The Forest from the Trees: A post for the stats nerds


#141

In my experience people who say this don’t understand stats, how to use them and this generally misinterprete what they are saying.


#142

Heres a stat that probably re-inforces what could be seen just by watching the game judging by comments on this forum. Yesterday Berger lost the hitouts 23 to 42. OK, games can still be won under this scenario, we have done it, many other sides do it. But Gawn won 23 to advantage, that is, half his hitouts were to Melbourne players. On the other hand Berger won just 5 to advantage.

Against Collingwood, TBell managed to break even Grundy in taps to advantage even if not around the ground. Grundy won 32 10 to advantage, TBell 20 with 8 to advantage,


#143

Melkshams 3rd quarter !
Gawn anihilated Berger in the 3rd quarter. Melbourne scored 6 goals. Coincidence?

Consistently above average over last 2 weeks Parish

Well above Average: Stringer, Zaha, ZMerret, Parish, Smith , Ambrose, Baguley
Well below average: Tippa, Hurley, Stewart, Goddard, McGrath,JMerret, Daniher

Consistently below average over last 2 weeks: Tippa, McGrath, Stewart, Daniher. Sore? Injured? need a rest?


#144

One positive, Ambrose demolished Hogan. I didn’t see that coming, but it’s amazing what happens when a defender keeps within 5 metres of their opponent


#145

melkshkam. lols.


#146

Yeah, it hurts when you spend so much time calling one of your players a spud, and when they move somewhere else they show you up. Deserved.


#147

Hawthorn game. aka Essendon vs. Sicily (in the only quarter that mattered.)

Players well above average: Hooker ( last quarter), Bellchambers, Stewart, McGrath, Myers, Ambrose, Mutch, Baguley
Average players: Smith Zaharakis Hurley, McKenna, Mc Niece
Players well below average: Goddard, Merret, Tippa, Danaher, Saad, Stringer, Laverde Heppel.

Pretty much confirms what you could see with your eyes, except Myers is in our top 6 players according to the stats assessment. Blitzers will be disappointed to realise he will be picked in the side again next week and probably Laverde will be dropped.


#148

Ha, the analysis says Stewart was markedly better than average… sure, he conned his man a few times on the wing to allow the 20m pass to him, but really?

(Didn’t see first half, but’s it’s ranked similarly to the half I did see.)


#149

What it shows is that of the 10 best players on the field 8 were Hawthorn.

Yes, its possible to pick holes in ratings systems like this but you can bet, when the coaches votes come out, Hawthorn players will get all the votes because the 3rd quarter was the only one that counted, except Tommy Bell will get 1 vote just because he played well across 4 quarters, something no other Essendon player did.


#150

Myers probably will survive (personally I think Woosh is punishing blitz for hyping him up the last ten years) but I don’t think the stats are that flattering. If you’re placed in a key ball winning position your numbers are always going to get a head start. Myers must be judged compared to other key onballer.


#151

Heppell went with Mitchell all day, had similar numbers yet influence was way below.


#152

Looking at the stats Myers had the most center clearances and most clearances for us. His disposal efficiency was only slightly greater than Heppells (68%). I mean if that is what he is brought into the side to do then he is doing his job. Though looking at the overall stats it’s not clearances that is the trouble for us (for a change).


#153

We’ve been the centre clearance kings this year, but simultaneously been smashed at stoppages around the ground. There’s a story there.

Would be interesting if they split the centre clearances into bounces and throw-ups.


#154

It was similar last year from memory.


#155

I have two theories and no way to verify either.

  1. Bellchambers has actually been top 5ish for hitouts to advantage this year but frequently gets worked over around the ground (Saturdays excellent performance not withstanding). I’d love to know the hitouts stats around the ground vs centre.

  2. The centre is less of a rolling maul than around the ground. I think this favours our smaller onballer more than around the ground.


#156

Brilliant and relevant article to our current predicament.

Tl:dr take aways

  • The loss of clearance specialists can redistribute work load for the better. I hoped this would happen post Watson, but we’re still playing a one dimensional clearance player in Myers. We shouldn’t.
  • Gun players matter. Nic Nat is currently West Coasts 2nd best player and 22nd best player (according to their own stat system PAV) in the league overall.
  • West Coast play two genuine ruckmen and they’re both killing it.

#157

This could be our own doing, we’re the team bringing more numbers to stoppages which also reduces the amount of players to kick to if we do win a stoppage.
I’m fascinated as to what would happen if wingers stayed on the outside and forwards stayed ahead.


#158

I cannot find a AFL statistical data source ANYWHERE. Obviously you can scrape the data from various places, but I wouldn’t mind a nice clean public facing API or database or whatever. Anyone have any idea? We have built a few statistical models for predicting games, predicting fantasy and brownlow, yet it would be much better to have an official data source before publishing them anywhere.


#159

Yeah we are way behind most major codes. Is the AFL behind afltables.com ?


#160

You can get it from champion data for the price of your first born and future earnings.