Good Game or Bad Game: Riley Bonner vs Essendon

[Go_Dogs]

In some ways it depends on conditions too. A stat line like that in the sopping wet would be quite good I imagine. But under Marvel, maybe less good.

Ultimately double digit clangers kills the good for me. That’s like 4-5x an acceptable level. Meters gained is like 2x a high archived level.

[jeemak]

Neither good nor bad.

Territory is important and you can only get the footy to whoever is presenting and I'd much rather a long/ up the ground turnover than a short one in the pressure zone.

If you're crowding your own defensive half then options further afield are going to be a bit limited. If team mates aren't running for you or presenting wide then you're going to turn it over long.

A mixed result is OK, good or bad is a bit too linear for mine. Stop being linear MJP!

[Hotdog60]

Metres gain on how they present it is poorly defined in that it should be the player would runs and takes the game on without kicking it.
Metres gained from dodging and waving through traffic and the old one two would be a more affective metres gained.
A player who can roost a ball would have more of an advantage over distance. Lukosius could out gain Caleb Daniel even if Caleb ran more.

But in saying all that its no good gaining a 1000 metres if you gift the oppo with 5 or 6 goals from turnover. So the question needs to asked how bad was the clangers and how many of them resulted in goals. That will tell you if its a bad game or not.

[Mofra]

I didn't see the game. I have one question.

Was it wet, or dry?
It's perhaps way off, but I feel like metres gained on a wet day matter far more than a dry day.

[Scorlibo]

AHA. So...player ratings are important?

Do player ratings include metres gained?

If you could only look at one stat to assess a player's performance, it 100% should be player ratings.

Metres gained is a component of player ratings, but not as part of a weighted compilation (as you'd see in Supercoach points).

Player ratings assess the change in game situation, specifically the change in probability of the next score.

Example >

The starting situation is a centre bounce. The ball is in the middle of the ground and not in either team's possession. Therefore the probability of either team scoring is 50%, and the probability of either team scoring a goal is something like 30%.

Possible outcomes are:

Team A scores a goal - 30% chance - 6 points - 0.3*6 = 1.8
Team A scores a behind - 20% chance - 1 point - 0.2*1 = 0.2
Team B scores a goal - 30% chance - 6 points - 0.3*6 = 1.8
Team B scores a behind - 20% chance - 1 point - 0.2*1 = 0.2

Average of all outcomes for each team is 0.

Now let's say that Team A's ruckman manages to grab the ball out of the ruck, brush a tackle and handball to their wingman who has run forward of centre and is in clear space. The ball is now in Team A's possession, not under any pressure (also measured), and forward of centre. Let's say that Team A's chance of scoring next is now 75%. Possible outcomes are:

Team A scores a goal - 45% chance - 6 points - 0.45*6 = 2.7
Team A scores a behind - 30% chance - 1 point - 0.3*1 = 0.3
Team B scores a goal - 15% chance - 6 points - 0.15*6 = 0.9
Team B scores a behind - 10% chance - 1 point - 0.1*1 = 1

Average of all outcomes is now that Team A scores 2 points. (2.7 + 0.3 - 0.9 - 0.1). So the ruckman is credited with 2 player ratings points for changing the game situation in his team's favour.

Tied up in the ruckman's play is a centre clearance, a contested possession, a broken tackle, an effective handball and maybe 20 metres gained. The player rating is not the sum of these parts but rather the scoreboard impact of the player's involvement.


Coming back to Bonner's game. I'd guess that the probability of the Saints scoring a goal when Bonner has uncontested possession in defensive fifty is going to be pretty similar to the probability of them scoring a goal from a contest on the wing. So if he's bombing long to a contest, that's not improving his team's chance of scoring next. His stat sheet reads like he was doing this all game, with a few clangers thrown in. It would be different if he was winning his own footy but 21 uncontested possessions says that wasn't the case. The player ratings clarify all those factors into one helpful number, which if you haven't watched the game is the best starting point (imo).

[jeemak]

You're dead now MJP. I took on Scorlibo on this topic twelve years ago and still have the burn marks to prove it.

[Topdog]

For me I'd have to dive into how many of those "clangers" were kicks to packs as opposed to just piss poor turnovers but I feel like nothing I look at can turn this into a good game. As hujsh said so many goals come from turnovers now so giving it away 19 times in a game is criminal.

[Topdog]

Bad game.

Would be interesting to see a metres gained with possession retained stat, because that's what matters.

He registered a player rating of 4.4 which was the 5th lowest for the Saints.

Which site gives this player rating? (Asking this after reading your detailed overview got me interested)

[Bulldog Joe]

OK. You know I 100% agree with all of what you wrote and this in particular...

But I would love you to put all this sensible stuff that muddies the waters aside for a second.

19 Clangers.
1000+ metres gained.

Same player.
Same game.

Good or Bad??
Gut feel.

Would it help if I told you the 2x main stats for determining the winner of a game in 2024 (so far) have been:

1/. Number of EFFECTIVE kicks.
2/. Total metres gained.

It's a very interesting stat line and one I have never seen before.

I am saying bad game.

On the metres gained to the deduct what is cost by the clangers?

1000 metres gained is huge but not if it has cost something similar in rebound.

[Scorlibo]

Which site gives this player rating? (Asking this after reading your detailed overview got me interested)

You can access it via the afl.com.au match day under 'player stats', but you have to go to the 'my stats' tab and setup your preferred stats. There's a bunch of stats only accessible this way including ground ball gets, pressure acts, hitouts to advantage.

Alternatively a hobbyist footy statistician named Andrew Whelan runs this third party site which is nicely formatted and great for historical ratings: https://www.wheeloratings.com/

[mjp]

The player ratings clarify all those factors into one helpful number, which if you haven't watched the game is the best starting point (imo).

Player ratings are a solid starting point, we can 100% agree on that.

[Topdog]

I still think AFL stats overall are quite immature and at times struggle to believe/understand some of them.

This Bonner game is a good example. 32 disposals, 21 of them effective disposals yet 17 turnovers with 6 clangers.

[Scorlibo]

I still think AFL stats overall are quite immature and at times struggle to believe/understand some of them.

This Bonner game is a good example. 32 disposals, 21 of them effective disposals yet 17 turnovers with 6 clangers.

There's definitely something awry with the public relations around AFL stats. It being such a complex game (compared to cricket for instance) makes it tough.

One oddity that is probably not widely known is that long kicks to a contest are counted as effective kicks. Bonner's stats you've listed above only make sense if you happen to know this.

32 disposals

10 either retained possession or were long kicks to a contest where the oppo didn't win the ball back straight away (effective disposals)
5 short kicks or handballs to a contest (ineffective disposals)
11 long kicks to a contest where the oppo won the ball back straight away (effective disposals, turnovers)
6 pure turnovers (clangers, turnovers, ineffective disposals)

[Topdog]

Brilliant, thanks Scorlibo!

[Mofra]

Bonner is a defender?
I still think 32 touches in the wet would be a better game, but if he represents 32 times the opposition forwards don't have the ball, well, he's done his primary job. I'm fairly certain he's not their 'escape plan' from opposition congestion.