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


#501

Look, to be fair - I’ve taken a bit of usual hyperbole and exaggeration with my comments. I think stats have a very important role in sport. But my opinion is that the role they have is twofold - raw data to aid coaches, medicos and players in bettering their craft, and molded data to suit click-bait. I have an abnormally high disdain for the latter.

On a related point to that - the age-old conundrum of the single gun forward vs the spray of lesser lights hitting the scoreboard. They have similarities in the way the play is focussed and directed.


#502

While I agree with the idea that this particular stat is a bit rubbish (the sample size is so small etc), I disagree with the rest of the premise. The raw data is no use of you can’t use it to draw conclusions. But I think it’s possible to draw conclusions (but not necessarily infer causality) in even very complex systems. Presenting that in a way that makes it digestible is the challenge, and one that gets people all wound up about CD over.


#503

Ummmm. Pretty sure there’s sh*t loads of raw data processing that goes on before any coaches or medicos get near it.


#504

Really? We’re in love with putting a forward at the back of the square at centre bounces.


#505

So DJR, obviously they do have Internet at Everest Base camp.


#506

Of course. To cater for all the rich buggers & influencers ticking it off their bucket list. ( Still waiting for the ‘DJR in Essendon jumper’ Instagram posing in front of Mt Everest, & the subsequent deluge of Blitz metaphors and puns cross referencing Premierships)

On DJR’s point re Walla & Fanta off the back of the square this was a late in the quarter speciality to take advantage of tired legs.
Interesting to see if we try a similar tactic using them as wingers for late centre bounces this year.


#507

Mate, I went higher than Everest when I took Mark of the Century™ at Glenferrie. Took three days for the ice in my beard to thaw.

But here’s one from earlier, go nuts.


#508

It looks a tad…ahm…how can I say this…compensatory…


#509

I think the grass is making the Qutab Minar look taller.


#510

It used to be bigger…


#511

As did most of us.


#512

37 observations should be enough for a relatively small margin of error… say +/- one standard deviation. However we did have a relatively soft draw towards the back of the season when that quartet would have done the bulk of their work.


#513

The bloke that wrote the piece in the sun said it was surprising that the essendon group was best. Maybe he made one of these mistakes.
A. He looked up the CD player ratings of our 4 players and saw they are just average or above average, with not one elite among them and he believed those ratings.
Or B. He believed the table that shows that group of Essendon mids are best.


#514

Let’s put this centre bounce stuff in numbers.

Our best combo resulted in a score 7 times out of the 550 centre bounces Essendon competed at.
Or, another way to look at it -
The best combo in the AFL resulted in a score 7 times out of 550 centre bounces.

Not particularly impressive.

25 centre bounces a game
22 rounds
We scored 7 times
Once every 3 games did this combo result in any score. Probably a behind*

Mind blowing stuff. This could revolutionise the way we play.


#515

Spot on.

The 6 6 6 rule is going to have little effect.


#516

Are you using a binomial distribution? Something else? I’d have thought you’d want to be coming up with a different number to be doing it right for this sort of case??


#517

Yeah, but how many games did we win when wearing red?


#518

Well they weren’t.

Gawn and Brayyshaw were involved in 80 events on those limited numbers you showed, vs 37 for EFC’s trio. (Of course I can’t see the lower figures, but those numbers you provided show why MEL is doing so well).


#519

Yes Albert, I fully realise Melbourne have 2 different groups of mids coming in the first 10 combos, meaning they have better “depth”, and we cannot see the other combos in rankings below 10th

However, my post omitted the “tongue in cheek font”


#520

Aah, the tongue-in-cheek font — a much rarer, though better appreciated, commodity in these parts than sarcastica.