NRL 2021 – how the game changed statistically this season

The year of our lord 2021 has not been a banner year, and for our great sport of rugby league it was especially troubling. Thankfully after a one-sided regular season we’ve seen some close and exciting finals games, but it’s difficult to cover up the changes and struggles fans have had to endure this season.

The Eye Test has already summarized all the ridiculous one sided and outlier statistics of the season, but what we haven’t covered at is how the game as whole looked in 2021 and what changed outside of those beatings.

I’m not going to do this on a team-by-team basis, as I couldn’t compete with the tremendous season recaps for each team that friend of the site Liam of PythagoNRL has been producing. They’re a must read for every fan, especially the ones that don’t involve your own team. Instead I’ll be taking the same path as last season, looking at the league as a whole on a macro level to see what trends occurred or what changes came out of the season.

Before we start let’s clear up the methodology. Firstly, as always, I’m using the Fox Sports data, other than time in play which is sourced from the NRL site. Secondly, all percentage changes are comparing per game averages (unless otherwise specified), as there were only 20 rounds last season compared to 24 games and 25 rounds in 2021. It’s hard to show the change in run metres from 2800 to 2900 on the same chart showing an increase in metres per run from 8.86 to 8.93.

The charts are set up showing the change from 2020 to 2021 (orange dots) or 2021 to 2020 (blue dots). Orange dots to the left means an increase on 2020, orange dots to the left means a decrease from 2020 with the bar in the middle showing the gap between the change from 2020 to 2021 or from 2021 to 2020. It’s not elegant but I find it’s an effective way of highlighting the largest changes year on year.

Scoring and attack

This is probably the least shocking chart in this post. Double digit increases in all scoring categories, with points and tries up 11% on last season. The biggest change though is field goals kicked, which increased by 20%. The two point field goal may just be the only decent rule change brought in by the current administration.

An easy way of visualising the change in attack this season is looking at heat maps of tries scored for 2020 and 2021, which can be seen below.

An increse in intensity of data points in the middle of the in goal area can be seen here. Traditionally under the posts hasn’t been a high scoring area for NRL teams, yet in the era of over fatigued middle forwards the number of tries scored by big bodies just waltzing through the centre of the park and over the line jumped significantly.

Time in play

For all the talk of wanting to speed up the game and reduce stoppages, the one thing that increased this year was stoppages.

The average time in play dropped by nearly two minutes, from 55.9 minutes of play down to an average of 54.1 this season. For all the talk of wanting to speed the game up, the average time in play in 2021 was almost identical to 2019 (54.2). Another example of fixing a problem that didn’t exist.

The startling part of this drop is that not only was it a sizable drop in ball in play, but since Round 6 only one round averaged more than 55 minutes of play in game (Round 16) and all but a handful of rounds averaged more than 54 minutes per game. The average ball in play since Round 6 was only 53.5, well over two minutes per game down on last season.  

Nevermind the fact that trying to keep the game flowing and trying to increase the amount of points scored are mutually exclusive. You can’t have more time in play and fewer stoppages with more tries, because the mere fact of scoring creates stoppages. The more tries you have the more stoppages there are, as you can see by the below chart plotting average number of tries scored per round (blue columns) with the average time in play (orange line) for the past three seasons.

You can have more time in play or more tries. Not both.

Possession

One change this season was that there were fewer play the balls on a per game basis compared to 2020. That’s after a 7% increase in 2020 and well above the the average back to 2014, so we’re still living in an era statistial inflation. Keep that in mind when you’re comparing 2021 to previous seasons.

Total play the balls declined by 2% per game, with teams attempting 2% fewer sets. In per game averages this equates to 288 play the balls per game (down from 295 in 2020) and 78 sets per game. The number of tackles per set remained constant however (3.7 per set), so when they did have possession, teams weren’t doing more with less, or less with more.  

Completion rates, which we established earlier this season as a myth and have no correlation with winning games, barely changed from last season. Completion rates were on the same level as 2020 at 78.4% this season. This comes after the completion rate from 2015-2019 had sat firmly in the 76% range before a big spike of +2% in 2020 with the set restart change.

With the increase in scoring and attacking plays this season, it makes sense to see a 7% increase in teams tackled in an opponent’s half and an 8% increase of being tackled inside an opponent’s 20 metre area.

Set Restarts & penalties

With now two infringements (Inside 10 metres joining ruck infringements), the number of set restarts called jumped by 6% this season from 7.1 to 7.5. With 10 metre infringements ranging from 20-30% of set restarts called per round, that means ruck infringement transgressions actually declined this season. I’m sure the players were much better behaved in 2021 despite being considerably more fatigued.

The below chart shows the average penalties awarded per round for the past two seasons.

You can see the Magic Round “crackdown” lasting for about five weeks, which seems about right. There was a corresponding “crackdown” later in the season as well which lasted about two weeks.

When you look at the same chart, you’ll see that despite the threats of a set restart crackdown for teams consistently flaunting the rules, there was no such thing this season. Especially not around Round 10 when the penalty crackdown started.

The only real trend was the absence of set restart calls over the final three weeks of the season, culminating in the final round of the 2021 season having the fewest six again calls of any round since the rule’s inception, despite 2020 having half as many reasons to call one. It seems like there was some sort of unspoken rule to reduce the impact set restarts had over the final few weeks of the season leading into the finals.

Running and metres

Unsurprisingly with so many points being scored, and fewer play the balls, there was a drop in runs and running metres. After all you can’t run the ball when you’re watching a teammate or opponent take a conversion. Total run metres per game declined by 1.2% after increasing by nearly 3.8% in 2020, with this season seeing 2,880 metres run per game on average.

This came from 2% fewer runs in 2020, which you’ll note is a larger decline than for metres gained. Therefore obviously metres per run increased, and they did by 0.8% to 8.93 per run, up from 8.86 per run in 2020. Most of this was from post contact metres per run, which increased from 2.94 to 2.98 per run, while post contact metres gained as well from 5.92 to 5.95 per run. It’s only a handful of centimetres but over the course of 320 runs per game those numbers can add up.

One pass runs, which increased significantly in 2020 with the introduction of set restarts, dropped back to 2019 levels as a percentage of total runs.

This looks to have come at the expense of dummy half runs, which increased as a share of total runs for the first time in multiple seasons.

With set restarts being awarded for offside players as well as ruck infringements, it appears that players are darting out of dummy half more to try and catch players offside or out of position at the expense of the traditional and slower one out run. Whatever the reason, the tedious one out running looks like it was quarantined to 2020. If only that was the case with certain other things.

Another running change this season was the change in positions taking one pass runs. The below table shows the average number of one pass runs by position for the last two seasons.

Middle forwards, interchanges and wingers are still doing the lions share of the work, but they’ve all seen fewer hit ups per game this season. The one position that has seen any meaningful uplift in one pass runs per game this season was centres, whose average moved from 20.5 to 21.3 per game.

As theorized earlier in the season, with players tiring earlier in games due to fewer stoppages and a faster pace, interchange players are spending more time on field due to starting middles not being able to play at the same rate. This also explains centres needing to run the ball more as a break for traditional hit up positions who need a break. Depth across the park mattered this year, and those who recruited big bodies with poor lateral quickness or agility suffered.

Tackle busts was the big change for running statistics this season, with nearly 18% more tackle busts recorded in 2021. That’s an extra nine per game, up from 51.5 to 60.7. Most of them were probably from Tom Trbojevic alone…

Passing & errors

If 2020 was the year one pass hit up, then 2021 was the year that spreading the ball and second phase play returned to rugby league. General play passes (not including passes from dummy half), grew by 3% this season, and offloads also increased by 4% after declining by 5% in 2020.

The safe, one out style of 2020 has been buried (at least by the top teams), and now the meta is exploiting tired or limited defenders with ball movement. The one surprising thing about the increase in ball movement this season is that errors dropped by 3%, something you wouldn’t expect with more players looking to offload.

The intriguing thing about the extra passing and attacking more is that it didn’t result in players being tackled any wider. Below are the heat maps for players tackled on tackles 1-5 for 2020 and 2021. You can see some slight variations from year to year but no significant changes, and even at a team level there’s not too many differences.

There’s a thing about errors though, which is pretty important if you’re basing your strategy around holding the ball or generating empty possessions.

The part of the field with the highest intensity of errors occur between 15 metres and an opponent’s goal line, which means if you’re reducing your errors there’s a chance you’re also reducing your chances of scoring. There’s no reward without risk.

Kicking

In 2020 we saw kicks dead decline by 6%, yet in 2021 they returned to their former state with an increase of 6%. Last season was the peak since 2014 with 41.5 kicks per game, with the 6% decline bringing the average back under 40 to 39.7 per game in 2021.

Total kicks dropped by 4%, interestingly again wiping out the increase of 4% in 2020. Long kicks, which had an 11% jump in 2020 were flat in 2021, with a 0% change. With teams making metres up the field due to restarts and tired defenses, kick metres also declined by 2%.

What didn’t change was the continued decline of weighted kicks, which decreased by 26% in 2020 and another 25% in 2021. The result of this was a continued decline in forced dropouts, which fell by 1% in 2020 and another 10% this season to just 2.9 per game, down from a peak of nearly 6 per contest in 2017. It’s clear as it was last season that attacking kicks aimed inside the in-goal area are a relic of the past.

Defense

If you’ve been paying attention, the fact there’s been fewer play the balls and runs would naturally mean fewer completed tackles. And you’d be right, with a drop in tackles made of 3% after an increase of 3% in 2020. There were still 600+ tackles made per game this season, down from the high of 630 per game in 2021, which was only the second time since 2014 there was more than 600 tackles completed on average per game.

What did change though is the massive increase in missed tackles, which was evident before with the huge jump in tackle breaks. Missed tackles skyrocketed by 18% in 2021, resulting in a big drop in Tackle Efficiency of 1.6% to 90.93%. That’s a big change, which hasn’t dropped below 91.4% since 2014 and had been above 92.4% in the past two seasons.

And as we saw with scoring, there was a big increase in points, tries and line breaks. To be expected there were double digit increases in try causes (12%) and line break causes (13%).

The ridiculous outlier statistics from a very normal NRL 2021 season

The first round of finals games over the weekend provided three close finishes from the four games, a pretty good record for a season that would like to be forgotten by many fans. As enjoyable as those games were, we wouldn’t want recency bias to hide the fact that the 2021 NRL regular season was underwhelming to say the least and had a significant number of outlier statistics.

And I’m not even talking about the standard ones, like Melbourne nearly breaking the consecutive win steak record and nearly scoring the most points in a season, Cody Walker trying Johnathan Thurston’s record for try assists in a season, or Reuben Garrick scoring more points himself than the Canterbury Bulldogs did as a team.

I’m also not going to talk about set restarts (seriously), other than to state that the cumulative effect of all rule changes in the past 18 months can be seen by the sheer number of outlier statistics from multiple teams this season.

It’s not unusual to see one (or maybe two) teams hit a high mark for one statistic in a season. For multiple teams to hit multiple records in a single season is a sign that something has shifted too far.

I’m not holding my breath for the proposed end of season review, which states that fans will be heavily involved. A simple change of removing set restarts for being offside and re-introducing a second referee would improve things, or removal of set restarts completely if we’re to remain with one referee. But I digress.

The statistics below are also evidence of just how embarrassingly one sided the regular season competition was this year, with one or two completely dominant teams, a handful more who were very good but unlikely to prevail, and a huge number of also rans. On to the perfectly normal outliers.

The increasing number of blowouts and amount of garbage time

Earlier in the season the Eye Test examined at the minute-by-minute margins of games, and it wasn’t a pretty picture. The average half time margin was 11.4 points, with huge increases in minutes played with a margin of 19+ points. Looking at the end of season numbers aren’t much better, as you can see below.

Up until the 30th minute, the average margin is pretty consistent with prior seasons. The yellow line for 2021 diverges at about the 30th minute where the average margin is 8.4 points, then goes to 10.9 for half time, 12.9 by the 50th minute, 15.0 by the 60th minute, 17.3 by the 70th minute and 18.8 points by the final whistle. The average NRL game this season had a margin in at the end of the game of more than three converted tries.

One of the points mentioned in the previous analysis of this data was that for a product that heavily relies on television revenue just to exist at this level, the increasing disparity between teams occurring even earlier in games is a huge cause for concern. Round 16 was one of the worst outliers this season, where the minute-by-minute average margin beat out all other rounds since 2016.

From the 34th minute onwards, no other round in the past six seasons was close to the average margin of Round 16, 2021. It would have been even earlier but Round 10, 2010 had some large margins from the 25th-35th minute.

More alarmingly the percentage of garbage time (minutes with 19+ margin) increased by 43% this season from 12% to 18%.

Let that sink in for a minute. The share of time played this season with a result virtually assured increased by 43%. Again, not a great look for any upcoming broadcast deals. Interestingly though, the change didn’t come from minutes played where the score was tied, it was the same as 2020 (15.1%) but down on the high point of 2019.

Not only that, 10% of minutes were played with a margin at least six converted tries the difference in scores. Once again, let that sink in – at least eight minutes a game were played with a margin of 30 points or more.

When you put this number into context as a total it becomes even more stark just how one-sided regular season games were. Over 2,800 minutes were logged with a margin difference of at least 19 points.

It’s easy to see the huge jump in total garbage time minutes from this chart, an increase of 68% on 2020. Those 2,811 minutes equate to over 35 full games played with a gap between teams of 19 points or more. Those 35 games make up over 18% of the NRL’s 192 game regular season schedule, nearly one in five games. Hide this one from any prospective media bidders.

All very normal stuff.

For and against outliers

Another example of the statistical outliers we saw this season is looking at average for and against by teams from the NRL era (1998 onwards), which was examined a few weeks ago on the site. Below is the end of season update.

Melbourne (and to an extent South Sydney) stand out as anomalies for the 2021 season, both scoring more than 32 points per game, a feat that had only been accomplished once previously (Parramatta, 2001). Melbourne have the best per game average in points scored, with Souths second since 1998.

Manly, despite a horrid start to the season still hold the fifth best per game average for points scored. However if you look at just games from round six, they’re actually ahead of Melbourne, with the three highest per game scoring averages in NRL era history coming this season.

Perfectly normal stuff, I know.

The other outlier we can pick from this data is by looking at the same data but colour coding them based on making the top eight or not.

Newcastle stand out as well here, as one of just four teams to make the finals with a below average attack and defense, although one of those teams (Canberra 2002) were only admitted due to the Bulldogs salary cap breach. It also shows that Newcastle had one of the worst attacks of a team to make the top eight since 1998, scoring 17.8 points per game which only puts them ahead of the 2007 South Sydney team that scored 17 points per game.

And the Knights were comfortably seventh. The fact they gave the Eels a scare on Sunday probably says a lot about the Eels chances going forward.

Margin versus running metres

Another outlier this season was just how dominant Melbourne and Penrith were, beating teams by 20.7 and 16.3 points per game respectively. Those numbers were the two highest average margins of victory in the NRL since 2014, beating the previous best of 14.95 by Ivan Cleary’s Panthers in 2020.

What is surprising is that both sides achieved this without having the best season for average run metres gained in a season. The Panthers and Eels of 2020 gained more metres than Penrith or Melbourne this season.

On the other end of the scale, Canterbury’s Trent Barrett had one of the worst seasons since 2014 for not only margin, (-15.4 per game), but also run metres. The Bulldogs averaged nearly 300 metres fewer per game than the Panthers and 200 fewer than the Storm. It puts them in an illustrious group of teams, including the 2020 Anthony Seibold Broncos and the 2016 Nathan Brown coached Knights for futility.

It looks even worse for Barrett when you look at the last three seasons of net post contact metres (post contact metres gained minus post contact metres conceded). The Dogs won just two post contact metre battles all season, which you can see on the below chart.

Theygave up 166 more post contact metres per game than their opponents, by far the worst mark in the NRL in the past three seasons. Even the wretched 2020 Broncos coached by Seibold only gave up 20 more than their opponents.

Very normal indeed.

Dominance of field position

Not only were the top teams scoring more points and winning by record margins, but they were also restricting opponents field positions. This year there were four teams who held their opponents to more than 60% of play the balls in their own half, a feat that only occurred once since 2014 (the Panthers in 2020).

There was nearly a 5th team, with teams facing Newcastle playing the ball 59.9% of the time in their own half.

Very normal, I know.

The Tom Trbojevic dominance

Say what you will about his disappearance on field in Friday’s loss to Melbourne, but Manly’s Tom Trbojevic had one of the best regular seasons statistically in NRL history. Some of that is due to stat inflation due to more time in play, but it doesn’t downplay just how he carved through the competition.

If you look at the Eye Test’s Net Points Responsible For statistic from 2014 onwards, not only does Trbojevic have the best season on record, but two other players from 2021 sit inside the best five seasons for NPRF ever.

Just behind Turbo at +13.1 points per game is Nathan Cleary, whose +12.4 would have shattered his own prior best in 2020 if not for the Manly #1. Cody Walker (+9.3) takes fifth spot all time for his impressive season as well.

A number of other players from 2021 sit inside the top 20, with Jahrome Hughes, Brett Morris, Mitchell Moses and Daly Cherry-Evans having comparable seasons with Johnathan Thurston and Cooper Cronk for NPRF. Jarome Luai and Adam Reynolds also sat just outside the top 20.

Totally normal stuff from a very normal season, your 2021 NRL season.

The NRL 2021 Eye Test Advanced Statistic leaders

With the 2021 regular season has drawing to a close it means we can take a look at who were the Eye Test’s advanced statistical leaders for the season. If you’re not aware of these advanced stats, I’d recommend a look at the glossary page for a quick rundown, and there’s further reading on Tackle %, Run % and Involvement Rate as well.

Before we start, I want to explain the minimum thresholds for these advanced statistics. For the first three (Tackle %, Run % and Involvement Rate) the cut off is having played 25% of games which is six appearances this season (last season it was five from a 20 game season). In addition to six games, there’s also a minimum 200 minutes played threshold to remove any players who have played enough games but not enough minutes to generate a decent sample size on the field.

By combining the games and minutes it eliminates players who don’t appear enough for decent sample size like Manly’s rookie Kurt De Luis who played the required 6 games but only 125 minutes (20.8/game) and I’d like to see a bigger sample size. It also removes players who may have played 200 plus minutes but in 4 or 5 games and had a few big games skew their data.

In the end the cut offs are mostly arbitrary however I feel they do a good job of recognizing the type of player these statistics were created to showcase.

For Net Points Responsible For (NPRF) and Error Rate the 25% cut off remains, however there is no minute restriction. In fact, NPRF only has the six game requirement, as low minute players don’t tend to feature. For Error rate, the only other minimum qualification is committing at least six errors. Again this number is subjective but passes the eye test for most cases.

If you’re curious as to what makes a good Tackle %, Run % or Involvement Rate, the average rates are shown below. Generally, they favour interchange middle forwards, but you do occasionally see some hookers and backrowers in the charts, although not typically for a full season.

The leaders below are all significantly beating these averages, making them elite by these metrics. On to the 2021 leaders. 

Tackle %

Billy Magoulias of the Sharks takes top spot for Tackle Rate in 2021 at 34.12%, meaning he is completing a tackle on at least three out of every play the balls Cronulla faces whilst he is on field. Second place goes to Reuben Cotter of the Cowboys at 33.77%, who may have taken top spot if he hadn’t spent the majority of the season out with a severe foot injury.

Rounding out the top three is Eye Test Hall of Famer Daniel Alvaro from the Dragons, with his usual high output at 31.43%. Alvaro would have taken the 2020 crown but only played 154 minutes during the season.

Previous winners:

2020 – Jai Whibread, Gold Coast (39.88%)

2019 – Daniel Alvaro, Parramatta (34.17%)

2018 – Daniel Alvaro, Parramatta (37.43%)

2017 – Daniel Alvaro, Parramatta (38.04%)

2016 – Siliva Havili, St George Illawarra (34.66%)

2015 – Christian Welch, Melbourne (42.88%)

2014 – Tim Robinson, Cronulla (36.65%)

Run %

Panthers impact sub Spencer Leniu held on after blitzing this statistic at the start of the season, finishing with a run rate of 15.98%, and one of just four qualified players in the NRL who had a rate higher than 15%. A pair of South Sydney forwards picked up the second and third spots, with Mark Nicholls (15.71%) and Liam Knight (15.51%) following Leniu. The only other player above 15% was Manly’s Josh Aloai at 15.35%.

Previous winners:

2020 – Andrew Fifita, Cronulla (17.34%)

2019 – Corey Horsburgh, Canberra (16.09%)

2018 – Martin Taupau, Manly (16.29%)

2017 – Nathaniel Peteru, Gold Coast (18.77%)

2016 – Jeff Lima, Canberra (18.60%)

2015 – Paul Vaughan, Canberra (16.98%)

2014 – David Klemmer, Canterbury (18.49%)

Involvement Rate %

The number one player in the NRL this season for Involvement Rate was Canberra’s (by way of Belmore) Corey Horsburgh, who ended the season at 22.68%. This means he completed a tackle or run on more than one in five possessions during a game this season, one of just seventeen NRL players to have a rate higher than 20%.

Taking the runner up spot for a second time is Cotter, with an Involvement Rate of 22.37%. This place him ahead of third place in Tackle %, with Alvaro sitting at 22.30%. Last year’s winner, Jaimin Jolliffe followed up his impressive rookie season by playing 34 minutes a game in all 24 of the Titan’s contests and finished fifth for the season with a rate of 21.56%.

Previous winners:

2020 – Jaimin Jolliffe, Gold Coast (28.04%)

2019 – Daniel Alvaro, Parramatta (23.59%)

2018 – Daniel Alvaro, Parramatta (25.85%)

2017 – Daniel Alvaro, Parramatta (24.21%)

2016 – Siliva Havili, St George Illawarra (34.66%)

2015 – Christian Welch, Melbourne (25.27%)

2014 – Tim Robinson, Cronulla (24.07%)

Net Points Responsible For

No surprises here with Tom Trbojevic’s devastating season topping the ladder here. The Manly #1 was responsible for +196 net points for his side, averaging out at +13.1 per game in his 15 appearances. Considering the Sea Eagles were +252 themselves, you could argue that Trbojevic was responsible for about 78% of their difference.

Again no surprises that Nathan Cleary was second, who was having his own amazing season at +12.4 net points per game, but was overshadowed by just how dominant Trbojevic was. In another non-surprise, third place was taken by Souths’ five eighth Cody Walker, who ended up with 37 try assists for the season and a net +9.27 points added per game.

Trbojevic’s +13.1 smashes the previous best of +9.7 by Cleary in 2020 by over 3 points per game, and Cleary himself beat his own record by nearly +3 points per game as well. Walker’s season was no slouch either, and is fifth overall since 2014.

Previous winners:

2020 – Nathan Cleary, Penrith (+9.72)

2019 – Luke Keary, Sydney (+8.63)

2018 – Cody Walker, South Sydney (+6.50)

2017 – Cooper Cronk, Melbourne (+8.21)

2016 – Mitchell Pearce, Sydney (+9.67)

2015 – Johnathan Thurston, North Queensland (+7.62)

2014 – Johnathan Thurston, North Queensland (+7.82)

Error Rate

The award for worst hands this season has been won by Panthers outside back and future Canterbury Bulldog Brent Naden, who committed an error every 7.9 times he touched the ball.

Naden was followed by Newcastle winger Dominic Young, who committed an error every 8.3 possessions, with another Panther in Izack Tago rounding out the top three with an error rate of one every 8.5 times he handled the ball.

Roosters fans will not be shocked to see Matt Ikuvalu on this list, having committed 27 errors this season at a rate of one every 11.65 possessions, the worst of players with at least 200 touches this season. He barely beat out yet another Panther in Viliame Kikau who had a rate of one every 11.89 possessions.

Tigers rookie Zac Cini would have topped this list, as he committed nine errors at a rate of one every 6.2 touches, but he only played four games this season. That may be a reason why.

Previous winners:

2020 – Shane Wright, North Queensland (7.75 possessions/error)

2019 – Lindsay Collins, Sydney (7.89 possessions/error)

2018 – Will Matthews, Gold Coast (7.89 possessions/error)

2017 – Joe Wardle, Newcastle (9.33 possessions/error)

2016 – Corey Denniss, Newcastle (7.36 possessions/error)

2015 – Blake Ferguson, Sydney (8.57 possessions/error)

2014 – Kyle Feldt, North Queensland (7.06 possessions/error)

Are Manly the third legitimate grand final contender ? – NRL Round 24 2021 stats and trends

Last weekend’s fiery Souths/Roosters clash seemed to strike a line through the Rabbitohs chances as a grand final outsider with the lengthy suspension (but not send off) of Latrell Mitchell. Parramatta meanwhile performed one of the cruelest acts that could be inflicted on an already traumatised fanbase – they gave them hope.

Here at the Eye Test, we’d already struck a line through Eels for being perennial finals disappointments, Souths due to their defense, the latter of which I’ll expand on this later. Most teams had a line put through them by myself this season, and not just because they’ve had 50 points put on them. One team that hasn’t been completely eliminated from contention in my view (yet) is Manly.

Why Manly? I’ll get to that later but let’s first set the stage for how obscene this season has been for the top teams.

The first chart this week is a plot of average points scored and conceded for all NRL era teams (1998 onwards), with 2021 seasons highlighted in red. Credit to friend of the site Rugby League Analytics who had the idea for this chart first, showing results for English teams.

The chart is set up into four quadrants to group teams, defined by the average points scored in a game this season (21.4). The top left “bad defense/bad attack”, the top right “good attack/bad defense”, the bottom left “bad defense/good attack” and finally the bottom right showing “good defense/good attack”.

If you’re scoring above 21.4 per game and conceding less than that, you’re in the “good attack/good defense quadrant”. If you’re scoring more than 21.4, but also conceding more, then you’re in the “bad defense/good attack” quadrant. It’s a quick way of visualising what the basic strengths and weaknesses are for teams and seeing how they compare historically.

Looking at this year’s numbers, it’s painfully obvious just how ridiculous this season has been. Not only is Penrith historically great defensively, Souths have the second highest points for average since 1998 and Manly have the fifth. This doesn’t even count Melbourne averaging 2 points per game higher than the previous best mark for attack and sitting as an incredible outlier.

Those of you noticing where Manly is will be wondering why I haven’t ruled them out if I’ve ruled out Souths, who sit in a better position. The answer lies if I filter the same chart for Rounds 6-24, instead of the whole season, as seen below with the current season shown in red.

When looking post Round 6, Souths become an average defensive team despite putting a lot of points on the board. Last year I posted a similar piece showing that teams outside the “good defense/good attack” quadrant rarely make the grand final let alone win it, which is why I’d be putting a line through Souths after looking at this. On their best day they can put points on anyone, even without Mitchell. But they’ve shown this season that keeping teams from scoring hasn’t been their forte, which will become a larger issue in September.

Secondly, it’s no surprise that Melbourne are again a staggering outlier, even more so than on the previous chart, outscoring opponents 3 to 1 during this period. I feel like sometimes we almost take it for granted how dominant Melbourne are, this chart does a great job at highlighting just how far ahead of everyone else they are.

And then there’s Manly. For as dominant as Melbourne has been, the Sea Eagles are right up there over this period. It’s been obvious how easily the Sea Eagles have been scoring points Round 6 when Tom Trbojevic returned and other changes were made (Haumole Olakau’atu playing his first game of the season for example), but not a lot of focus has been on how difficult they’ve been to score on.

From Round 6, Manly has scored 36.2 points per game but only conceded 16.9, third in the competition behind the Storm and Panthers. It’s incredible to see the turnaround one player can make, but it would be disrespectful to ignore the job Des Hasler has done with what many considered a team lacking talent.

It’s an indication (from my perspective at least) that it took Hasler just over a month that the new “normal” in the NRL required some changes to his game plan, something that other coaches seem reluctant to admit.

Using one of my favourite charts, If you compare margin versus net set restarts (awarded minus conceded) over the same rounds, you’ll see that Melbourne and Manly are playing the same game and everyone else is lagging behind.

Another thing to note from the previous chart is that there’s no team sitting in the “good defense/bad attack” quadrant this season and three teams in the “bad defense/good attack” quadrant. One of the teams in that latter quadrant, the Wests Tigers, are having the third worst defensive season of any team scoring above league average, only behind the 2003/2004 Manly sides.

I know what you’re thinking – all perfectly normal.

It’s still hard to talk myself into anyone other than Melbourne or Penrith playing on the last Sunday in whatever month the grand final is played in due to Covid. But if I had to put money on one other team to be there it would be the Sea Eagles, as they’re the only other team that seems to understand just how much rugby league has changed this season.

Having the most impactful player in the game on your team doesn’t hurt either.

The worst defensive top 8 team ever?

Another way of segmenting the above for and against data is by highlighting the teams that make the top 8. The results are below, and for Newcastle fans it’s not pretty.

The Knights will join an illustrious group after this round, being just the fourth team in the NRL era to make the top eight from the dreaded “bad defense/bad attack” quadrant. They’ll join the 2002 Raiders, 2009 Broncos and 2008 Warriors once this round concludes, as Newcastle’s average scored (17.65) and average conceded (23.3) can’t be improved enough to jump into a better quadrant.

Newcastle is also by far the worst defensive team (by average points conceded) to make the top 8 during the NRL era. And they’re sitting in comfortably in seventh place.

Again, all perfectly normal stuff.

The biggest edge black hole in the NRL?

Last season I did some analysis on which edge players were the biggest black holes. I looked at which second rowers and centres were running the ball the most and passing the least, to identify which players were tucking the ball and running rather than spreading it to their outside men.

The reason for this was watching Euan Aitken play, then at the Dragons, and his incredible ability to ignore his outside man and run the ball himself. And no surprises, he was the biggest black hole in the NRL last season.

Fast forward to 2021, and friend of the site Jason Oliver of the essential Rugby League Writers noted that Aitken hadn’t thrown a pass in three weeks.

That might be more palatable as an edge backrower this season at the Warriors than it was as a centre at the Dragons previously, but it still starves the outside backs of opportunities to attack.

Given this observation, I thought it was a great time to revisit this and the results for 2021 are below. Data is from players starting a game as a second rower or centre, and not from those as an interchange player but playing there. Average number of runs is plotted against average number of general play passes, with number of possessions indicated by the size of the data point.

For the Aitken fans, you’ll be pleased to hear he’s not the worst offender this season. He’s still not passing the ball much but he’s not running it as much either. The largest black hole honour in 2021 would go to the Sharks Teig Wilton, who averaged 16.4 runs per game and just 1.4 general play passes in his eight games in that position this season. He’s slightly ahead of the improved Isaiah Papali’i from the Eels, who runs the ball 16.2 times per game but passes nearly twice as often as Wilton at 2.6 per match.

The Bulldogs Aaron Schoupp is by far the worst offender at centre, averaging 15.3 runs per game but only 1.2 general play passes. Cronulla’s Jesse Ramien has a similar number of runs but passes the ball nearly 4 times per game.

On the other end of the scale, the most prolific passers this season from an edge player have been Josh Schuster (6.1 general play passes with 8.9 runs per game) and Tevita Pangai at the Broncos (14.2 runs with 6.1 passes).

The equal fewest set restarts in NRL history – Round 23 2021 stats and trends

A week off was all I needed and I’m back on my bullshit. Or is that the NRL’s bullshit? Stay in your lane, kids.

Round 23 saw the equal fewest set restarts awarded in NRL history, going back to the rule being implemented from Round 3, 2020. Just 46 set restarts were awarded, at an average of 5.75, beating the previous (full round) low of 47 in Round 3 which had an average of 5.88. Below is the chart for average penalties and set restarts awarded by round this season.

You will note I said equal fewest, but the round that shares that honour wasn’t mentioned. That’s because it occurred in Round 4, 2020, which was the last round before coaches realised the new rule could be gamed to their advantage.

Those three rounds (4 in 2020, and 3 and 23 in 2021) are the only rounds thus far to have an average number of restarts awarded fall under six. You can see below the average of set restarts called per round in 2020 (top row) and 2021 (bottom row), with the three rounds with an average under six sticking out. You can also see the previous 5.8 average in Round 4, 2020 right before the average nearly doubled to 9.4 and then 10.4 in Rounds 5 and 6 of that season.

The fact it’s equal with Round 4 from last season is ridiculous, considering we have six again now being called for being inside 10 metres at the play the ball. This would require me to believe that teams are infringing in the ruck less, which the Eye Test will tell you is a tall tale.

We’ve mentioned previously how set restarts are inconsistently called between halves, rarely called late in games and if so only to losing teams, are only awarding 2.3 extra tackles and are called mostly in teams own halves. The Sydney Morning Herald had a recent story showing data that teams leading don’t win set restart counts.

I can believe that they’re inconsistently applied and there may be some unconcious bias towards losing sides. What I don’t believe is that players are interfereing in the ruck at a lower rate than they were twelve months ago. If anything it’s gotten signifcantly worse.

I’m also not going to draw conclusions on the integrity of referees. Despite being a set restart truther, I feel that the referees are (largely) doing a good job but are hamstrung by some poorly implemented and untested rules that were thrust upon them alongside a 50% reduction in their workforce. And they’re regularly thrown under the bus but their own administration. Given the hand they’ve been dealt they’re doing a great job undermined by some inconsistent interpretations.

The other trend that’s appeared over the past few weeks, which matches up with last round’s record low, is the slow decline of restarts being called over the past month. The average has dropped from 8.4 per game in Round 19 to 7.4 in Round 22 and 5.8 in Round 23.

Still, it’s an interesting trend as we progress towards the pointy end of the season. A more cynical person than myself might suggest that they’re deliberately being downplayed to prevent any drama in the finals. I’d like to see another week or two of declining restarts before saying it’s a larger trend, but I’ll certainly be keeping watch on this moving forward.

One thing we can take solace in is that the six again king has returned. Thanks to Grant Atkins calling just three restarts in the Souths v Penrith match on Friday evening, Gee has now pushed ahead of Atkins for the highest average of set restarts awarded among NRL referees this season. The full chart is below.

Tackle busts and long run analysis

Earlier in the season there was a near regular section on the Eye Test every week talking about how great Brian To’o is. It’s been a while since we looked at those charts, and with the season winding down and To’o on the sidelines it’s a good time to check back and see if he ended up dominating the competition as it looked like he would back in May.

First up is average runs plotted against average tackle busts. This chart is a great way to identify high volume runners of the ball (further to the top of the chart), impact runners (further to the right), or players who do both (the top right). The size of the data point indicates the amount of run metres, the larger the circle the more metres gained over the season.

It’s evident here just how dominant To’o was and it’s shame his (regular) season performances were curtailed by a syndesmosis injury a few weeks back. At 22.4 runs per game, he’s at least 3 runs per game higher than his nearest rival, Roger Tuivasa-Sheck. To’o’s 6.13 tackle busts per game place only behind David Fifita, James Tedesco and Tom Trbojevic. That’s incredibly elite company.

It’s also worth noting that Latrell Mitchell breaks just one half tackles fewer than To’o, but does so on more than half as many runs. This is what’s known as picking your spots, and when Latrell picks one, he usually makes sure it counts. Kotoni Staggs isn’t far behind Mitchell, but his 5 tackle busts per game from 7.5 runs occurred in only four starts, which is bordering on small sample size territory.

Moving on to the average number of long runs (more than 8 metres) v average number short runs (less than 8 metres) this season and To’o is still streets ahead. Performances for this are split between interchange and starting spots, which is why David Klemmer appears twice.

75% of his 22 runs per game pass the eight metre mark, which places him inside the top 30 among all NRL players this season. None of them complete the same volume of runs as To’o though, the closest being Payne Haas, at 82% of his 15.6 runs per game passing 8 metres.

The other big story from this chart is just how much of a down season Jason Taumalolo is having. This chart used to be owned by the Cowboys backrower, who averaged 15+ runs per game over eight metres in 2020. This year that has dropped by almost a third, down to 10.8 per game, with his sub eight metre run average nearly identical.

Regular Eye Test readers will remember Marcelo Montoya being possibly the least effective runner in the NRL last season. This year he’s improved to 45% of his runs over 8 metres, up from 41% in 2020. He’s not even the worst outlier anymore, that crown has been taken by Bulldogs rookie Falakiku Manu, who passes eight metres on just 32% of his runs.

Is Greg Marzhew in for a huge 2022?

One last name to note on the previous runs v tackle bust chart is the Titans Greg Marzhew. His numbers sit near the top at 17.6 runs per game and 3.9 tackle busts in his 8 appearances this season. Why am I bringing this up?

Because if you look at the same chart for 2020, his numbers aren’t that different from To’o, as seen below.

To’o averaged 15.9 runs and 5.0 tackle busts in 2020, which is about 2 runs and 1 tackle bust fewer than Marzhew this season.

You could also bring up their similar physical dimensions, but the similarity in numbers doesn’t stop there, because when you look at the percentile ranks of To’o for 2020 compared with Marzhew for 2021 compared to all outside backs from 2014-2021, the similarities continue.

Their radar chart shows a very similar profile in per game averages for runs, metres, tries, line breaks, tackle busts and one pass runs. In 2020, To’o sat above the 90th percentile for runs, run metres, metres per run, tackle busts and one pass runs. Even their try assists and line break assists sit in similar areas, which would be expected playing on the wing.

Marzhew in 2021 sits above the 90th percentile for runs, run metres, tries, tackle busts and one pass runs. He’s actually in the 98th during that time span for percentile for runs, run metres, one pass runs, which sits him firmly among the game’s elite.

There’s one massive difference between them though and that’s defensively, with Marzhew  which probably should be expected. To’o has been an integral part of one of the best defences in the NRL over the past two seasons, while Marzhew sits on the outside of a team that has given up some very large scores of late. As always there’s context needed for every statistic and these try and line break causes are usually more a factor of team defense than necessarily any individual deficiency. They’re indicative rather than representative.

It’s unlikely that Marzhew has a break out season in 2022 like To’o is having this year, especially with the chasm of talent between the Panthers and Titans. Although if you look at with that gap in ability in mind, maybe Marzhew is actually having an even better year than previously thought? Either way, if he continues his current run of form he may end up as exciting to watch in 2022 as To’o has been this season.