In the pre-Origin lull, let’s slip in a few stats. I’m sure you’ll thank me later …

Looking through the NRL data yields an array of both fairly obvious results, as well as relationships you might not expect. There are certainly some that don’t jump off the page at you until you take a close look and add a good dose of interpretation! I would include the halves percentage of a team’s defensive workload as one of these (covered in a past blog).

Similarly, while the amount of metres a team carries the ball during the game correlates positively to the points they score generally (though not Rabbitohs and Storm, strangely), how does that correlation behave if you take into account the opposition’s metres and points? ‘Net’ differences, in other words.

As you can see in the examples below, it can cut many ways – the difference in metres gained doesn’t necessarily translate predictably to ‘net points’:

Clearly, the metres gained matters considerably to the team above when it comes to scoring points relative to the opposition. The relationship is astonishingly predictable, and if you were to tell me this team, let’s call them the Knights, shall we, would have a +500 metre advantage over their opposition in a particular game, I’d expect them to win by about 25 points. Or, a -200 metre deficit would translate to a loss of about 15 points.

Attacking prowess and thrust, alongside defensive techniques and attitude can really dilute this relationship. For a team that doesn’t rely as heavily on possession and territory for their points, the relationship looks more like this:

Let’s call this team Manly.

You can see that I can’t make the same kind of back-of-the-envelope quantitative judgement based on these results, and the R-squared (a measure of data fit) of 0.179 tells me I’d be a fool to try. It’s almost completely random!

With the Knights, though, an R-squared of 0.915 (where 1 is a perfect fit) tells me I could be quietly confident of the nature of the correlation of net metres into net points.

Unless you’re gambling, predicting what the difference in metres gained will be in an individual game is irrelevant. Its application is better reserved for planning your approach to playing a specific team.

In this case, it would be quite useful for a team playing the Knights. You know that if you take an even share of the ball and translate it to a relatively equal performance in metres gained, then you’re a better than even chance of winning the match. Your game plan might therefore adjust to focusing more on ball control over and above intricate set plays and other aspects of the game plan.

As of Round 15, here are the R-squareds for all NRL teams, showing quite emphatically the teams that rely more on possession and metres in order to outscore the opposition than others.

The two immediate observations are:

– The Knights’ net points trace an almost unbelievably regular path based on metres gained against the opposition. Their attack is good enough to post a lot of tries, but clearly needs to get a roll on beforehand, and needs a lot of ball. This is one reason why even in the unlikely event they make the finals, they won’t last long.

– Manly and the Storm, as top-4 teams, can win ugly. Even if you are taking 55% of the ball from them and making significantly more metres during the game, they can find a way to win.

The ability to remain in the top-4 based on stats such as these says a lot about the Storm and Manly, and makes them dangerous in a finals situation.

The Rabbitohs are mid-field in this statistical category, highlighting that they are a bit of a mixture. It is worth noting, however, that there have only been three games where they have lost the metres battle this year, the solitary loss coming to the Storm. The worst net metres performance actually resulted in a win against Manly.

For the other top-4 side, the Roosters, their net metres performance is also fairly regular and strong. For a team who have struggled to post points in recent years, this is the year you don’t want to give them a lot of ball and get a roll on.

A not-so-obvious observation is that the data works in reverse too. Net points can also be negative, and there are many teams, like the Tigers for example, that are the complete opposite to the top teams – the Tigers have only had three games where they won the metres dual, while the Rabbitohs have won all but three. Here’s the Tigers’ chart followed by the Rabbitohs:

Then there are other cases such as the Dragons, who have a habit of winning the battle of metres, but who still have an enormous dispersion of results. Clearly their attack is missing some vital ingredients. They simply should not be losing games with 200-400 metres advantages. The three losses you see below are the difference between their current position and sitting relatively comfortably in the top-8.

Anyway, just a few observations.

Enjoy your pizzas tonight, one and all!

Nicely done Dr.

I’ve noticed a trend in your statistical analyses – the stats that are most mentioned by commentators appear to be the least important in terms of winning football games. Perhaps they need to read you blog?

On a similar note, where do you uncover your statistical data? I’m assuming you don’t freeze-frame every match, every round and add up each tackle and metre.

yes, I do …

Just kidding. I do compile them, though. You can find all the stats you want on the score.com.au and nrlstats.com.

There are far more stats I’d like to keep, but they’re either too time consuming (individual players, for example), or require the sort of effort you alluded to. In this category, I’d like to track the flow of penalties, but there’s no way I’m watching every second in order to do so.

If you know, or anyone else does, how to this, I’d be eternally grateful. The penalty count tells me only the accumulated tally, and says nothing about the flow.

Re stats in general, I like to take a bit of a left field approach and see what works. I’m sure no one has looked at an r-squared of a net difference between metres and points before! Only a lunatic …

This is how I uncovered the halves tackle % … Originally I ‘assumed’ making the smaller guys do more defence hampered their attacking performance. So i looked into it.

Turns out, it’s not true at all. Look at the number of actual runs from many of these guys and you’ll conclude that they play more of a conduit role, while giving the big, lazy forwards a rest:)

Thanks for Q. , and I’m glad you liked.

Sent from my iPad