Monday, May 28, 2018

AFL Power Rankings: Round 10 2018

Melbourne has won its past five matches by an average of almost 70 points. Even if the Demons have played more ‘below average’ teams than ‘good’ teams over that stretch, that is amazing.

This week I originally thought of highlighting the GWS Giants and how injuries by themselves can’t explain why, over the past four weeks, the Giants have been playing at the level of Carlton and Gold Coast.
But I can’t ignore the rankings’ favourite team over the past month – the rapidly rising Melbourne Football Club.
The Demons have gained a whopping four goals worth of ranking points since their Round 5 loss to Richmond. Put another way, the rankings think the Demons are a four goal per match better side than it rated them just five weeks ago.
They’ve shot up from tenth to second in the rankings. Their average net margin, over the past five weeks – even after adjusting for estimated opponent strength and home ground advantage – is +61 points. Over the past three weeks it is +81 points. Melbourne has destroyed sides. Sure, two of their big wins were against Carlton and Gold Coast, but they beat them by far more than you would expect them to (see chart below). And their latest win was a 91 point thrashing of last year’s runners-up the Adelaide Crows, who are still a decent side even with their injuries. 
Melbourne has the highest inside 50 ratio of any team in 2018, gaining 30 per cent more inside 50s than their opposition. That is up there with the very best midfields since 2000. Over the past five weeks the Demons have had 50 per cent more inside 50s than their opponents. Superstars Max Gawn in the ruck and Clayton Oliver are the backbone upon which the midfield is built, ranking second and ninth in SuperCoach ranking averages over the season to date.
But it’s their forward line that has really taken off over the past five weeks – particularly the quartet of Jake Melksham, Alex Neal-Bullen, Jesse Hogan, and Tom McDonald. During that period Melbourne has averaged 20 goals per game, in a season where teams have only been averaging 12 per game. They took over 18 marks inside 50 metres in every one of those weeks, compared to a league average of 11 per game.
Those forward line statistics are not just due to more entries into attack. Over the past five weeks Melbourne’s scoring once inside 50 has also been about 30 per cent better than the league’s season average. According to the AFL Player Ratings, Melksham had the highest-rated performance for six years against Carlton. For a forward line that would have generally been rated as no better than average before this run few would have seen this coming.
Sceptics may still point to the fact that once you get a bad team down by about 50 or 60 points the extra few goals don’t matter. Some AFL rating systems, unlike mine, do put less weight on extra points once you get past a certain margin. If you chopped off the maximum net margin at, say, 50 points Melbourne would rank about a goal per game lower.
That still doesn’t overturn the fact that Melbourne has vastly improved in the past month to an extent not often seen. Beating up sides for weeks on end – good or bad – is exactly what many premiership contenders do.


Act II Ministries said...

You're not the only one.

The Demons were 11th on our ELO-Following Football ratings after round five (45.3, where 50.0 is the mean). In the last five weeks, consecutively, they've gone up about four points, three points, six points, 10 and 10 points. Today, after round 10, they sit at 79.3, six points above Richmond for first (the Tigers have fallen to 73.6; West Coast is up to 72.2 in third).

In this system, rating points translate directly to predicted margin (discounting HFA, short weeks, etc). So Melbourne is theoretically a 34-point better team than they were in April. And if it weren't for the fact that we have a circuit-breaker in place when a blowout gets too big, they'd be farther ahead. (We've only had to use that circuit-breaker thrice this year - Melbourne's last two games are two of the occasions!)

Troy Wheatley said...

I suspect they’ve shot up many people’s ranking systems!

Apart from ‘circuit breakers’ for blowouts some main things that could slow them down on a rankings system are scoring shots (they have been very accurate the past fortnight) and a relatively slow adaptation to recent performances.