Pulling strings
Pulling strings
Where do the best playmakers play in today's NRL?
Welcome to the fifth edition of Plot the Ball — a newsletter where I offer data-driven answers to interesting questions I have about the world of sport.
This month — after taking in the first half of the 2022 NRL season and enjoying the recent State of Origin series opener — I’ve been thinking about attacking play in elite men’s rugby league, and wondering how the balance of teams’ playmaking responsibilities are distributed.
Where do the best playmakers play in today's NRL?
An interesting (if rather provocative) way to think about the two codes of rugby is that, for fans of skilful ball movement over run-it-straight tactics, rugby union provides better entertainment value for three-quarters of the length of the field — but rugby league really comes into its own at the business end.
In the 15-a-side game, the range of attacking possibilities often seems to contract rather than expand as teams advance the ball up the pitch.
Think of the scenes at the end of the recent Champions Cup final, where winners La Rochelle battered away at the Leinster try line for the last 10 minutes of the game — with little interest in spreading the ball to width — until they eventually ground their opponents down and scored.
In the 13-player code, however, things are turned on their head.
Teams in possession inside their own territory are so determined not to turn the ball over to the opposition in dangerous areas that attacking sets remain very conservative, as they look to achieve maximum yardage on each carry with minimal risk.
Only once the ball has been moved beyond the defensive team’s 20-metre line is their full range of attacking shapes finally unfurled.
Strings of intricate passes behind decoy runners often end with the ball in the hands of touchline-hugging wingers, ready to explode into the air and finish acrobatically next to the corner flag — like this effort from the New Zealand Warriors’ Dallin Watene-Zelezniak last weekend.
While league players can add value all across the park in myriad ways, when it comes to these attacking sets in the red zone it is the two halves — the five-eighth and the halfback — who clearly run the show.
It’s unsurprising, therefore, that data from the last 10 seasons of Australia’s National Rugby League (NRL) — the code’s preeminent club competition — shows that players in these positions have been consistently the most involved and effective playmakers on the field.
Per the chart above, players starting at halfback and five-eighth have — on a per-80-minute basis — provided more try assists to their teammates than those in any other position in each of the last 10 seasons.
In three seasons — 2014, 2020 and 2021 — players playing fullback have put up half-like numbers, averaging 0.5 try assists or more per 80 minutes played, but in each case their assist rate was lower than those recorded by starting halfbacks and starting five-eighths.
(Manly Sea Eagles fullback Tom Trbojevic’s 2021 season was one of the most remarkable in recent NRL history, as he notched up 28 total try assists at a rate of 1.6 per 80 minutes; ABC’s excellent journalist Jack Snape broke down the attacking threat he posed in a piece last September.)
At an individual level, the identity of the best playmaking half in the game over this period is also entirely unsurprising. Despite this data only being available for the tail-end of his storied career, NRL legend Johnathan Thurston is way out in front of his positional peers.
Aggregating minutes played in starts at both positions (halfback and five-eighth), the former North Queensland Cowboy averaged 1.1 try assists per 80 minutes between the start of 2013 and his retirement at the end of the 2018 season.
For context, among other players with more than 10,000 minutes in starts in the halves since 2013, the highest rate is Cody Walker’s 0.8.
(There’s unlikely ever to be another Thurston, but current Sydney Roosters pivot Sam Walker — who only turned 20 years old on Wednesday — is setting a strong pace in the race to be the outstanding playmaker of his generation. Through approximately 2,000 minutes as a starting NRL half, he’s already averaging 0.8 try assists per 80 himself — and is still years away from his likely peak.)
Looking at the top players by this metric in other positions does, however, reveal some more interesting conclusions.
Among players currently active in the NRL, the Cronulla Sharks’ Nicho Hynes (1.0), the Canterbury Bulldogs’ Matt Burton (0.5) and Josh Schuster of the Sea Eagles (0.4) lead the way in try assists per 80 among fullbacks, centres and second rows respectively.
Intriguingly, each of the three might now spend the bulk of their careers playing away from those spots and moving to — you guessed it — the halves.
Hynes broke through with the Melbourne Storm before moving to Cronulla for the 2022 season, and an injury to star fullback Ryan Papenhuyzen in 2021 gave him an opportunity to start a series of games at the back in his place.
The Sharks, however, signed him with the intention of playing in the halves — and he’s started every game at halfback for Cronulla so far this season.
Burton’s career has followed a similar trajectory. While he did start a number of games in the halves for the Penrith Panthers early in his career, he was unable to displace their stars, Nathan Cleary and Jerome Luai — and it has taken a move away this year for him to get consistent opportunities at five-eighth.
Schuster, meanwhile, has not yet been afforded such a sustained chance.
He has played only a handful of games so far this year due to injury — and it’s only in the last few rounds that he’s been thrust into the five-eighth position.
Whether Schuster will get what he wants in Manly — or whether he’ll need to move away to be entrusted with the keys to an NRL team’s attack — remains to be seen.
What the early career paths of all three players do show, however, is the value that NRL teams continue to place on selecting their best playmakers at halfback and five-eighth in order to get the ball in their hands as much as possible in dangerous areas.
If you watch a single game of league, you might well arrive at a similar conclusion about the importance of these two positions. Look at the numbers for the last decade of NRL action, and you almost certainly will.
Further reading
Jack Snape of ABC on Tom Trbojevic’s fullback play and Cody Walker’s pursuit of Johnathan Thurston’s single-season try assist record
Technical notes
You can find the code for this piece on GitHub here
Since I’ve been putting together this newsletter, there have been editions where I’ve been led by my interest in a very specific question I wanted the answer to — and others where I’ve set myself the challenge of extracting a certain dataset, and the theme of the piece has been settled later in the process. This month’s was very much one of the latter editions. I knew that NRL.com had player data going back as far as 2013, but that they did not summarise it in full in the Stats section of their website — and so I wanted to try and build a complete database myself. The way the site is constructed meant that scraping it with
rvestwasn’t going to cut it, so I spent a fair amount of time getting to grips withRSelenium— an R package which allows you to access Selenium’s browser automation capabilities. The set-up was fairly laborious, but the time investment was certainly worth it, as learning how to use the package has now opened the door to scraping a much wider range of web content thanrvestalone can manage.In a similar vein, I enjoyed Giorgio Comai’s distinction between ‘data-hunters’ and ‘data-gatherers’ in the most recent edition of Giuseppe Sollazzo’s great weekly newsletter, Quantum of Sollazzo:
Doing the work for this edition felt much more like hunting than gathering.
I was really pleased with how the heatmap in the first graphic — made using
ggplot’sgeom_tile()function, with labels overlaid usinggeom_text()— turned out, even if there wasn’t a really striking trend in the data to convey. I’ll definitely be returning to it as a chart type for depicting time-series data — and this R Charts guide was useful in working through some of the customisation options available.
Next month — to coincide with the end of the 2022 Wimbledon Championships — I’ll be looking at the early career trajectories of some of the up-and-coming stars of both men’s and women’s tennis.