Sometimes, numbers don’t lie. That’s true. At other times though – like in the recent article at nrl.com proclaiming Bulldogs prop Aaron Woods as the statistically superior footballer to Panthers forward Reagan Campbell-Gillard – they evidently do. Or in the very least, they tell a truth that’s been completely misinterpreted by an author with no understanding of what the numbers actually mean (we’re giving the author, Michael Chammas, the benefit of the doubt here, and assuming that he’s simply ignorant of the nuances of different stats, and not being deliberately misleading).
Taking it from the top
Let’s start from the beginning. How on earth could somebody make an argument that Woods – the highly-paid but widely-panned Bulldogs recruit – is outperforming the leader of the Panthers’ outstanding engine room? Chammas makes his argument based on the following statistics:
|Post Contact Metres||386.8m||581.2m|
According to Chammas, “it is Woods who is on the dominant end of those figures”. That may be true; the problem, though, is that those figures are largely meaningless, and cherry-picked to ignore the numbers in which Woods is worse (which just so happen to be the areas that actually matter).
A conscious uncoupling
There are two things that are immediately obvious about the offensive stats listed. The first, is that they’re all volume stats – overall totals, rather than values per carry. This is important, as the overall volume is largely irrelevant to the performance of an individual – while it’s critical for teams to total as many run metres as possible (for example), it makes no difference to a team whether all of those metres come from one player, or are evenly dispersed among all 17. It’s the team’s overall total that’s relevant. As a result, an individual’s overall total is of less importance than his per carry totals – since a team will only have so many carries in a match, the overall total will be at it’s highest if they can maximise the value of each individual carry.
Subsequently, it’s worth pointing out that none of the offensive stats listed are mutually exclusive. The first line asserts that Woods averages more carries per game than Campbell-Gillard – that’s a fact. But from there, all the other offensive numbers – metres per game, post contact metres per game, and offloads – are all directly influenced by the fact that Woods is carrying the ball more in the first place. Since we already know that Woods is making 4.5 more runs per game than Campbell-Gillard, a higher metres total isn’t a revelation, it’s an inevitability.
So, if we’re going to compare the two players’ offensive impact, apples to apples, we need to unpack these numbers, and start comparing the players’ per carry impact. And by that measure, it’s Campbell-Gillard who comes out on top (numbers via Fox Sports Lab):
You’ll notice that we left off ‘Offloads’ from the table. This is for two main reasons.
Firstly, because it’s irrelevant to offensive impact. There’s no correlation between total offloads and points scored, so the ability to produce more offloads is of debatable value at best (and not all offloads are created equal anyway; a player who attracts two defenders and pops a ball for a line break has clearly done more for his team than a player who backs into the defensive line and drops the ball on the ground). Consider the current NRL ladder; the top 3 teams in offloads – the Warriors, Dragons and Sharks – rank 7th, 1st and 12th, respectively in points scored (and Woods’ own team, the Bulldogs, rank 15th – despite featuring Woods and his apparently invaluable offloads).
Secondly, if Woods’ penchant for offloads does has a substantial impact on his team, it mightn’t be in the way he’d hope. Yes, he makes more offloads, but perhaps as a result, he also makes more errors – a lot more. Woods has so far made 9 errors, compared to RCG’s 4, and on a per carry basis, that translates to 0.6 errors per carry (1 in every 18 carries) vs 0.4 (1 in every 27).
Oh, and while we’re comparing offensive contributions, it’s probably also worth mentioning that despite making less runs, Campbell-Gillard has more try involvements (2 v o), line breaks (2 v 0), and tackle breaks (7 v 4). The only relevant stat in which Woods has the upper hand is line break assists (2 v o).
And on the other hand… RCG is still better
Ok, so Campbell-Gillard is the more valuable on offense. But what about defense? Chammas referenced two defensive statistics, so Woods is definitely better at that, right? RIGHT?!
Well, sort of, but not in a significant way. You see, again, the two numbers listed aren’t mutually exclusive. One number is missed tackles, and the other is tackle efficiency – a number produced by deducting the percentage of missed tackles and ineffective tackles from the total attempted tackles. In effect, they’re two numbers representing the same thing – the fact that Campbell-Gillard misses more tackles than Woods (true).
However, again, it’s debatable how relevant a statistic that actually is. The correlation between missed tackles and points conceded is incredibly weak – in fact 2 of the top 3 teams to have missed the most tackles (the Panthers and Tigers) actually rank in the top 3 for least points conceded (and only 1 of the 5 teams to have missed the most tackles is also in the top 5 for most points conceded). There’s a much stronger correlation between line breaks conceded and points conceded (of the 5 teams to have allowed the most line breaks, 3 of those are also in the top 5 for points conceded). And how does Woods fair compared to Campbell-Gillard? Well, he’s directly conceded 4 line breaks, while Campbell-Gillard has conceded just 1.
Of course, there’s more to defending than simply making tackles, right? Like, for example, you would surely prefer a defender who does their job without conceding penalties, and handing the opposition an extra possession (since time in possession is a crucial factor in determining the outcome). In this respect, Woods falls short again, conceding 12 penalties (1 in every 22 tackles) to RCG’s 7 (1 in every 38).
Drawing a long conclusion
Finally, as though the overall misleading message of the article isn’t enough, Chammas even manages to draw the wrong conclusion from it. Having deployed his statistics in favour of Woods, Chammas asserts that “the fortunes of respective teams somewhat dilute perception of how a player is actually performing”. While we’re not saying that on occasion this isn’t necessarily the case, in this particular example, it’s actually Campbell-Gillard who is the victim of his own team’s ability, rather than the other way around.
Let’s reflect on Chammas’ numbers – offensively, in particular, these are all volume stats, so Woods is made to look better by having had more runs. But why do you suppose that is? Let’s compare the two teams’ expected forward packs for this week:
Panthers: Trent Merrin, RCG, Corey Harawira-Naera, Isiaah Yeo, James Fisher-Harris, Viliame Kikau, James Tamou, Jack Hetherington.
Bulldogs: Woods, David Klemmer, Josh Jackson, Raymond Faitala-Mariner, Rhyse Martin, Adam Elliot, Clay Priest, Renouf Toomaga.
If you want to know why Campbell-Gillard is taking less runs per game than Woods, the answer is contained in the players above. In short, every time Campbell-Gillard takes a run, that’s a run that’s being denied to one of the many other international-calibre forwards at the Panthers’ disposal. It’s in their best interests to share the carries around, regardless of how well Campbell-Gillard is traveling. The same isn’t true at the Bulldogs, where if Woods doesn’t take the hit-up, then who does? Josh “7.3 metres per carry” Jackson? Adam “7.5 metres per carry” Elliot? The volume numbers that have been selected are actually boosted by the fact that Woods plays for a weaker team, where he’s required to take on a larger share of the workload, rather than the opposite, as has been asserted.
So, there you have it – Aaron Woods is not better than Reagan Campbell-Gillard. The numbers don’t lie (as long as you’re speaking their language).
If you’re wondering who the true forgotten man of New South Wales is, it’s Panthers forward James Tamou, who is either as good or better than both the players listed in almost every relevant statistic, besides line breaks conceded (5). Unfortunately for Tamou, in this instance, Chammas would be right – the overall performance of the Panthers is diluting the perception of Tamou, leaving him the victim of being just another high performer in an excellent forward pack.