The long 27 round slog that was the NRL 2023 season has finally concluded, which means we can take a look at which players topped the charts for the Eye Test’s suite of advanced statistics.
If you’re a recent visitor to the Eye Test, I would recommend checking out the explainers on of them to get an idea of what type of player behaviour I’m trying to measure with these metrics. These aren’t your typical counting statistics where everyone is trying to fill up a bucket and the more things you put in your bucket the better. For these statistics we’re looking at incidence rates or trying to attribute team performance to specific players.
All of these metrics are possession adjusted of course. Penrith averaged 160 play the balls per game, 10 more than any other team and nearly 30 more than the Bulldogs this season.
That means Panthers players have at least 7% more chances to accumulate statistics than any other team, and 20% more than a Canterbury player. So to combat this, all statistics are adjusted for possession to ensure that players are on equal footing.
I’ll also remind everyone that there’s a first half advanced statistics wrap on the site, but I figure the only people who want to specifically celebrate the first half of the season at this point of the year are Souths fans.
Let’s dive into the numbers.
Eye Test Player Contribution Ration (ETPCR)
First up we’re going to look at the sites player attribution metric, the Eye Test Player Contribution Rating (ETPCR), which assigns a weighting for statistics that correlate well with winning and positive margins (eg tries, try assists, line breaks, tackle busts, run metres) and also those that correlate with losing and negative margins (eg errors, penalties conceded, missed tackles, causing a try or line break).
Then we apply those weightings to each player’s statistical output for a game and adjust it for possession and you’ve got an estimate of how many points (positive or negative) that player was “worth” for that game. Add them together and you’ve got their season value.
A small housekeeping note for ETPCR before we move forward. I’ve set the threshold for games played at 13, meaning everyone on this list has played at more than half a season. For ETPCR, the only player inside the top 50 disadvantaged by this is Tom Trbjoevic who would place 7th at +2.142 from his 11 games.
Below is the top 20 for 2023. All numbers used are from Fox Sports Stats.
Shaun Johnson takes top spot for 2023, with an ETPCR of +2.717. This estimates he was worth an extra (possession adjusted) 2.7 points per game for the Warriors this season. It was his best season for this metric since I’ve been tracking data in 2014, and a huge improvement on a substandard 2022 campaign.
It was a very close race in the end. Johnson was leading into Round 27, when he wasn’t playing. Second placed Nicho Hynes entered the weekend with a season average of 2.622, and only needed to increase it by .010 to take top spot. He only moved his average by just under 0.05, meaning Johnson takes the crown for the season. The year of #ShaunHub continues.
In addition to this, Johnson would have also been the winner of my proposed Dally M medal voting system based on ETPCR, as he had an unassailable lead of 11 points with three games to go. If you want to read about it, the post is here. Will Johnson do the same later this season for the real thing? One can hope.
Moving further down the list, Reece Walsh rounds out the top three at +2.587, with Nathan Cleary fourth at +2.515 and Jahrome Hughes fifth at +2.495.
Ponga’s numbers in sixth spot are incredible if you take into account he was either ineffective or inactive until Round 13. From that point on, when he was moved to fullback, he would have led the NRL in ETPCR at +3.436, ahead of Johnson at +3.403.
Another Knight in Greg Marzhew had an outstanding season placing 18th spot at +1.376. That is incredible if you consider he was the 11th worst player in the competition last season at -1.360. A lot of things Adam O’Brien has done this season deserve credit, but taking Marzhew from being unable to secure a spot on the Titans to one of the best wingers in the competition might be some of his best work. Jackson Hastings also cracked the top 20 at +1.338.
Everyone down to Kalyn Ponga has made the top 20 season average for ETPCR since 2014. The list of the top 20 over the past 10 seasons is below.
Back to 2023, Charnze Nicoll-Klokstad looks like one of the buys of the season here, placing 9th and third among eligible fullbacks with an ETPCR of +1.756 per game. Payne Haas unsurprisingly takes the top spot for forwards at +1.692, with Scott Sorensen the next forward just outside this list at +1.256. Addin Fonua-Blake (+1.228, 23rd) was just outside as well while Isaah Yeo (+0.943), Toafofoa Sipley (+.978) and Pat Carrigan (+0.808) were inside the top 35.
One of the reasons Haas was so important for Brisbane is the sheer absence of negative play in his game. He had the lowest negative component of his ETPCR in the NRL at just -0.120 per game. This part is made up of missed tackles, errors, penalties or set restarts awarded, line breaks or tries caused and sin bins or send offs. Haas is down for just 1 line break cause and zero try causes according to Fox Sports Stats, missing just five tackles, committing just three errors and conceding just four penalties all season. Given the minutes he plays these figures are astounding.
It’s also a reason why Sipley ranks inside the top 30, having the 12th lowest negative component of his ETPCR at -0.384 per game. Below are the 20 players with the lowest negative contribution to their ETPCR, so you could say they’re the “safest” players.
Part of that is due to tries and line breaks usually being conceded around the edges. But if you look at the players who have the highest negative contribution there are still some forwards among the centres and wingers, most notably Reed Mahoney, Corey Waddel and Elliot Whitehead. Villiame Kikau and Billy Burns would be there too but he didn’t play enough games (9) to qualify, although Burns is mostly there as collateral damage as we’re about to find out.
On to the bottom 20 players by ETPCR this season.
Mathew Feagai’s -2.672 was the worst in the NRL this season, and somehow played in all 24 games. The thing about the bottom end of this list is that most of these players only end up playing half a dozen games. If I take out the 13 games played threshold there’s another 12 players below Feagai, and not a single one of them played more than 7 games. Generally bad players get a short leash and are far more likely to be hooked than given the chance to turn things around.
That also applies to the second worst average, which belongs to Paul Alamoti with an ETPCR of -2.244. Again, somehow he played 19 games for the Bulldogs despite drawing the ire of From the Couch’s Nick Tedeschi every time he stepped foot on the field.
The Gold Coast’s Jojo Fifita had the third lowest number at -1.818 from his thirteen appearances. Kikau would have placed fourth here if he’d played enough games, whilst the other notable Bulldogs recruit in Mahoney had the actual fourth worst number at 1.635. That was only slightly worse than the man who replaced him at Parramatta, with Josh Hodgson averaging -1.605 from 12 games, and again just missing out on the cut off.
Isaiah Papali’i’s appearance in this list is mainly due to the dearth of attacking stats he picked up for the Tigers this season. His negative numbers weren’t that bad but without the line breaks, tackle busts, offloads or tries he was picking up with the Eels he didn’t have enough numbers to lift his overall ETPCR.
Feagai also ended up with the worst average since 2014, beating out the previous low of -2.546 from Pat Mata’utia from one of those Nathan Brown coached Newcastle sides in 2016. Jojo Fifita, Alamoti, and the Tigers duo of Asu Keapoa and Junior Tupou also ended up in the bottom 20 in the last ten seasons based on their 2023 ETPCR.
Again I’d like to point out there are plenty of players with a season ETPCR lower than Feagai, but none of them have played more than 13 games in a season.
Now we’ll move on to the rate based metrics, starting with Error Rate.
Error Rate
I’ve defined Error Rate as the average number of touches (or possessions/receipts, pick your own nomenclature) that occur for every error a player commits. I’ve set a different threshold here, mainly because players who regularly make errors generally don’t play a lot of football. Unless they’re a Dominic Young or Reese Walsh where their overall value outweighs a lost ball or an errant pass.
So the threshold here is just seven errors and six games played. I could have made it even lower to accommodate the Bulldogs Declan Casey, who would have ranked first with seven errors from 33 possessions at a rate of 4.71 touches per error, but he only played the three games. And that’s probably why he only played those three games.
With that out of the way, here are the worst error rates in the NRL this season.
Souths winger Izaac Tu’itupou Thompson claimed the worst error rate of the season, commiting 14 errors in 7 games from 109 possessions for a rate of 7.79. All of those errors were in his first six games of the season, his seventh game in Round 27 was a clean sheet, meaning this number was even worse a week ago. Last year’s worst hands went to Super League bound Jayden Okunbur, who committed an error every 9.44 receipts in six games for the Dogs.
The only other player with a rate lower than 10 touches per error in 2023 was Gold Coast winger Alofiana Khan-Pereira, who committed an error once every 9.52 touches of the ball. As error prone as he was, it goes back to my comment before about players needing to have a high overall value to outweigh any handling issues.
And Khan-Pereira showed that this season, ranking equal sixth in tries scored with 20 from his 23 games, on the same number as Maika Sivo, Greg Marzhew and Ronaldo Mulitano. If you’re among that company then the odd handling issue isn’t a problem.
The second leading try scorer this season in Young didn’t have the best hands, with 35 errors and an error rate of 10.97, 4th worst in the NRL. Maybe it’s an English thing and not a Burgess genetic trait? Sivo wasn’t much better than either Young or Khan-Pereira, with an error rate of one every 13.2 touches. The point is, if you can score tries regularly, a coach will put up with errors.
The rest of the top five comprised of Billy Burns from the Dragons, who shows up for the second time in a negative statistic and Brent Naden from the Tiger both averaging an error roughly every 11 touches of the ball.
This is Naden’s third season in a row that he’s placed inside the top 5 for this metric. But at least he’s improving. This year’s average of 11 touches per error was ahead of his 9.93 from last season (9.13 in eight games with the Dogs, 10.79 in 10 games with the Tigers) and 7.92 in 2021 with the Panthers.
Tackle Rate
This was one of the first advanced metrics for rugby league that the site used, and the easiest way of describing it is an estimation of how often a player completes a tackle, adjusted for possession and minutes played. Or tackles per minute adjusted for possession.
After all, you can’t complete a tackle if your opponent doesn’t have the ball. And if your opponent has the ball a lot, and runs in your direction, you’re going to make a lot of tackles.
I’ve set the threshold here at 250 minutes played for these rate based metrics as we’re looking at specific events within a game, rather than a number of games. There’s no scientific reason for 250 minutes, other than it appears to give good results.
Here’s the top 20 for 2023.
Cameron McInnes takes top spot this season, the first time a player who had played more than 1,000 minutes has led this metric in a long time. It’s usually led by a lower minute bench forward, playing between 30-50 minutes per contest.
However this year McInness made 777 tackles at a rate of 32.54%, indicating that he completed a tackle on one third of all play the balls the Sharks defended. The only player in the top 20 who spent more time on field than McInnes was Max King from Canterbury, but even then it was just one more minute and his tackle rate was 3% lower.
Canberra’s Pasami Saulo ranked second with a rate of 32.21%, but played 22 fewer minutes per game, which puts McInnes’ impressive workrate in perspective. Canterbury middle Ryan Sutton rounded out the top three with a rate of 31.56% in his 13 appearances.
I’d also like to make note of Eye Test inaugural Hall of Famer Christian Welch ranking 21st and just outside the above list at 28.23%, just behind teammate Bronson Garlick at 28.27%.
Ball Runner Rate
This measure is similar to tackle rate, but looks at how often a player completes a run with the ball, adjusted for minutes played and adjusted for possession. Keep in mind this is not a qualitative measure, just how often they make a run and not how useful that run is.
Here are the top 20 players for 2023.
Souths middle Tom Burgess takes first place this season, with a Ball Runner Rate of 15.89%, completing a run on three out of every 20 play the balls the Bunnies had this season. Last year’s top spot was taken by Rhyse Kennedy at the Broncos with a rate of 15.86%.
Burgess was followed by a pair of Eels middles, with Wirimu Greig ranking second at 15.72% and Ofahiki Ogden in third at 15.30%. Eels interchange middles usually rank high in these metrics as Brad Arthur usually plays Junior Paulo and Reagan-Campbell Gillard high minutes, so their bench forwards can expend extra energy in shorter stints on field.
Manly’s forward Sipley placed fourth with a ball runner rate of 15.13% and Dragon/Dolphin Josh Kerr was fifth at 15.11%. These were the only five players to have a rate higher than 15%, last year there were six players above that rate and just four in 2021.
The 250 minute threshold applied to these rate based metrics has robbed South Sydney’s Daniel Suluki-Fifita of first place here, as he had a rate of 17.13% but only played 246 minutes, meaning he was 5 minutes short of claiming top spot from his teammate. He would have ranked first in 2022 as well with a rate of 19.34%, or a run on one in every five Souths’ play the balls, but only played 231 minutes.
Total Run %
This builds on the Ball Runner Rate above, by adding in option runs – decoys or supports – to get a better picture of how often a player completes a run whilst on the field. The same adjustments (minutes played and possession ) are made here.
Below are the top 20 for 2023.
Once we add in in option runs, we have *checks notes* A TIE? There’s a first time for everything.
Burgess still claims top spot for this metric once we go down to three decimal places. The Souths prop completed 104 decoy or support runs in his 10 games, for 336 total runs at a rate of 23.017%, or a run on nearly one in every four Souths play the balls whilst he was on field. Last year’s winner was Griffin Neame of the Cowboys with a rate of 23.38%.
Burgess’ teammate Shaqai Mitchell ranked second by the tiniest of margins with a rate of 23.016%, with a 1.67:1 ratio of runs to option runs. I’ve never seen one of these metrics decided by such a small margin, it’s usually a few hundredths of a percent, not thousands.
Sipley again ranks highly in another Eye Test metric, with his 188 runs and 95 option runs in 40.4 minutes per game giving him a Total Run % of 22.77%.
Blake Lawrie’s 190 option runs ranks 2nd of all players with at least 250 minutes on field, only behind Melbourne middle Josh King, who was recorded making 203 support or decoy runs this season.
Kerr’s 21.03% and fifth spot comes down to how he was used at the Dolphins. He was completing a run at just 18.8% for the Dragons in eight games with an average of 29.4 minutes on field. Once he moved to the Dolphins, his rate increased to 22.3% from just 28 minutes per game.
Involvement Rate
This last advanced statistic combines the running and tackling rates above to create an overall metric to measure the quantity of their workload. Like with the previous statistics, it’s purely quantitative, and as the great John Wooden said, “never mistake activity for achievement”. Although in most cases Involvement Rate usually matches the eye test.
Here are the top 20 for 2023.
Whilst Mitchell may have been cruelly robbed of first place for Total Run % by a hair, he takes out first place for Involvement Rate once we add in his defensive effort. Mitchell completed a tackle or run on 26.64% of all plays during Souths games this season, or once every four play the balls in attack and defense. Last year’s winner was Fletcher Baker, with a rate of 26.45%.
Cronulla’s Thomas Hazleton was second and the only other player above 25% at 25.87%, with the Roosters bench middle Terrell May rounding out the top three at 24.72%.
Melbourne interchange player Alec McDonald ranked fourth at 24.60% while Warriors cult favourite Bunty Afoa just finished 5th with a rate of 25.54%.
The volume of work McInnes does in defense compared to attack can be seen here, as he ranked first for tackle rate but 13th once we add in any type of run, at a rate of 23.65%.
I’d again like to note Eye Test Hall of Famer Christian Welch ranking 20th here at 23.13%, one of the higher rates in the NRL for players averaging around almost 50 minutes per game or more.
