Statistical Analysis of Patrick Dangerfield

[Liverpower]

Statistical analysis without any stats?

At least some kind of readable table would be nice.

The other thing when assessing Dangerfield that would be required is time of game spent forward. Adelaide's midfield strength v forward line strength particularly since Walker went down surely means Dangerfield plays more forward than most other mids. That's going to change all his stats.

I fear your collection of stats is not only difficult to comprehend but in the end adds little to the subjective analysis.[/QUOT

You make a lot of mentions to his lower position rankings in various stats but sometimes that kind of data can be misleading when you don't know the actual value difference between those rankings.

e.g. (just pulling numbers here) Dangerfield might be ranked 80th in some stat while other midfielders are ranked around 20th but what isn't shown is that the actual difference between 20th and 80th is only an average value of 0.8, so in reality there isn't really much of a difference there.

Not saying you're trying to mislead people with how you're presenting stats but anyone who looks at it objectively might question why you wouldn't actually include the totals and averages there. Unfortunately for you, being a Port Adelaide supporter in this case, people are going to assume you're pushing an agenda.

Also worth pointing out that the other thread was specifically about whether Dangerfield does or will in the future rank in the top-5 in the competition, so I don't think it is unreasonable to refer mainly to his ranking according to various metrics.

 

[Jcpdragonx]

You underestimate contested ball, has a lot more weighting.

 

[Damon_3388]

Sometimes viewing perceptions are consistent with statistical evidence, sometimes not. Always worth checking it out. Also, it makes the debate more evidence based rather than just a collection of people's perceptions.

I'm still trying to think where/what the actual debate is.

 

[clogged]

Chris Judd - 4.01 tackles per game over his career (3.56 per game from seasons 1-6 as well, which is the age Dangerfield is at)
Patrick Dangerfield - 2.99 tackles per game over his career

Dangerfield isn't great defensively, but there no doubt Judd is pretty average in that area for a midfielder himself.

Fairly ordinary analysis. Judd started in 2002, Dangerfield in 2009 (effectively).

The number of tackles per game in that time has increased by 50%.

 

[master bate]

Also worth pointing out that the other thread was specifically about whether Dangerfield does or will in the future rank in the top-5 in the competition, so I don't think it is unreasonable to refer mainly to his ranking according to various metrics.

But as far as I see it all you are doing is listing where he ranks. You use the term metric but I don't actually see any form of metric.

However simple it may be the AFL have a rating system at Dangerfield ranks at number 3

http://www.afl.com.au/afl-player-ratings

 

[manboob]

TL;DR

 

[Liverpower]

You underestimate contested ball, has a lot more weighting.

I don't underestimate it - it is very important and one reason why Dangerfield is a better player than his raw numbers might otherwise suggest. However, effective disposals, tackles and goal assists are also rather important don't you think?

 

[Liverpower]

But as far as I see it all you are doing is listing where he ranks. You use the term metric but I don't actually see any form of metric.

However simple it may be the AFL have a rating system at Dangerfield ranks at number 3

http://www.afl.com.au/afl-player-ratings

If you read the OP you will see that there are numerous occasions where I list the actual numbers, as well as the rankings. In the next day or so I will put together a table of the actual numbers so that you can see for yourself how he compares to the other players.

 

[Damon_3388]

Fairly ordinary analysis. Judd started in 2002, Dangerfield in 2009 (effectively).

The number of tackles per game in that time has increased by 50%.

Even so, in the heavier tackling era, Judd's tackling numbers vary from inconsistent to unprolific.

 

[GoCROWS!]

Part of why people rate Dangerfield is entertainment- which is why we all watch footy. In my eyes (heavily biased) few players attack the ball like he does.

 

[rose jam]

Sounds like Chris Judd.

Honestly when all the ridiculous fanfare goes away from Judd he is pretty much about the 30th nose effective midfielder in any year of his career from age 24 onwards.
Danger field is very much a one way player which means you tend to remember the parts about him when he has ball in hand, but he isn't great at bringing others into the game. Can be influential at times.

 

[clogged]

Even so, in the heavier tackling era, Judd's tackling numbers vary from inconsistent to unprolific.

Well, I'm not saying he's a master tackler, but at the same age Dangerfield is now, he was in the top 10 tacklers.

http://afltables.com/afl/stats/2006s.html#sh7

As the OP has pointed out, Dangerfield has never even been in the top 100 tacklers. That is poor for a guy who spends the majority of his time on the ball.

 

[strobe00]

Statistical analysis without any stats?

At least some kind of readable table would be nice.

The other thing when assessing Dangerfield that would be required is time of game spent forward. Adelaide's midfield strength v forward line strength particularly since Walker went down surely means Dangerfield plays more forward than most other mids. That's going to change all his stats.

I fear your collection of stats is not only difficult to comprehend but in the end adds little to the subjective analysis.

Thank f*ck for that.

Was hoping I wouldn't reach the end of the thread before common sense made an appearance.

Good work for putting in the hard yards... but a table would have been nice.

 

[nerRADIAN]

It’s not really a statistical analysis if you don’t cover all the stats, you’re picking them at random it seems. I'd be interested in his contested marking stats compared to other mids. Though a lot of what he does there’s no stats sheet for. Look at his intensity and pressure around packs and contested ball situations.

 

[Liverpower]

But as far as I see it all you are doing is listing where he ranks. You use the term metric but I don't actually see any form of metric.

However simple it may be the AFL have a rating system at Dangerfield ranks at number 3

http://www.afl.com.au/afl-player-ratings

The AFL player ratings are interesting. If you look at the methodology, in principle it should be far superior to the fairly basic analysis I have put together. It attempts to control for where a player receives the ball, precisely what they do with it, and how that type of posession would be expected to influence a team's score. However, their system does throw up a number of anomalies. Has Brent Harvey been the 6th best player in the competition over the past 40 games? Has Rioli really been Hawthorn's best player? His poor start to the season aside, how many people would claim that Travis Cloke was only the 171st best player in the competition over that time frame? Like most systems it is heavily biased towards midfielders over defenders and forwards. Their ranking system relies heavily on their assessment of the value of particular posessions as they contribute to a player's total scoreboard contribution. If that assessment and weighting mechanism is incorrect or biased then so will the player rankings. Dangerfield comes out in 3rd place in those rankings, implying that he adds value in a way that is difficult to assess using more standard and easy to understand performance metrics. However, given the near impossibility of replicating the AFL's system (without considerable time and resources), that ranking is ultimately a black box, making it hard to independently evaluate.

 

[Liverpower]

It’s not really a statistical analysis if you don’t cover all the stats, you’re picking them at random it seems.

For example you’re comparing numbers but not mentioning the amount of time he spends in the forward line which is higher than most mids. It was even higher in 2013 due to the lack of key forward option after Tex went down. It’s not coincidence his numbers went down slightly that year. You also make no mention of contested mark stats to which he is very good at for a mid.

A lot of what he does there’s no stats sheet for. Look at his intensity and pressure around packs and contested ball situations.

They are not picked at random. They are statistics that are widely regarded as important. However, I also acknowledged at the beginning that the analysis was incomplete as there are data that I don't have access to. On the question though of intangibles I would ask you two questions. First, if we should rank him higher because of his time up forward, why does he perform so poorly in goal assists? He ought to do better on that metric not worse if he spends more time close to goal. Similarly, if his pressure around the packs is so high, why is he such a poor tackler compared to other mids? His intensity is reflected in his dominant contested posession rankings, so I haven't really left that out at all.

 

[strobe00]

The afl rankings system has well known faults. Any kind of broad brush ranking system will throw up meaningless results.

Every single weighting system will have some kind of bias in it, because you need to weight the various factors to be able to combine them into a score.

Rating systems are a waste of time.

It's almost impossible to compare apples with apples, especially when it's more important how the guy contributes to what his team expects of him, rather than some kind of averaged weighting that assumes one size fits all.

 

[Liverpower]

The afl rankings system has well known faults. Any kind of broad brush ranking system will throw up meaningless results.

Every single weighting system will have some kind of bias in it, because you need to weight the various factors to be able to combine them into a score.

Rating systems are a waste of time.

It's almost impossible to compare apples with apples, especially when it's more important how the guy contributes to what his team expects of him, rather than some kind of averaged weighting that assumes one size fits all.

It is not weighting that introduces bias, but inappropriate weighting where appropriate is judged against the objective that you are trying to evaluate.

 

[Bard]

Also why are all Port fans obsessed with the crows?

You just had to go and say it, didn't you?

 

[nerRADIAN]

They are not picked at random. They are statistics that are widely regarded as important. However, I also acknowledged at the beginning that the analysis was incomplete as there are data that I don't have access to. On the question though of intangibles I would ask you two questions. First, if we should rank him higher because of his time up forward, why does he perform so poorly in goal assists? He ought to do better on that metric not worse if he spends more time close to goal. Similarly, if his pressure around the packs is so high, why is he such a poor tackler compared to other mids? His intensity is reflected in his dominant contested posession rankings, so I haven't really left that out at all.

My guess on the low tackle numbers are that he bursts open packs and gets the ball before the other players around him can. If you watch his stoppage work you'll usually see him diving in and getting his own ball instead of waiting for another player to get it and then tackle them. I don't really rate the goal assist as a stat, why pass it if you can take the shot your self?

Some other stats you missed that I think he would rate very highly in for a mid

-contested marks
-clearances
-goal kicking accuracy
-set shot accuracy
-intercept marks
-hit outs

 

[Del80659]

great read. we always hear people call dangerfield "explosive". i think a fair summary would be to say that perhaps dangerfield needs to work on his endurance, so he can get to every contest, so he can be in every contest. that's how he can produce goal assists and get more disposals of course. this is stopping him from becoming an all-round elite player, he is still a great midfielder as it stands, but we are looking at what he needs to work on here.

he makes me think of jack wilshere of arsenal.. both players so hyped up, both great midfielders but not quite there yet in terms of conducting games and really taking them by the balls consistently. both have immense pressure on them as young lads, too much, over hyped like I said. but if you can pull through being this hyped and become a seasoned player in your mid 20s, the world is yours for the taking. I think dangerfield will end up better than wilshere, if you know what I mean. end up having a better career on his scale.

 

[Liverpower]

My guess on the low tackle numbers are that he bursts open packs and gets the ball before the other players around him can. If you watch his stoppage work you'll usually see him diving in and getting his own ball instead of waiting for another player to get it and then tackle them. I don't really rate the goal assist as a stat, why pass it if you can take the shot your self?

Some other stats you missed that I think he would rate very highly in for a mid

-contested marks
-clearances
-goal kicking accuracy
-set shot accuracy
-intercept marks
-hit outs

He is a very good clearance player - 8th ranking in 2012, 14th in 2013, 30th so far in 2014. Consistent with his being an exceptionally good mid, but not a top 5 mid.

 

[PKnox]

The obsession continues .

 

[pavilion103]

Dangerfield, exposed yet again.

 

[strobe00]

It is not weighting that introduces bias, but inappropriate weighting where appropriate is judged against the objective that you are trying to evaluate.

True. But try to define inappropriate or appropriate weighting when the question is which player is better.

You tune the weightings and see what the rankings look like. If a guy like jack steven, or redden, or rockliff, or neale end up above guys like beams or swan or goddard, then they re tune the weightings to get the outcome they want.

That's my point. They're tweaking the system to bring the players they like to the top of the list. If they are with an eye on order in coaches votes, then that's a fairly solid approach. But if they're just going on who they like most, then there's an inherent bias in that.

The afl ranking system can't rank defenders properly, especially if they aren't strong rebounders.