the pace? Win against Port and the Hawks and they'll nearly have a fifty-fifty finals shot; drop their next two and it's likely to be all over.

Return to top


After a hot-and-cold start to the year, the Bombers have come roaring back into form in recent weeks. The Elo model still doesn't quite give them an even-money shot at September, putting them at 40 percent after Round 12. However, a win against the Swans would send them well above the 50 percent mark.

Return to top


The Dockers have spent some time in the top eight this season, but they're quite unlikely to be there at the end of the year. Even if they can beat Geelong, they'll still have less than a one-in-five chance of making the finals.

Return to top


For the Cats, the main game now is stitching up a top four spot - the eight isn't really in doubt. Even if they lose their next three games, they'll still be odds-on to make the eight; it's unlikely they'll dip below a 90% shot from here.

Return to top


The Suns have strung together a couple of wins in a row, leaving them with a nearly one-in-five chance of snatching a spot in the finals. A win against the lowly Blues won't do much to improve those odds, but their games against St Kilda and North Melbourne give them a chance to knock off fellow mid-table aspirants. Win all three and the Suns will have a fifty-fifty shot at September.

Return to top


The Giants haven't quite been the all-conquering unstoppable force that some foresaw for this year, but they're still an extremely good football team that is very unlikely to miss the finals. If they lose to the Lions and the Cats they'll still have a five-in-six chance.

Return to top


.

2017 will be the first season this decade in which the Hawks don't make the finals. They still have a mathematical shot (1.8 percent) but it would be a little crazy to cling to that. If they lose to Adelaide and Collingwood, that chance will be well below 1 percent.

Return to top


Melbourne's finals chances are the closest thing to a toss-up in the league at the moment. After Round 12, the Demons have a 54% chance of playing in September. The next few weeks will really shape the Demons season, as they're facing three beatable teams with their own finals aspirations. If they win two of the three, the Demons will have a better than 75 percent chance of playing finals footy for the first time since 2006.

Return to top


After aggressively flipping the switch to 'rebuild' at the end of 2016, few expected the Kangaroos to be contenders this year. They've been far from woeful though, so their shot at the finals is still in the double digits. A loss to St Kilda would push their finals chance below the 10% mark.

Return to top


At times, Port Adelaide looks like a true Premiership threat; at others, they lose by more than ten goals to Essendon. The Elo-based simulations currently give the Power a 78% chance of making the finals. Wins against the Lions and Magpies would send that figure to around 90%. Port are very unlikely to drop out of the eight.

Return to top


The Tigers' season so far has been a rollercoaster, but they're still in a very good position after Round 12. They have an 80% chance of finishing up in the top eight from here; if they win two of their next three games their chance will be over 90%.

Return to top


The Saints had real finals hopes this year, but they're fading fast. After Round 12, our Elo-based simulations give them a 20% shot. Still, they have three winnable games coming up - victories in all of them would leave the Saints with a fifty-fifty shot at playing September footy.

Return to top


The Swans have possibly been the most perplexing team of 2017. Very few would've tipped them to have just four wins at this point of the year. Still, the Elo ratings continue to be impressed by the Swans, and the simulations give them a better than one-in-three chance of clawing their way into the finals from here. The three games against likely top eight teams coming up could define the Swans season; they could end this three-game run with anywhere from a 2% to 77% chance of playing in September.

Return to top


Indifferent form and the quirks of the bye weeks have pushed the Eagles out of the top eight, but they're still quite likely to claw their way back for September. A spot in the bottom half of the top eight looks likely for West Coast, which would be a good result for many teams but perhaps not for one with their ageing list and closing window.

Return to top


The reigning premiers haven't looked very regal of late, but they still have a three-in-four chance of making the finals. If they beat the Demons, Kangaroos and Eagles, their chance will soar to 93%; drop all three upcoming games and September will look quite unlikely.

Return to top


Every month, we'll keep you updated about how your team's rating -- and chance of making the finals -- evolves.

Matt CowgillMatt Cowgill is an economist by day and footy stats obsessive by night who thinks there's no such thing as too many graphs. He uses data analysis to tell stories for ESPN's Footy Forensics, as well as his own blog The Arc.

More Stories