Tag Archives: statistics

It’s not just the six again

The move to do away with the ‘unlimited’ tackle rule that had been played in Australia for 59 seasons was the brainchild of the urbane secretary of the Rugby Football League in England, Bill Fallowfield. The move was designed to counter what Jack McNamara in the Manchester Evening News described as the ‘evils of the almost never-ending possession’. In both countries, a grinding, physical style of play had developed, with teams holding possession for long periods. Australian centre Bob Hagan… told of a game in which Huddersfield kicked off against Hull Kingston Rovers, and then touched the ball only twice in the first half. Hagan reckons it was this game that killed off the old rule.

Centenary of Rugby League (2008), Ian Heads and David Middleton

In the late 1950s, rugby league faced a problem. It was boring. Teams could hold the ball indefinitely and, using the favoured tactic of one out hit-ups, could maintain possession provided they didn’t make an error or do something silly like scoring.

The problem identified, a radical solution was borrowed from American football to introduce a ‘use it or lose it’ ethos to the game. Initially limiting possession to four tackles, and extended to six in 1971 in New South Wales (and later elsewhere), the new rule was trialled in pre-season competitions to test its impacts. The new rule had its detractors and the style of football it spawned was chaotic – dubbed “panic football” – but better than it had been. It forced teams to attack. Four tackles didn’t seem to provide enough time, so six tackles became the solution. The game was better for it.

The NRL website has every Sydney grand final from 1966 onwards and if you watch them like I have, you can see the evolution from unlimited possession to four tackle and then six tackles in the space of a few hours.

The key thing here is the process. A problem is identified. A solution is proposed. The solution is tested and evaluated outside of the main premiership. If successful at resolving the initial problem, the solution is implemented. The solution is adjusted as required in response to feedback

This is the basic framework of common sense decision making. Further, it is evidence that the sport of rugby league collectively and consciously decided it was not a game that valued possession of a football but one that valued attacking play and, as perhaps an unintended consequence, became a game of field position

Rugby league has a history of making these rule changes – introducing the play the ball and reducing from 15 to 13 a side in 1906, reducing the points value of goals in 1897 and increasing the points value of tries in 1983, introducing and then gradually increasing the offside rule over the years – to support the scoring of tries, considered the most interesting part of the game.

As it currently stands, the set restart will not be joining that pantheon of innovative rule changes.

The governing bodies for Queensland, NSW, French, PNG and British rugby leagues, as well as the international board, have all joined the Australian body in adopting the set restart. The rule sets still aren’t completely harmonised – the two point field goal remains an Australian-only feautre as far as I can tell, scrums are used at fewer points (or not at all) to restart play in England and the rules are different again for the women’s game for some reason – but largely everyone is now on the same page after Peter V’Landys, Project Apollo and the ARLC unilaterally changed the sport of rugby league during the 2020 covid off-season.

As more time passes, and the NRL bogs down into a mire of repeated blowout scorelines, it becomes clear what the sport has decided to sign itself up for. In a sense, it continues the tradition of attacking play but it appears to be decidedly one way. Concerningly, even as fewer six agains are called, the blowouts remain.

Worse still, the sport seems to have regressed more than half a century to being a matter of possession. Phil Lutton put together a fantastic piece in the SMH on the rule changes that, somewhat novelly, involved speaking to the actual players to see what they thought was happening, instead of merely regurgitating talking points from the administration. I believe this is called journalism.

The truly fascinating thing about is that these well-paid professional athletes cannot explain precisely what’s happening. This is not a reflection on them but rather reflects the complexity of the problem the NRL faces.

We can see the obvious. There are a lot of games decided in 2020 and moreso in 2021 by larger scorelines than we are used to. We can watch the games and see that if one team has a noticeable advantage after ten, fifteen, twenty minutes, then it won’t be long before they race out to a twenty point lead and it’s game by half time. The second forty minutes is effectively irrelevant. I’ve written about the subltety of the impact on the game before but do you really need any more evidence than this?

Previously, I compared the impacts of the rule changes to that of climate change. It’s small, it’s consistent and there are other things hapenning but it’s there. Some people’s complete inability to parse this reality – that several things can happen simultaneously to affect an outcome – leaves me baffled on a regular basis, however, if a substantial population seem to have an unbending love of simple, monocausal explanations for the complexities of the world, that does seem to explain much of history.

Not every game is a blowout, just as each day is not necessarily hotter than the last, but the data paints a picture of the overall situation that is as alarming as it is obvious. To repeat bullshit talking points about development, pathways, roster management and whatever else is insulting to the collective intelligence of the NRL fanbase. If anything, the complete absence of any thorough explanation of how the sport’s mechanics actually work on the field belies the idea that rugby league is a simple game for simple people (a subscription to Rugby League Writers will dispell that notion for you) and yet simple people insist on talking as if it were so.

Confounding this further is the abject refusal of the same blowouts afflicting the NRL to turn up in Super League, which adopted the six again after their own covid break in 2020, or either State Cup, which implemented the six again at the start of this season.

To unpack this, we’re going to have to go back to first principles.

*****

I wrote about how this all came about in June of 2020 and managed to reasonably well predict most of what’s happened since. I was wrong about the impact on margins (that only became clear later) but otherwise, the cliff notes follow.

The suspension of the 2020 season due to the coronavirus pandemic put the NRL in breach of contract with its free-to-air broadcast partner, Nine. Nine, a relic in a dying industry, felt that it had overpaid for the rights to the NRL and Origin and used this opportunity to make the NRL take a significant cut to its broadcast revenue while forcing the NRL to bring in changes that would make the game more entertaining, thereby increasing Nine’s ratings. Australian netball did somethinig similar at the behest of Nine.

The NRL introduced packages of rule changes between the 2019 and 2020 seasons, between the first two rounds and remainder of the 2020 season and then another between the 2020 and 2021 seasons. These changes tend to all be lumped together and include the captain’s challenge, reducing from two to one referee, the set restart for ruck infringements and later offside, a reduction in scrums but greater flexibility in how they are used, the 20/40 and the two point field goal for attempts beyond the 40m line.

The changes made under the V’Landys regime, most of the above, were sold on the basis that it would improve the pace of the game, the flow of the game, bring back fatigue to the game, which would allow the little man to flourish. All of this was deemed to be more entertaining, taking us back to a time of purer, less robotic football.

When the changes were first premiered, on 28 May as Parramatta defeated Brisbane 34-6, it was clear that panic football had returned after lying dormant for half a century. As the weeks progressed, it was not at all clear to me that the product was better but there was definitely more of it and I was told by the media, breathlessly and relentlessly, that it was better.

Ratings for the first two weeks of the resumed competition were record breaking. The dogshit defence of the Broncos was seen by more than 1.3 million people. But after a few weeks, ratings fell back into their usual rhythm and while the season seemed to finish with improved overall ratings, it was on fewer games and the big four matches of the grand final and State of Origin were well down on previous years.

This season, we’ve seen Origin bounce back a little, leading to suggestions that its ratings are up, which is true as long as you don’t look at what the ratings were five years ago. For the regular season, ratings seem to be down but not significantly enough that you wouldn’t be able to point to streaming and make up some stuff about historically bad teams to explain it. If the on-field product is leading to people turning off their TVs, it hasn’t been significant enough that the NRL won’t be able to duck and weave taking any responsibility for what’s happening.

All of this leads us to the inevitable question: what was the point?

I understand that one must occassionally suffer in the short term, in order to gain in the long term. The penalty crackdown in 2018 was a perfect example. In principle, we simply had to ensure a few months of penalty-ridden games to ensure that players and coaches understood that the “wrestle” would not be tolerated and the game would then speed up. Players and coaches knew that the administration did not have the stomach for it, nor for the fight in the media, and would buckle eventually. Sure enough, the administration did and we watched a lot of penalties get awarded for nothing.

This is the opposite. If there isn’t a clear goal to be achieved, as I suspect the current administration does not know what they’re doing or why they’re doing it, then why do we need to suffer? None of the rules supporters seem to be able to answer that question, preferring to deflect onto a Victorian bootstraps philosophy that every team simply must do better.

Ratings are, at best, flat and, at worst, down. That rather suggests that the rules haven’t had their intended impact. If the rules don’t serve their ostensible purpose – to entertain and to increase the audience – then there shouldn’t be anything that stops the NRL from rolling them back. There seems little point in “tweaking” the rule changes to mitigate the worst outcomes because the game is less entertaining and no more popular than it was previously.

As it stands, we have had to sit through a lot of terrible football for no gain.

The problem is how to unravel what’s been done. I don’t think it is as simple as getting rid of the set restart.

*****

I’ve got to say, even with the blow-outs the games are entertaining.

Before they were robotic, they were predictable. Now, they’re entertaining. Even the blowouts over the weekend were entertaining. For the viewer.

Peter V’Landys, 17 May 2021 (Poor roster management, not faster game, to blame for growing number of blowouts: V’landys, WWOS)

Teams that have adapted well to the new game don’t seem to mind where they start their sets. Using fast play-the-balls, a reliance on metre-eating backs gaining ground early in the set and a narrow passing game to add just enough variation to keep the defence guessing, they are able to keep their attacking line moving fast enough to regularly cover the best part of a length of the field in a set. The threat of the six again is enough to keep poorly organised defensive lines scrambling, unwilling to risk sitting in the ruck too long and extending the time that they have to defend. Ironically, this plays exactly into the better teams’ hands

Should the weaker team survive the set and regain possession, the well adapted teams are flying off the line and pinning down their opponents. This increases pressure on the team with the ball, forcing them to accept a paltry gain on their set or forcing them into an error, either from an ill-advised pass or from the sheer impact of defensive line.

Now the better team has field position and the ball. From there, any team with a competent halfback should be able to string together two or three repeat sets. Failing that, they can rely on their defensive linespeed to crush the opposition until they have had enough attacking opportunities to put points on the board. Then they get the ball back from kick-off.

And that’s it. Fifteen minutes of this and most teams crack – good and bad. Players do not have the aerobic engines to compete at that intensity for long and the poorer teams do not have the defensive structures to resist. Once the players are gassed, it’s trivial for the team with the upper hand to start running through and over teams no longer able to organise themselves or make tackles. Once they’re up by twenty, it’s game over but unfortunately, there’s often up to an hour still to go.

It gets worse later in the game when the fatigue causes handling errors, turning the ball back over and resulting in more energy-sapping defence, leading to a negative feedback loop whose destination is a blown out scoreline. On the rare occassions where the losing team manages to string a set or two together, they are too fatigued to run with any intensity. It becomes laughably easy to defend their insipid attacks. In desperation and running on empty, their fifth tackle options fall apart as players de-sychronise their timing, lose cohesion and begin to rely on individuals going it alone.

In short, once you are on the backfoot, you start to play a lot like the 2020 Brisbane Broncos. More often than not, you lose like them too.

What you may notice is that nowhere in that platonic ideal of the NRL in 2021 was the team with the upper hand awarded a set restart. Indeed, it was at most the threat of a set restart that got defensive teams scrambling. This, I think, explains why Penrith can have a positive set restart difference and Melbourne can have a negative difference and both can completely dominate the competition.

You might then be wondering what the difference is between 2019 and 2021. Isn’t this just a description of good rugby league gameplay? The answer is yes but also no. The differences are subtle and mostly rooted in the rule changes that have been brought in.

Relieving penalties are a thing of the past, replaced by the set restart. Bad teams relied on these penalties being awarded, sometimes seemingly at random, as a means to get a lift down the field, an opportunity take a breath and reset their organisation or to score an easy two points with another possession to follow.

This shouldn’t be seen as a bad thing. It’s how the game has been played for more than a hundred years. Importantly, it interrupts the otherwise continuous possession of the better team and offers a foothold for the weaker team to work their way back into the game. Otherwise, the better team knows it has the ability to move the ball as required to score points, they simply need to hold the ball until the lesser team cracks. As Phil Lutton put it:

This style of play, with its heavy emphasis on holding possession in lieu of gaining field position, wasn’t possible previously because the conditions that allow it to exist weren’t in place.

The reduction from two on-field referees to one has had huge and completely unexplored implications. Most of the discourse at the time of the change was about working rights. The ARLC gave some token concessions and since then, the referees have shut up and gotten on with it. The problem is that between halving the number of referees and the increased pace of play, the referees are now worse at their jobs.

This is the genesis of the debacle of the high shot crackdown. High shots that were obvious enough on TV were being missed on field because the referees, like the players, were gassed and looking for too many things. That crackdown, like all of its previous editions, was quietly shelved when it turned out it was poorly thought through and ruining the spectacle of the game even moreso than the blowouts. The solution seems to have been more Bunker involvement, which I’m fairly certain was decried by segments of the media a few years ago but goes unremarked in V’Landys’ NRL. If they haven’t already, the referees are going to reach overload.

Under a two referee system, one referee set the line and the other policed the ruck. Under a one referee system, one referee does both jobs but neither of them well. The better teams are able to get off the mark a few tenths of a second earlier because the referee’s attention is elsewhere and that’s enough to get to the opposition slightly sooner and hit slightly harder. The cumulative effect starts to hurt after a while.

Meanwhile, on the other side of the ball, teams like the Panthers can simply lie in the ruck, either waiting for a signal from the referee or their teammates to get up. If the wrestle has been eliminated, it’s been replaced with an even more blatant ruck infringement, largely daring the referees to blow a penalty that will never come. Instead, a lesser punishment in the form of a set restart may come, and the Panthers will simply shrug their shoulders and set their line, confident in the knowledge that their defence can withstand the insipid, exhausted attack of their opposition.

Considering all of this, I believe any attempts to tweak the rules, to apply another band-aid, are misguided. The suggestions to date will not disrupt this paradigm. One common one is to re-adopt the Super League rule and let the scoring team kick off. This would prevent the scoring team from regaining possession immediately after scoring but merely delays the inevitable. If the better teams can move the ball 70 metres downfield in a set, more if they get a set restart, and can manipulate the other team to dominate possession for an extended period, then we are back to square one.

Giving penalties in your own half and giving set restarts while in the attacking half doesn’t help either. Penalty goals are a key means for lesser teams with weak attacks to keep within touching distance of their opposition. By effectively eliminating the penalty goal from the game, you force weaker teams to attack the line. Penrith, who currently have the best defensive record of any Australian club of the last twenty years, would simply laugh.

In a season with a record number of shutouts, this isn’t going to redress the balance. Weaker teams need tools to keep their opposition within reach and to be able to work themselves back into the game, as they did 1895 through 2019. Without that, the blowouts will continue.

While we’ve examined one factor in detail but there are many at play and it would take a thesis to unpack everything. Some of the teams currently playing are simply bad at football and could probably stand to improve. But the teams on the wrong end of pastings have included the pre-season favourites and the club that won two premierships in 2018 and 2019. At the other end of the ladder, the Bulldogs have already won more games than the 2016 Knights and aren’t conceding anywhere near as many points as the 1999 Magpies. There’s always been bad teams but there’s never been scorelines like this. Not against good and bad teams. Not in a wet La Niña year. Not in a salary capped, full-time professional league. Not when it’s been fourteen years since the last expansion team entered the competition.

However, like breaking the four minute mile, the ceiling of what’s possible in rugby league has been raised and, even if the old rule set was reinstated, elements of the new style of play would remain. Unless the game is slowed down, it may well remain a game of possession. While I’m normally in favour of pushing the barriers of what’s humanly possible, we can only speculate as to where this might lead the sport over the coming years.

*****

So why don’t we see these blowouts in other leagues?

Despite the margin heading higher but not exceeding previous year, there’s signs in the English game. The closest game in round 12 was a forfeit. Round 8 had the fourth highest margin in Super League history. St Helens are currently conceding points at just 7.6 points per game, better even than Penrith and the best in Europe since Wigan in the 1986-87 season. The league leading attack of Warrington is less impressive, merely being the best since the 2017 Tigers. Salford’s attack has only been good for 11.5 points per game, the least since 2-23 Swinton in 1991-92. Leigh’s 0-12 record speaks for itself and their 39.1 points conceded per game is the seventh worst in the entire history of English rugby league dating back to 1895, sitting behind six Super League teams who managed to combine for 15 wins and 2 draws from 155 matches.

In Queensland Cup, Wynnum’s 34.0 points scored per game and 18.2 points conceded per game are only 16th and 52nd best marks in that competition. The Capras’ 16.1 points scored per game and the Cutters’ 33.6 points conceded per game are 26th and 27th worst, respectively. In all, it seems rather balanced. Despite this, the two leading teams each have more competition points than the bottom five combined, which includes three unaffiliated clubs.

It may just be that not enough set restarts have been called for it to be a sufficiently significant threat to fluster teams even in the absence of the six again call.

But we see a similar decline in penalties awarded.

Albeit, in percentage terms, it’s less significant in state cups than the NRL. Comparing the average penalties across 2016 to 2019 to the number called in 2021, 52% of penalties have been eliminated in NRL, compared to only 33% in Queensland Cup and 44% in NSW Cup.

It may be that in these leagues, the players aren’t athletic or skilled or coached well enough to implement the strategies used by Melbourne and Penrith in the NRL. Most of the NRL – and the Queensland Maroons, for that matter – haven’t come to grips with it yet, so it’s questionable whether we would have expected reserve graders to have mastered the new game. It may be that the spread of talent across the leagues is greater than the NRL and the effects are concealed within the typically higher scoring. It could be that Super League never adopted the second referee and state cup did so only sporadically and so the “change” back to one ref has had comparatively less impact on those competitions.

It may, as in the NRL and Super League in 2020, take time for the changes to fully percolate through. It will be worth watching other leagues to see how they adapt to the play on a year’s delay. More data should help isolate what’s happening and prove or disprove any hypotheses.

It is, as I said, complicated.

*****

Rugby league is an entirely artificial space. There is no natural order. We can decide what we want to see and there is no reason why would should have implemented these rules. They do not serve their intended purpose and the second order impacts, unknown at the time of implementation due to a lack of trialling, have made the NRL demonstrably worse.

Any defence of the rule changes starts with the presumption that they need to remain and only require modification but this flies in the face of the common sense decision making framework used previous administrators to improve the game. We must ask ourselves what purpose the changes were meant to serve, if that’s been accomplished and if not and what reasoning there is to continue with it. If the goal was to punish teams for losing, then it’s mission accomplished but if the goal was to entertain, it’s been sadly lacking.

The concern is that in a World Cup year, assuming it goes ahead at all, will see minnows will face off against Australian and New Zealander sides principally comprising players from Penrith and Melbourne. If full time professionals get blown off the park by these players, what hope is there for nations whose teams are made up of part timers and amateurs? It could get ugly and it will be embarrassing and it will reflect the shortage of common sense decision making at all levels of rugby league adminstration.

But until someone smarter than me works out what’s really going on, the best we can hope is the World Cup is refereed as the lower tiers of the sport are and hope that the next round of band-aid solutions, rumoured to already be in the works, somehow fixes the problem by accident.

2021 Super League WIP Report

We’re about a third of the way through the Super League season (maybe? Depending on covid and weather, I suppose) and it’s time to review how each team and player are performing. We’ll be looking at this using the following analytical tools:

  • FormElo ratings that reflect short term performance.
  • Production – The accumulation of valuable work on field, as measured by statistics that correlate with winning. Production is measured in Rismans.
  • Disappointment Line – The minimum number of wins for the season to not be considered disappointing by fans, as calculated by the pre-season class (long term performance) Elo rating.
  • 1st order winsPythagorean expectation calculated by points for and against.
  • 2nd order wins – Pythagorean expectation calculated by SCWP.

There’s more detail at How It All Works.

I won’t be providing much analysis for individual teams or players, instead preferring to let you do the work for your favourite team or player with the context about how the tools work provided.

WIP graphs

Comparing win percentage of the Disappointment Line, actual wins (0th), Pythagorean wins (1st) and 2nd order wins.

Note that the zero to one scale for winning percentage can make significant differences look minor; a .100 change in winning percentage is worth up to 2.5 wins at season’s end.

Some of the games have no or incomplete stats, so the second order stats might not move in line with the first order or winning percentage. Some teams haven’t played all of their games between round 1 and 11, in which case the lines do not change through this round. Statistics do not include Castleford’s forefit to St Helens.

Significant divergences between actual and Pythagorean wins are usually indicative of mean regression in future. That is, teams underperforming will improve their actual win percentage and vice versa so that the two numbers tend to converge. However, there’s always one or two that manage to avoid this convergence. We call them lucky/unlucky, depending on the direction that they miss and the ladder is currently a mess.

The outlook for each team:

  • Actual wins understating year to date performance / potentially tending better – St Helens, Leeds, Huddersfield, Leigh
  • Actual wins overstating year to date performance / potentially tending worse – Catalans, Wigan
  • Actual wins reflective of year to date performance – Warrington, Hull FC, Hull KR, Castleford, Wakefield Trinity, Salford

For teams with less than .500 percentage, 2nd order wins is always better than actual wins and vice versa. The trick here is to not focus on this specific value per se but to look at it as a prediction for next year’s performance. The next season’s performance for most teams will fall within one standard deviation (plus/minus .130) of their 2nd order win percentage for this season.

Form

Form Elo ratings at the end of 2020 and the end of round 11

1st Order Wins

Winning percentage estimated by Pythagorean expectation of actual points for and against

2nd Order Wins

Winning percentage estimated by Pythagorean expectation of SCWP for and against

Player Leaderboards

Rismans

Rismans are the unit of measure for the amount of valuable work done (production), equivalent to Taylors in Australia. Due to the state of the dataset, not all games and appearances will have been captured.

The career leader (2017 – now) for total Rismans is Jermaine McGillvary with 4402. The single season record holder is Jackson Hastings in 2019 with 1768. 2020 was topped by Bevan French with 1114.

Rismans per game

As above but averaged per game (minimum 5 games)

The career leader (2017 – now) for Rismans per game is Peta Hiku with 74.1 (minimum 10 games). The single season record holder is (minimum 5 games) Craig Hall in 2018 with 76.7. 2020 was topped by Ash Handley with 68.0.

2021 Queensland Cup WIP Report

We’re over half way into the Intrust Super Cup / Queensland Cup and it’s time to review how each team and player are performing. We’ll be looking at this using the following analytical tools:

  • FormElo ratings that reflect short term performance.
  • Production – The accumulation of valuable work on field, as measured by statistics that correlate with winning. Production is measured in Taylors.
  • Disappointment Line – The minimum number of wins for the season to not be considered disappointing by fans, as calculated by the pre-season class (long term performance) Elo rating.
  • 1st order winsPythagorean expectation calculated by points for and against.
  • 2nd order wins – Pythagorean expectation calculated by SCWP.

There’s more detail at How It All Works.

I won’t be providing much analysis for individual teams or players, instead preferring to let you do the work for your favourite team or player with the context about how the tools work provided.

WIP graphs

Comparing win percentage of the Disappointment Line, actual wins (0th), Pythagorean wins (1st) and 2nd order wins.

Note that the zero to one scale for winning percentage can make significant differences look minor; a .100 change in winning percentage is worth 1.9 wins at season’s end.

There’s no projections this year as we didn’t have any information from last season to work with.

Significant divergences between actual and Pythagorean wins are usually indicative of mean regression in future. That is, teams underperforming will improve their actual win percentage and vice versa so that the two numbers tend to converge. However, there’s always one or two that manage to avoid this convergence. We call them lucky/unlucky, depending on the direction that they miss and the ladder is currently a mess.

The outlook for each team:

  • Actual wins understating year to date performance / potentially tending better – Blackhawks, Tigers, Jets, Cutters, Capras
  • Actual wins overstating year to date performance / potentially tending worse – Wynnum, Devils, Tweed, Redcliffe, Pride
  • Actual wins reflective of year to date performance – Falcons, Bears, Magpies, Hunters

For teams with less than .500 percentage, 2nd order wins is always better than actual wins and vice versa. The trick here is to not focus on this specific value per se but to look at it as a prediction for next year’s performance. The next season’s performance for most teams will fall within one standard deviation (plus/minus .130) of their 2nd order win percentage for this season.

Form

Form Elo ratings at the end of 2020 and the end of round 11

1st Order Wins

Winning percentage estimated by Pythagorean expectation of actual points for and against

2nd Order Wins

Winning percentage estimated by Pythagorean expectation of SCWP for and against

Player Leaderboards

Wins Above Reserve Grade

WARG is a volume stat that compares the total amount of valuable work done (production), when compared to a replacement level player (or fringe first grader in the vernacular) at that position, irrespective of the time on field. A replacement level player has 0 WARG. It’s interesting to correlate who is the top of this leaderboard and who has been promoted to the NRL this season.

2019 was topped by Harry Grant with 2.6 WARG. The career leader (2016 – now) is Jonathon Reuben with 7.8 WARG. The single season record holder is Scott Drinkwater in 2018 with 3.0 WARG.

Taylor Player Rating

TPR is a rate stat that compares the amount of valuable work done (production) per game, factoring in time on field, to the average player at that position. An average player has a rating of .100. Minimum 5 games need to be played to qualify for TPR.

The 2019 regular season was topped by Harry Grant with single season record TPR of .266. The career (2016 – now, regular season only, minimum 15 games) leaders is also Harry Grant with a TPR of .195.

2021 WARG by position

WARG as generated when the player is listed in that position so far this season.

2021 NRL WIP Report

We’re just a bit over half way into the 2021 NRL season and it’s time to review how each team and player are performing. We’ll be looking at this using the following analytical tools:

  • FormElo ratings that reflect short term performance.
  • Production – The accumulation of valuable work on field, as measured by statistics that correlate with winning. Production is measured in Taylors.
  • Disappointment Line – The minimum number of wins for the season to not be considered disappointing by fans, as calculated by the pre-season class (long term performance) Elo rating.
  • 1st order winsPythagorean expectation calculated by points for and against.
  • 2nd order wins – Pythagorean expectation calculated by SCWP.

There’s more detail at How It All Works.

Unlike the season preview deep dive, I won’t be providing much analysis for individual teams or players, instead preferring to let you do the work for your favourite team or player with the context about how the tools work provided.

WIP graphs

Comparing win percentage of the Disappointment Line, pre-season projected wins, actual wins (0th), Pythagorean wins (1st) and 2nd order wins.

To be perfectly honest, I’m not 100% sure what these are meant to look like, having been the first time I’ve presented this information in this format. For example, I’m surprised that the pre-season projections and Disappointment Line are often pretty much the same because I’d never thought to check previously. Note that the zero to one scale for winning percentage can make significant differences look minor; a .100 change in winning percentage is worth 2.4 wins at season’s end.

The projections are naturally more conservative (less likely to predict outliers) to cover the potential spread of results. Think of, for example, Penrith as having 15.6 projected wins plus/minus 3 wins. While some projections are right, it tends to overshoot really bad teams and undershoot really good teams.

Significant divergences between actual and Pythagorean wins are usually indicative of mean regression in future. That is, teams underperforming will improve their actual win percentage and vice versa so that the two numbers tend to converge. However, there’s always one or two that manage to avoid this convergence. We call them lucky/unlucky, depending on the direction that they miss.

The current status of each team:

  • Way out – Cowboys, Rabbitohs
  • Actual wins understating year to date performance / potentially tending better – Titans, Panthers, Dragons,
  • Actual wins overstating year to date performance / potentially tending worse – Knights
  • Actual wins reflective of year to date performance – Broncos, Raiders, Bulldogs, Sharks, Sea Eagles, Storm, Warriors, Eels, Rosoters, Tigers

For teams with less than .500 percentage, 2nd order wins is always better than actual wins and vice versa. The trick here is to not focus on this specific value per se but to look at it as a prediction for next year’s performance. The next season’s performance for most teams will fall within one standard deviation (plus/minus .130) of their 2nd order win percentage for this season.

Form

Form Elo ratings at the end of 2020 and the end of round 14

Production

Pre-season projected and average year-to-date Taylors per game

1st Order Wins

Winning percentage estimated by Pythagorean expectation of actual points for and against

2nd Order Wins

Winning percentage estimated by Pythagorean expectation of SCWP for and against

Player Leaderboards

Wins Above Reserve Grade

WARG is a volume stat that compares the total amount of valuable work done (production), when compared to a replacement level player (or fringe first grader in the vernacular) at that position, irrespective of the time on field. A replacement level player has 0 WARG.

2020 was topped by Daly Cherry-Evans with 2.1 WARG. The career leader (2013 – now) is James Tedesco with 11.3 WARG. The single season record holder is Jarryd Hayne in 2014 with 2.3 WARG.

Taylor Player Rating

TPR is a rate stat that compares the amount of valuable work done (production) per game, factoring in time on field, to the average player at that position. An average player has a rating of .100. Minimum 5 games need to be played to qualify for TPR.

The 2020 regular season was topped by Cameron Smith with a TPR of .229. The career (2013 – now, regular season only, minimum 10 games) leaders are Harry Grant and Robbie Farah with a TPR of .164. The single regular season record holder is Robbie Farah with a TPR of .244 in 2013.

2021 WARG by position

WARG as generated when the player is listed in that position so far this season.

2021 TPR by position

TPR as generated when the player is listed in that position so far this season (minimum five games).

Primer – SCWP

People familiar with my philosophy will know that I put less stock in wins than most people. The binary nature – you either take everything or get nothing – means that a simple win-loss record is not a particularly nuanced and, unless you have a very long timeframe to work with, doesn’t necessarily reflect teams’ actual talent over shorter timeframes. Points difference and by extension, Pythagorean expectation, does a better job of reflecting true team ability but even that can be affected by luck or odd results. Does a 50-0 scoreline really tell you any more than a 30-0 scoreline about the relative disparity in talent? If a team scores more tries but loses the game, what does that tell you?

Baseball and college football analysts have developed a metric called “second order wins“. The actual win-loss record are considered to be zeroth order wins (nomenclature that I use and probably no one else). Pythagorean wins, the number of wins expected based on the team’s Pythagorean expectation, are considered first order wins. Second order wins calculates a Pythagorean expectation, not based on actual points scored, but utilising advanced stats to calculate expected points. The idea is that these expected points are more repeatable, and less subject to good/bad luck, and provide a less wrong basis for estiamting teams’ true talent and forecasting teams’ performances on that basis.

For use in rugby league, I propose the following hierarchy:

  • 0th order wins – actual wins
  • 1st order wins – wins as calculated by Pythagorean expectation of points for and against
  • 2nd order wins – wins as calculated by Pythagorean expectation of SCWP (Should-a Could-a Would-a Points) for and against

Note that third order wins are second order wins adjusted for strength of schedule. I’m not really concerned with this right now, given that everyone in the NRL plays each other once and then mostly twice so it is of marginal value.

SCWP is not what I would call an advanced statistic because there’s only so much I can do with the data I have. I have taken two metrics for rugby league – metres gained, representing field position, and line breaks, representing playmaking – as our key statistics on which to estimate expected points. I briefly toyed with including tackle busts but it did not improve performance and I suspect we would get a similar result with other stats.

In a similar process to building up the Taylors system, I took every NRL game (2013 – 2021 rd10), QCup (2016 – 2021 rd 7), NSW Cup (2016 – 2021 rd 10) and Super League (2017 – 2021 rd 5) and calculated the running metres and line breaks for each team in each game. I put these games into buckets and then calculated the average score for the bucket with a minimum of five games. The net result is a near 1:1 relationship between metres/line breaks and points scored.

The trendline for these graphs allows us to calculate the Should-a Could-a Would-a Points (i.e. the expected points) that we would expect the team to have scored given the metres and breaks made. We take the (basically) average of the points expected by metres and the points expected by breaks, resulting in the SCWP for the game.

The question then might be, why? The 2nd order winning percentage, based on SCWP, has a lower mean absolute error (MAE) when compared to next year’s actual winning percentage than 0th or 1st order winning percentage. Over the 2013 to 2020 NRL, 2016 to 2019 state cup and 2017 to 2020 Super League seasons (n = 221), we find:

  • 0th order winning percentage has a MAE of .149 when compared to next season’s winning percentage (equivalent to 3.6 wins over a 24 game schedule)
  • 1st order winning percentage has a MAE of .132 (3.2 wins)
  • 2nd order winning percentage has a MAE of .122 (2.9 wins)

Each iteration lowers the error by 10% when forecasting. There’s an additional layer of linear regression that could be applied over the top and this might replace the now defunct Poisedon ratings in pre-season sims.

The decreasing error is partly due to an in-built regression to mean, as SCWP typically has a lower margin than actual points which reflects the fact that teams always put in some effort, even when they get shutout on the scoreboard, and partly because SCWP reflects repeatable statistics, whereas the scoring of actual points can be somewhat prone to randomness (“we would’ve won that if he hadn’t dropped the ball three times/missed those conversions”, hence the name).

The current state of SCWP in the NRL (round 11), compared to actual for-against:

For Super League (round 6):

For Queensland Cup (first part of round 8):

For NSW Cup (round 11):

There’s an additional layer of efficiency to consider. I don’t know if the ratio between actual points scored and SCWP will prove meaningful but if a team is consistently outscoring what we would expect considering the fundamentals, that might either give us a clue about their style of play or it might signal regression to mean. This is something to keep an eye on.

There’s every chance that a SCWP v2 might come forward in the future, based on actual advanced statistics. I of course reserve the right to tinker with my own systems but I’ll let you know when I do.

Primer – Rismans

A big thank you to Lorna Brown (@_Lorna_Brown) who provided me with the dataset and whose ongoing updates to the same mean that we should be able to do some form of Super League player analysis. She has – presumably through some sort of black magic and/or competence with programming – managed to scrape a far more complete dataset out of the SL website than I managed to in previous attempts.

In short, a Risman is an English Super League equivalent to a Taylor. That is, it is a unit of measurement of rugby league production. Production is the accumulation of valuable work done on field as measured by traditional statistics.

The Risman, as a unit of production, is named for Gus Risman. He is a player whose name has largely stuck in my head due to Tony Collins’ podcast, Rugby Reloaded, wherein Collins makes the case that Risman is one of the all time great footballers of any code.

Gus Risman was one of the greatest of Cardiff’s rugby codebreakers. The son of immigrants who grew up in Tiger Bay, he played top-class rugby league for more than a quarter of a century, was a Championship and Challenge Cup winner with two clubs, and captained the 1946 Lions. Not only that, but he also captained the Wales in war-time rugby union internationals while a rugby league player.

Rugby Reloaded #138

As with Dave Taylor, the unit of production is named for a player who can do it all.

The Risman is derived by running linear regressions to confirm which statistics from the Super League dataset correlate with winning percentage. The stats get distributed in to buckets and we review the success of teams achieving those statistics (minimum ten games in the bucket). The result is that tries, try assists, missed tackles, tackle busts, metres, clean breaks and errors (negative) have significant correlations with winning. This is considerably less than the NRL dataset offers, which is why I’ve opted to give these production units a different name; Rismans don’t quite measure the same stuff as Taylors.

We multiply the stat by the slope of the trendline calculated in the regression and a weighting proportional to its correlation to winning (higher the correlation, the higher the weighting) and then by 1000.

Through this product of slope and weighting, we develop a series of equivalences between stats and can compare this across leagues. The following shows the quantity of each stat a player needs to accumulate to be equal to the same production as scoring one try for the 2021 season. The NRL’s values are calculated on the dataset of the five previous seasons, while the others are based on the three previous seasons (State Cups just roll over what the weightings should have been for 2020 to 2021, given they didn’t play last year).

For the record, a try is worth 8.7 Taylors in the NRL, 8.4 Ty in QCup, 7.5 Ty in NSW Cup but a whopping 17.3 Rismans in Super League. This, of course, doesn’t mean anything as Taylors and Rismans have no real world value.

Due to the limitations of the dataset, we can only calculate raw production. Without positional information or time on field, it is not possible to calculate more exotic ratings like an English TPR equivalent, Wins Above Reserve Grade or undertake pre-season projections.

Raw production is still somewhat useful and if nothing else, I think it will likely come in handy for assessing squad strength at the next World Cup. Teams with superior production, as calculated post-game, win 90% of their games.

The average player generates approximately 20 Rismans per game and for players with fewer than ten games, this figure is used until a reasonable sample size can be drawn upon for that player. Based on the actual 17 fielded, teams with the superior expected Rismans, as estimated pre-game by the sum of each player’s prior career average Rismans per game, has a 63.8% successful tipping rate (n=131). This is comparable to using Taylors in the NRL. Using the same formula for the NRL and the above, we can estimate a pre-game winning probability for a given line-up (re-deriving this formula based on the small SL sample meant that the team with more expected Rismans had a lower winning probability when teams were closely matched, which doesn’t make sense).

I posted a leaderboard of players by total Rismans up to round 5 of the Super League. As a not particularly close observer of that part of the game – I still perhaps have a better idea of what’s going in England than in NSW Cup – most of the top twenty at least rang bells as players I’d heard of.

I would have included an update for round 6 but the Super League website does not have any stats listed for the Leeds-Wakefield game, except for who scored the tries. So we must bear in mind that the dataset has some fairly significant limitations, not just in scope, but in completeness. For example, some of the Qualifiers games have been included but a lot, particularly those involving Championship teams, were not. Stats avilability for finals games seems to be hit and miss.

There’s also probably something to be said for different positions accumulating different typical quantities of production but without an independent arbiter of who plays what position, I’m choosing to be blind to this because I refuse to do this manually.

Nonetheless, here’s the all-time (2017 to 2021 round 6) Risman leaderboard.

As a couple of reference points, George Williams’ 43.1 Rs/gm has translated into a TPR of .119 at NRL level. This should be exciting for Tigers fans, as Oliver Gildart will presumably perform at a similar level when he joins Wests next year based on his 43.3 Rs/gm. Undermining that somewhat is Jackson Hastings’ 56.4 Rs/gm, the second highest of any player with at least 50 games, behind Greg Eden, compared to his career .052 TPR in the NRL and .080 TPR in NSW Cup. Hastings is also en route to the Tigers in 2022. Whether a real correspondence between different leagues’ ratings can be derived will probably depend on sourcing more information to bolster the dataset but it should be interesting to see how those signings pan out in the meantime.

The Dally Ms are a wank, let’s use…something else

Wait, haven’t we done this before? Absolutely, in a post that hasn’t aged super-well, I replaced the Dally Ms with the old player rating system to assess the best player in each season. That system has been improved upon with TPR and WARG, so the conclusions drawn there can be safely discarded.

Today, I specifically do not want to replace voting with a statistical rating. I cannot stress enough just how little I want to do that. Of the two half decent, publicly avilable player rating systems, TPR and the League Eye Test’s Net Points Responsible For, neither account for everything that happens on the field and attempting to quantify all aspects of sport dehumanises the experience of watching and enjoying rugby league.

Awards should be given partly on emotion because it’s stupid to assume there’s a purely rational way to hand them out. Rationally, we should only care who the best teams are, seeing as that’s the point of the sport, and who the best players are is a sideshow that should only concern hacks padding out column inches. Indeed, I believe the flaws in any system are ultimately good because that creates fuel for the content machine and keeps the sport in the news, particularly as the higlighting of flaws comes toward the end of the season when there are fewer games to talk about (this is my pet theory as to why college football doesn’t do away with its ridiculous system for anoiting a national champion). Moreover, these flaws reflect our own flaws as humans and that’s one of the things that makes life interesting. Irrational emotions are part of us and part of the sport.

The Dally Ms are back in the news cycle with a proposed change of voting system courtesy of a Buzz Rothfield shit-stirring Sunday column. Under the current system, the judge has to be at the ground to vote the best player on field three points, the second best two and the third best one. The player with the most votes at the end of the season wins. The current system typically favours good players on average teams who are able to sweep up the points on a regular basis, albeit the award has generally been given to the right player, or close enough to, at the end of the year.

Under the proposed system, each player gets a rating out of ten in every game. The flaws in this are obvious. Even the keenest observer would struggle to give every single player in a match an objective and justifiable rating out of ten. I definitely couldn’t do it without making up at least a few. People naturally, when presented with this kind of problem, tend towards an average of seven, and not five, because people are generally kind of nice. Your five and my five are not likely to be the same and there seems to be no way to calibrate for this, other than intensive training, something that is not an issue under the current system because while we might disagree who is the best, we both understand what the best generally means. What is average is another kettle of fish.

There’s also a scaling problem, which affects both systems, wherein a player who absolutely plays everyone off the park is given the max score but the score rarely reflects how far ahead the player is compared to the rest. If Taumalolo runs for 300 metres, scores three tries and pots a field goal, he will get either a three under the old system or a ten under the proposed. The next best player gets a two for his efforts, implying that Taumalolo’s performance was only 50% more valuable, even though he probably would have earned that three points after the first try and 200 metres. Similarly, if he’s awarded a ten, is everyone else a six at best by comparison? Meanwhile, in the next game, the best player scores one try and assists another but gets an equal three points or ten rating.

Bearing all of that in mind, here’s the top ten from the five most recent Dally M votes.

As a point of comparison, here’s the TPR and WARG champions for each of those seasons.

What I want to do is compare how different voting systems impact on the final results. We’re going to look at four different voting systems –

  • System A: after round 9, 18 and the end of the season, the top ten players are awarded votes from 10 for the best player through that part of the season, down to one for the tenth best. Votes are tallied at season’s end and the player with the most votes wins.
  • System B: is the same as system A but players are assessed after every round based on their performance in that round.
  • System C: the current Dally M voting system, as a control.
  • System D: the proposed system.

Rather than go back and watch every game for the last five years, close to 1,900 hours of entertainment, I’m using my player ratings as a proxy for a (non-existent) rational voter. Systems A, B and C are assessed on WARG while system D is assessed on TPR. I think this difference is justifiable based on how I think people would approach the different voting systems. When when assessing players individually on a 0-10 scale, I’d expect judges would compare them to players at the same position and account for time on field, as TPR does, while WARG does a better job of assessing who has done the best by minimising those factors in favour of raw production.

To translate TPR into a 0-10 rating, I tried to put ratings into buckets that more or less reflects a normal distribution, as follows.

One would expect voters to award their ratings in a likewise fashion but will probably skew it based on their personality and what they’re expected to judge.

Here are the new Dally M results for each system.

If we just compare the results from the different systems, system A generates three or four defensible winners out of five, system B only a single one in 2019, system C three or four and system D two to three. In my subjective opinion, the current Dally M system seems to perform the best, even if sex pest Blake Ferguson was to be awarded the title of “the best” in 2018.

You can review the top tens but I watched most of these seasons and I couldn’t possibly remember who was seventh best in 2017. The point is less about who is in what position but rather how the different voting systems affect the outcome. In none of the five years, despite having the same information, did the systems uniamously appoint a winner.

All systems are going to have pluses and minuses. The validity of the result comes from two things. The first is a widespread understanding what the purpose of the award is. Is it for the best and fairest or the most valuable or something else? Each of those mean different things and the system to award the winner needs to reflect its purpose.

The second is capability of the voters. The average NRL twitter user (and/or person reading this) is going to assume that the judges are idiots, because they produce largely terrible commentary, and because they are susceptible to the same groupthink, biases and laziness as the rest of us. That, too, is very human, as is hubristicly assuming you would do a better job over the long run. When I see two people on the timeline have the complete opposite understanding of what just occured in a replay, I know that the individual punters will not do a better job and that the utopia of a perfect player award system is as far away as ever.

The gimmicks will continue until flow improves

In their infinite wisdom, the NRL introduced rule changes over the second off-season. The two major changes were to reduce from two referees to one and the six again set reset introduced for ruck infringements. In the froth over the six again, going back to one referee has almost been an afterthought but has likely had a similar scale of impact.

It’s difficult to keep track of the purported benefits – pace, flow, fatigue, consistency and bringing back the little man – thanks to their vagueness. Anything that’s changed in the last few weeks has been ascribed to the rule changes, whether it made sense to do so or not.

With three rounds complete, it’s time to take an early look at how these changes have altered the way the game is played.

Embed from Getty Images

The stats

We can use stats to cut through what we see on the screen and delve into some of what’s happening.

Points

The perception that the rule changes have led to more blowouts or points is wrong. The average margin for the rounds played are higher than average but not unusually so (last three rounds in red):

Average round margin

Even the rolling average margin over three rounds played is higher than average, but still not unusual.

Average margin over three rounds

If the current average margin of rounds 3 to 5 is somehow maintained for the whole season, then it would be on par with the 2002 season but this seems very unlikely.

Average season margin

If the pace that points have been scored over the last three rounds is maintained at 38.3 per game, that will put the balance of this season on par with 2011, the lowest scoring year in the NRL era. The current rate is two points per game below the average of the last decade and ten points below the record of 48.9 per game set in 2001.

Average points per game

The points appear to have gone missing in penalty goals.

Penalty goals

To me, the results we’ve seen are well explained by good teams beating the shit out of bad ones. If anything is at play, the season suspension has created a disparity in physical fitness across the comp, which in tandem with the new rules reducing penalty goals, might explain the mildly strange combination of higher than average margins with relatively low scoring.

Time will tell, but as the fitness disparity closes, it will be replaced by a disparity in effort as teams are gradually ruled out of contention.

Running metres

Running metres have totally blown out.

All running metres

Note that I use the stats on NRL.com, which indicate that the total running metres of rounds 3-5 has increased by roughly 9% compared to rounds 1-2. Fox League’s stats also show an increase, on the order of 10%.

There seems to be some thinking that this phenomenon is because teams don’t get a lift down field from the kick for touch that they used to get. Kicking metres, however, if anything have gone up. Perhaps stuck in their own end and without a penalty to assist, teams are kicking for distance more frequently at the end of sets.

Kicking metres

Some of these increases might be explainable through chance, as well as evolving measurement methods but, in general, more stuff is happening, as counted by the stats, in the same game time.

[Note carlos uses Fox League’s stats]

Penalties

The main change is that the NRL has increased the amount of field covered by the players in exchange for a reduction in the number of penalties.

Penalties

But if you include the number of six again calls, then the refs are as involved as ever.

Penalties + six agains

Funnily enough, teams that were giving away calculated penalties in order to gain a defensive advantage are still doing so, it’s just being swept under the rug of a rebrand. I, for one, am shocked that savvy coaches and smart players who are famous for gaming the system would work out how to game the system.

It is clearly preferable to give away a new set on the first or second tackle and set the defensive line than to attempt to keep up with the pace, especially as referees are not giving attacking teams the usual leg up to get out of their own half that they have come to expect. Refs seem to have caught on to this strategy in round 5, issuing 50% more six agains more evenly spread across the tackle count, than in rounds 3 and 4.

Six agains

The six again is an intentional compromise between stopping the game for penalties and policing ruck infringements that slow the game down. It succeeds in removing penalties but it does not effectively help police the ruck, due to reasons that will be elaborated later.

Playmaker contributions

“The little man is back” is the most mystifying response to the rule changes. Trying to untangle what this actually means is an intellectual exercise on par with understanding quantum chromodynamics.

Using Taylors, the proportion of production generated by playmakers, defined as those wearing 6, 7 and 9, compared to the rest of the starters is the same now as it has been for the last few seasons.

Production by position

The little man is back, in the sense that he never went away. If anything, he came back a year early somehow.

Play the ball speed

I’m not a huge fan of the play the ball speed metric because it doesn’t seem to reliably mean anything about winning games of football but it can be at least help us identify a narrative.

Average play the ball speeds

Or perhaps not. It’s possible that the speed increase caused by six agains is offset by eliminating the second referee yelling at players in the ruck, so we kind of end up back where we were anyway.

****

When commentators talk about the “pace” of the game, I think they mean more stuff is happening in the same amount of time. When commentators talk about the “flow” of the game, I think they mean play the ball speed and minimising interruptions caused by awarding penalties. Insofar as these very generous interpretations hold up – the little man angle remains hard to fathom – and players are inarguably more fatigued, it’s questionable whether this is better.

Embed from Getty Images

The feels

No one can tell you what your aesthetic preferences are, how rugby league looks and feels to watch and enjoy, or offer a certain judgement as to what is better. Aesthetics are pure subjectivity.

I can offer my opinion and justify it for what it’s worth to you, which is that the style of gameplay is not sufficiently different to justify the enormous volume of plaudits that were thrown around when the season resumed. Where the gameplay is different, this hasn’t actually improved my enjoyment of the game.

The immediate response to the rule changes was for players to do everything more, which created the illusion of filling air time with action. Round 3 reminded me of the frenetic period in the late 60s following the adoption of limited tackles. Players hadn’t optimised their tactics so responded, disoriented and panicky, by running the ball.

We’re seeing this settle relatively quickly. I expect that the amount of stuff done per game will find a new level, higher than we were used to previously. In the long run, we will get used to this but I’m yet to be convinced that more is better. If nothing else, more running metres per game cheapens the value of each metre made.

The irony is that the subtleties of the game’s structure were already happened so quickly that they were easily lost in the motion on screen. This has created a stereotype that the game is solely one-up hit-ups when nothing could be further from the truth for a well-drilled team. Speeding up has made it more difficult still to see the underlying shape of the game.

Conversely, it is entirely possible that this distortion is created by a newfound decreased emphasis on structure. Why execute complex plays when you can simply wait for your opposition to tire out and then run over the top of them? It might be easier, possibly even funnier, but I don’t believe this is more entertaining.

Embed from Getty Images

The gimmicks will continue until flow improves

Discussing the effectiveness of the rule changes merits a discussion of the context in which they came about. Glossing over the fact that the chairman has to date grossly overreached his job description and acknowledging the lack of transparency in the broadcast negotiations, the next biggest issue V’Landys’ leadership has introduced is making reactive changes in response to “problems” – largely the invention of a few nostalgic boomer hacks – and then using weak justifications after the fact.

For example, most people don’t care if there’s one referee or two on the field but for some reason, it grinds the gears of a particular subset of the NRL audience that we had two refs. So be it, but if the argument is the ruck will be policed better by one referee who’s also setting the line than two referees, one at the line and one at the ruck, then that is obviously bogus. Justifying it by cost savings that evaporated as soon as they came under any scrutiny and then by reference to a Daily Telegraph fan poll does not pass the smell test. So what was the point of the change? If we can’t publicly acknowledge who or what is driving these changes, why do they get accommodated?

The application of the six again in two games shows why this administrative approach is flawed. While Parramatta were defending a last minute Penrith attack on Friday, players were still lying about in the ruck to waste time and stymie the offence. Yes, the Eels’ Dylan Brown was sent off but why not just do that in the first place? Why introduce a new rule that prevents the refs from cutting to the chase and more effectively refereeing the game? Penrith lost the game.

Arguing that refs can still penalise ruck infringements if they so choose ignores the obviously political environment in which the referees operate. We know penalties are not acceptable to management because that’s what the referees used in an attempt to clean up ruck infringements in 2018.

The refs were castigated for it because the childishly cranky part of the NRL audience that the decision-makers listen to had a whinge that they were getting bored with the stoppages. Whether the stoppages would have the desired impact in the long-term never got a run because in the short term, tantrums were being thrown about blown whistles ruining the flow and, in a supreme act of psychological projection, referees trying to make themselves the centre of attention.

In an ideal world, perhaps the commercial and judicial arms of the sport would be separate but they aren’t in rugby league. Greenberg told the refs to lay off and they did. Players and coaches were not incentivised to change and the so-called wrestle continues.

In fact, referees now have more responsibility and potential impact on the game, having to use their discretion as to whether a ruck infringement justifies no penalty, a six again or an actual penalty. Teams who find themselves in situations where they would prefer a penalty than a set restart are not given an option, as demonstrated in Thursday’s Manly-Brisbane game where two set resets were given within twenty seconds of the last five minutes of the game, in lieu of a game-tieing shot at goal. The Broncos lost.

These consequences would have been obvious if they were thought about before implementation. Famously, the incredibly named Project Apollo’s innovation committee only had one hour to consider the changes. Once the novelty has worn off, questions will be asked, not just by unimportant nobodies with a WordPress account, but also by people who are actually listened to as the fallout become impossible to ignore. The Peanut King has already fired a shot across the bows, although I refuse to read what what he’s actually said.

In the rush to be seen doing something, V’Landys risks either looking foolish in rolling back the changes or worse and more likely, he will double down. For example, the suggestion that next season the scoring team will kick-off to prevent teams from getting a roll on is a dire sign but we’ll see what actually comes to pass. The slippery slope argument is that if he chooses to double down, V’Landys will apply band-aid gimmick after band-aid gimmick until the sport is barely recognisable or enjoyable to watch, satisfying no one and leaving everyone wondering how we got here.

While rugby league has a tradition of innovation to attract the mass spectator, it is also extremely questionable whether the ends will justify these means. The alternative – to take some time to consider changes, think through the second-order impacts, trial at lower levels, implement between seasons and transparently state the justification – is there to be utilised.

After a huge surge driven by a palpable sense of relief at the return of the footy, TV ratings are back to where they were pre-coronavirus. Nonetheless, the rule changes are likely here to stay.

A deep dive in to the 2020 NRL premiership

This is my third season preview and I have got some things laughably wrong in the previous attempts (see 2018 and 2019). This year’s will be a slightly different format to previous years but undertaken in the same spirit of considering each team’s strengths, weaknesses and opportunities, as well as assessing the changes made since last year and their potentially positive or negative impact on performance.

However, I plan to have fewer laughably wrong predictions in 2020 simply by making fewer predictions. After all, if you want to see laughably wrong rugby league analysis, you can just pick up a copy of the paper.

Last season in a nutshell

2019 was a weird season and completely different to its equally weird predecessor. In 2018, eight teams finished within a win of each other and then were systematically dismantled by the Roosters and Storm in the finals. In 2019, we had three teams that could clearly play football, another couple that were adequate and a bunch of losers that didn’t want to make the finals. The round 17 golden point field goal shoot-out between the Broncos and Warriors, leading to a draw after multiple botched attempts, encapsulated the lose-at-all-costs mentality that defined positions seven through fifteen on the ladder. In the end, the Roosters emerged victorious in a manner that still infuriates me, with the Raiders running out of points and the Storm running out of steam when it counted.

A relatively quiet off-season – dominated by Latrell Mitchell’s signature, the Tigers’ warchest, Melbourne pollinating the landscape with overpriced talent and what the second Brisbane team should be named – has seen most teams turn up to 2020 in roughly the same shape as they approached 2019. It makes it very difficult to get a grasp on how this year might pan out, without just repeating pretty much what happened in 2019. And, no, neither the Nines nor pre-season trials will provide any insight.

How it all works

I appreciate that it’s difficult to keep up with the Pythago NRL Expanded Universe™ of metrics and ratings. Not only are they generally more complicated than standard stats, I tweak them almost every year based on what I learned during the previous season. I created a short reference guide to what it all means.

2020 team projections are based on round 1 lineups, taken as a mix from NRL.com and League Unlimited. 2020 roster composition is based on the listed signings on League Unlimited (as of 28 February) but 2019 roster information is based only on players who played at least one game.

Jump ahead

Embed from Getty Images

nrl-bne Brisbane Broncos

2020-bne

2020-bne-2

2020-bne-3

2020-bne-4

Brisbane deserved to finish ninth or tenth last season. The Broncos were the second most heavily biased to their forwards, behind the Cowboys and the immutable Taumalolo. The strong and young forward pack means that the Broncos are projected to have the third most production in 2020 but there’s diminishing returns in having powerful forwards if the other parts of the team continue to struggle to execute. The reality is that Brisbane needs less stupidity out of the forwards, more offence out of the backs and an all round improvement in defence.

I assume we will see more of the same from last year because nothing has changed significantly enough to suggest otherwise. Giving the captaincy to Glenn over Boyd doesn’t change the fact that neither should be on the field. If Boyd plays anywhere, that side of the field will shut down in attack and one or two players will have to cover his defensive workload. None of the talk out of the club has really addressed this or any of the many other problems, so I don’t see how they could have fixed them.

As to what question Brodie Croft answers, I don’t know but it isn’t halfback production. Ironically, I think the team would perform better if Milford’s TPR was lower and he didn’t have to waste time carrying so much dead weight, both undercooked rookies and overcooked veterans.

Embed from Getty Images

nrl-cbr Canberra Raiders

2020-cbr-1

2020-cbr-2

2020-cbr-3

2020-cbr-4

Possibly more than any other team, the Raiders have lost the most talent in the off-season. Taylor is down on their prospects but expects Canberra to still perform above average. Elo and Poseidon, carrying through from 2019, expect them to return to premiership contention. The Raiders’ defence wasn’t quite enough to win them the premiership (as a rule of thumb, the Poseidon defence rating should be at least +50) and it would be unlikely to not see some reversion towards mean this year. With luck, it won’t be as disastrous as 2017 and 2018 following 2016.

While Canberra’s defence was good, the attack completely dissipated in the finals. Bringing in an English half is a risk, but so was bringing in English forwards, and it paid handsome dividends. By all accounts, George Williams is the goods and might be the missing piece of the puzzle. Leilua, Rapana and Sezer have all left in the off-season, to be replaced by Curtis Scott, who celebrated by punching some cops. After being mired mid-to-lower-table for so long under the decade-long dual dominance of Sydney and Melbourne, it would be genuinely surprising to see a team turn a corner and transform into perennial challengers.

Embed from Getty Images

nrl-cnt Canterbury Bulldogs

2020-CNT-1

2020-CNT-2

2020-CNT-3

2020-CNT-4

The Bulldogs are behind, way behind.

With Kieran Foran missing most, if not all, of the 2020 season, the Bulldogs either need significant development out of their relatively young squad or to land some signatures. Neither seem likely, especially as the club is likely to still be paying freight on players from the Castle-Hasler era and the current squad do not have the track record to suggest any superstars are emerging (perhaps Renouf To’omaga excepted). The players signed to development contracts do not have particularly impressive stats from the NSW Cup. With last year’s significant outperformance of the fundamentals, reversion to mean would likely mean a wooden spoon.

However, we’re now into our second full season of rebuild at Belmore and the signs have been promising. Late surges of form in 2018 and 2019 when other teams start to switch off towards the end of the season have often been timely, snagging wins that Canterbury have no right to and desperately need. This defiance indicates that Dean Pay can coach (“Dogs of war”, etc, etc) and jag the seven or eight wins required to avoid the spoon. I’m comparatively bullish on the Bulldogs but they need to resolve their cap issues to get some talent on board if they want to really progress.

Embed from Getty Images

nrl-cro Cronulla Sharks

2020-Cro-1

2020-Cro-2

2020-Cro-3

2020-Cro-4

With their home games moved to Kogarah, we may finally get an answer to the age-old question: what do the Cronulla Sharks actually do to justify their place in the NRL? 

The Sharks’ 12-12 record and seventh place belied how well they played last season. Let down significantly by their goal kicking, the Sharks lost a record five games despite scoring more tries. While that’s a NSWRL/NRL record, I doubt that’s ever happened at any other time in football. The odds of it are simply astronomical. Tack on a couple of extra wins to last year’s total to appropriately set your expectations.

Cronulla should have the talent to comfortably make the finals in 2020. We probably won’t see anything much more interesting than that out of them unless a couple of the top clubs stumble.

With Paul Gallen retired, the team will have to adjust their production bias away from the forwards. I still have question marks on Bronson Xerri but his production last year was impressive and Braden Hamlin-Uele should probably be starting.

Embed from Getty Images

nrl-gct Gold Coast Titans

2020-gct-1

2020-gct-2

2020-gct-3

2020-gct-4

Perhaps the most significant thing to happen to the Titans last season was being surpassed by Newcastle, to be left at the bottom of the league in class Elo ratings. It might be recalled that the Knights were the worst NRL team of all time in 2016 and since then, the Knights have gotten better and the Titans so much worse.

Last season, you would have only taken a handful of players from the Titans to your own club given the opportunity: Arrow, Fotuaika, Brimson (who has a surprisingly low TPR) and maybe Tyrone Roberts if you were feeling generous. The Titans managed to hang on to them, except Arrow who will be departing for Souths next year. The rest of the roster under Garth Brennan was a joke, hence the 4-20 record, so hopes are pinned on the incoming Justin Holbrook, having left the best Super League team for the worst NRL team. Indeed, last season the Titans were ranked lower than half of the Super League.

With the number of experienced veterans and the talent pool on their door step, the Titans really should be better than they are. They are not and the sims reflect it. Fans will hope the new coach can get more out of the squad. Appointing Kevin Proctor captain is not the most auspicious start to turning around the club’s culture. Sick 9s jersey though.

Embed from Getty Images

nrl-man Manly Sea Eagles

2020-man-1

2020-man-2

2020-man-3

2020-man-4

The Taylors bear out how low expectations were for Manly in 2019, only for those expectations to be obliterated. The Sea Eagles were one of the few teams outside the big two that could win regularly. I went out on a limb pre-season and suggested Manly would make the finals. While that was pure luck on my part, they managed to do it. It turns out Des Hasler can still coach, even after taking some shine off his reputation while at the Bulldogs.

Backing up without the element of surprise and the reversion to mean will be challenging. Reversion to mean is a harsh mistress and often a huge outperformance is punished with an equally severe reaction in the opposite direction in the following season. The law of averages demands its tribute. For now at least, Manly’s prospects for 2020 appear to be good and based on sound fundamentals.

It hasn’t been discussed nearly enough how costly Manase Fainu missing some (most? all?) of the upcoming season will be. He was one of the big unknowns that stepped up last year and with Api Koroisau now at Penrith, Manly are bereft of options at hooker. It is too early to discuss Cade Cust as a long-term successor to Daly Cherry-Evans but he had an impressive debut season.

Embed from Getty Images

nrl-mel Melbourne Storm

2020-mel-1

2020-mel-2

2020-mel-3

2020-mel-4

The Storm and Craig Bellamy, as they often are, were the biggest outperformers of their projections in the league. Melbourne finished the season with a 20-4 record, a record only bettered* by the Storm’s 21-3 2007 season. Unlike 2017, where it seemed inevitable that the Storm would win the premiership after winning 20 games, they never seemed to get much credit for what was still a very impressive season in 2019.

Melbourne just have the knack of taking extremely talented young men, putting them on the football field and winning games. Positions don’t seem important, neither do the names. It will likely continue forever because there is plenty of talent pushing through in reserve grade. Even the departure of several reasonable quality players doesn’t seem to have made a dent in their prospects.

So yeah, they’re pretty good. If I’m lucky, I may live long enough to see the next Broncos win over the Storm, an event about as frequent as Halley’s Comet.

Embed from Getty Images

nrl-new Newcastle Knights

2020-new-1

2020-new-2

2020-new-3

2020-new-4

The Knights will be glad to see the back of the 2010s, where they were the worst team in the NRL and nearly went broke. The good news is that the Knights might legitimately make the finals this year.

The Knights massively outperformed in 2018, which then led to many talking heads predicting serious success in 2019. Success wasn’t forthcoming because the fundamentals weren’t there. Instead, we had a heady mix of nostalgia, over-excitement and Blue bias that completely crippled the predominantly Sydney-based media’s capacity to objectively analyse (I have the same problem in the opposite direction but at least I’m aware of it).

Mitchell Pearce had a career season in 2019, at least until I wrote about it, but otherwise the team struggled to meet expectations. I’m more of a numbers guy than a culture guy, but even I could see that the team was often not trying. Results from round 16 through 21 last year bear that out. Their thrashing at the hands of the Titans in round 5 was more typical of the season than the six wins that followed.

The finishing touches to the “rebuild” have now been applied, not least Adam O’Brien replacing Nathan Brown as head coach, to bring the Knights back in contention for finals places. Newcastle are still a way off challenging for the premiership.

Embed from Getty Images

nrl-nzw New Zealand Warriors

2020-nzw-1

2020-nzw-2

2020-nzw-3

2020-nzw-4

People think the Warriors are bad. They haven’t been. New Zealand’s club embodies average-ness with every fibre and loves to squander an opportunity. The thing about the median is that it’s not last place, so I’m always wary of any prediction that gives the spoon to the Warriors.

The loss of Shaun Johnson was not well compensated and the team is now overly reliant on Roger Tuivasa-Sheck and the back line to generate production. The forward pack has not been impressive as a whole. The lack of star power – currently projected to be zero players – is concerning, although not damning. Kodi Nikorima is, at best, a below average halfback and Chanel Harris-Tavita is apparently too young to start but he’s far better bet (.098 in 2019 compared to the .085-ish range Nikorima has played in the last three years). The Warriors will chase eighth place with the Broncos, Tigers and Knights until they get tired and slump down the ladder.

More worryingly, the Warriors are on the precipice of falling full-time into the ‘bad’ category and once that happens, I don’t know how the club will pull itself out. The Auckland Rugby League should be a conveyor belt of talent and the Warriors should be at least Broncos-calibre, if not the Storm. Until that gets worked out, New Zealand will probably bounce along the bottom of the ladder.

Embed from Getty Images

nrl-nqc North Queensland Cowboys

2020-nqc-1

2020-nqc-2

2020-nqc-3

2020-nqc-4

A good showing at the 9s tournament in Perth has clouded judgement about what the Cowboys are capable of. Consider their stacked halves options of Michael Morgan, Jake Clifford and Scott Drinkwater. Drinkwater is only a thought there because Valentine Holmes is obviously the fullback. The ever-reliable Vaa’i Taumalolo will put the team on his back and Kyle Feldt will finish in the corner.

It sounds good in principle but most of these pieces have been available for the last three years and, other than limping to the grand final in 2017 and avoiding the spoon in 2018 and 2019, those three years have had little to celebrate. After all, we’re projecting a team with some well-known players to only be twelfth best. Without Taumalolo, a certified freak and statistical anomaly, that number would be a lot closer to the bottom.

Paul Green seems intent on stifling the creativity of his playmakers and/or was overly reliant on Johnathan Thurston to make plays. Either way, he has to adjust to the new Thurston-less world where scoring six to twelve points is not going to be enough. Despite delivering the premiership in 2015, a bad 2020 might be the end of the road for Green.

Embed from Getty Images

nrl-par Parramatta Eels

2020-par-1

2020-par-2

2020-par-3

2020-par-4

I think this is it for the Eels. They are due for their once-a-decade (give or take) tilt at the premiership.

The Taylors are not too crash hot on the Eels. There are holes in key TPR ratings: Reed Mahoney at hooker, Dylan Brown nominally at five-eighth and, to a lesser extent, Clint Gutherson at fullback. The forward pack is slightly above average but none are exceptional. Reagan Campbell-Gillard might be one of those high-TPR, low-impact players, like Aaron Woods. On the other hand, Parramatta are capable of outperforming their projections which, for their top players at least, seem conservative. Last season’s hiccups only came when meeting the Storm, a hurdle that has felled better teams in the past.

The Eels are one of the better set up football clubs in Sydney. They have a good new stadium in the heart of their community, not too far from their leagues club. They’ve had a reasonable amount of on-field success the last few years if we ignore the total and inexplicable collapse that was 2018 (which might explain the conservative projections). It will be worth keeping an eye out to see if the club an build on this and win two premierships this season to complete their five year plan.

If not, 2021 will probably be a tear down, followed by a firesale clearance, and then a rebuild.

Embed from Getty Images

nrl-pen Penrith Panthers

2020-pen-1

2020-pen-2

2020-pen-3

2020-pen-4

The numbers suggest a tough year ahead for the Panthers, with not much to look forward to. The projected team is only two Taylors per game better than the Bulldogs. TPR lists only four guys worth a damn, roughly the same as the Titans. The sims have ten wins and eleventh place on the ladder picked out for Penrith, a re-run of 2019.

My gut says Penrith could do a lot this year. The grand final might be a step too far but it wouldn’t surprise me to see them scrapping for a second week final nor would it surprise me if we wrote them off as finals contenders shortly after Origin. The risk is there is plenty of potential but not a lot of proven execution, as last year’s rookies become this year’s sophomores and the pack that was bulldozing the league a few years ago slowly being whittled away.

It might not matter if this year is a write-off for the Panthers if they can channel the experience into development, making this squad better in future campaigns. Ivan Cleary and a Gould-less Panthers will have to take better care of the next generation than they have done in the past.

Embed from Getty Images

nrl-ssr South Sydney Rabbitohs

2020-SSR-1

2020-SSR-2

2020-SSR-3

2020-SSR-4

I have a quiet confidence in Souths’ premiership aspirations but questions remain unanswered.

Souths’ spine is projected to be a full five Taylors per game better than Melbourne’s, which is next best, so it is little surprise that the Rabbitohs are the mostly heavily biased to their playmakers. Damien Cook is the keystone of the spine and has been the league’s most productive player by TPR two years running. The .200 barrier hasn’t been broken since Robbie Farah did it back-to-back in 2013 and 2014, years that the Tigers won a combined seventeen games. After two years of wrecking the league, have coaches finally watched enough tape of Damien Cook to put a lid on him? More pressingly, will Damien Cook turn up this postseason?

Latrell Mitchell’s mooted move to fullback returns him to a position he hasn’t officially played since his 2016 season for the Roosters. He put up an average TPR of .087 then. Mitchell is projected to carry through his (famously quite lazy) productivity at centre and bring .120 of production to fullback. I am loathe to make individual manual tweaks to my systems, so that seems like a bad assumption that is worth adjusting for. 30 pips of production at fullback is worth about 10 Taylors, enough to move Souths from fourth best squad to outside the top eight. Questions: will Latrell at fullback work? Will Latrell put his full back into working?

If they fail, it is not clear if the rest of the team will be able to pick up enough slack to keep the Bunnies in the premiership hunt. Adam Reynolds and Cody Walker form a potent pair. Cameron Murray looks ready to go up another level. But is the forward pack good enough without numerous Burgii? Edene Gebbie looked a little lost at the 9s, so who else is waiting in the wings if needed?

Is Wayne cooked?

Embed from Getty Images

nrl-sgi St George Illawarra Dragons

2020-sgi-1

2020-sgi-2

2020-sgi-3

2020-sgi-4

I didn’t want to make any specific predictions but wooden spoon, anyone?

It would be the first for Illawarra since 1989 and the first for St George since 1938. The reality is that Paul McGregor’s head is already on the chopping block. Since taking the reins, the Dragon’s class rating has dropped nearly 100 points, an untenable position and one no major league coach of the last two decades has been able to drag their team out before their time was up. No improvements to the roster, no improvements to coaching… wait, didn’t the Dragons sign Shane Flanagan as an “assistant”? That will be an interesting play and may well push the Dragons up the ladder.

The squad itself isn’t magic but should be better than last place. New signing Isaac Luke has always been a productive player but he will presumably be second fiddle to Cameron McInnes when he returns from injury, reducing the potential volume of work Luke could be doing. Indeed, St George Illawarra are extremely reliant on their spine to perform. While Hunt, Norman and McInnes have been productive, I don’t think they’ve been especially effective. The Dragons are also still searching for a fullback. Lomax may or may not be it.

If Flanagan really is the de facto, if not de jure, head coach, then he should be able to coax that performance out of the roster. If McGregor is still in charge, then a 5-0 start will turn into a 7-17 season and the cycle will begin anew.

Embed from Getty Images

nrl-esr Sydney Roosters

2020-eSR-1

2020-eSR-2

2020-eSR-3

2020-eSR-4

I’m quite comfortable assuming that the Roosters won’t go three in a row. They’re still good though, probably even still good enough for a minor premiership. A projected 15+ wins and the second best squad on paper is not going to have trouble reaching a preliminary final. The Storm are the only team superior on paper and they share the equal best class Elo rating.

When we talk about the trinity of rugby league – hungah, pashun and desiyah – do the Roosters still espouse these values? Cooper Cronk’s retirement and nominal replacement with near-rookie Kyle Flanagan is the kind of loss of edge that turns premiership winners into runners-up, as the Storm have amply demonstrated.

After all, it’s not just about production. Yelling at other players to get them organised is a rare and extremely valuable commodity. Luke Keary may have it but it will be the first time in his career that the 28 year old will be the elder of the halves pairing. But to put this supposed weakness into context, the Roosters will absolutely be a top four team come September.

Embed from Getty Images

nrl-wst Wests Tigers

2020-wst-1

2020-wst-2

2020-wst-3

2020-wst-4

The Tigers continue to defy my predictions of a wooden spoon to instead finish ninth. Last year, they really should have been eighth and the 12-12 record the year before should have seen them in the top eight. Basically, bad luck has kept them from breaking the NRL’s longest finals drought.

Still, you make your own luck. The Tigers were the biggest movers in the off-season and showed unusual astuteness in their acquisitions: Leilua times two, Adam Doueihi, Walters and maybe Harry Grant (.266 TPR in 2019’s QCup) will land.

The projections and the sims lock in a knife-edge battle for the Tigers to take that final step from ninth to eighth. Exactly 50% chance of making the finals, exactly 12.0 wins projected and an average finishing position of 8.6. I’m not ready to make them a lock but this is the best chance Wests have had in a long time.

All they had to do was spend their money wisely. Now they just need to lock down a home ground.

Primer – TPR

For the third season in a row, I’m changing the player rating system. We mourn the passing of Statscore (not really) and PPG (again, not really) as we slowly converge on to a system that I can take for granted and don’t have to refine any further.

The core of the system hasn’t changed. The proposition is that there are important and unimportant statistics and that counting the important ones provides information about players and teams and can be predictive.

PPG was useful, and development and application through 2019 demonstrated that:

The last one should be taught in universities as a perfect example of ringing the bell at the top. Sheer narrative power subsequently forced Pearce back to mean and Brown onto the compost heap.

The mechanics of PPG have been preserved through TPR. My biggest issue is that when I wrote about production (that is, the accumulation of useful statistics), I didn’t have any units to work with. I originally didn’t think this would be a problem but it would make some things clearer if I did have units. So I took a leaf from the sciences and landed on naming it after the man that could do it all, David “Coal Train” Taylor.

Embed from Getty Images

“PPG”, which was Production – and not Points – Per Game, doesn’t make much sense now, so that’s been punted and replaced with TPR, or Taylor Player Rating. There has been a substantial change in the way I’d calculated WARG in the primer at the start of 2019 and the way I calculated it in Rugby league’s replacement player at the end. The latter method is now canonical but the name is going to stick.

In brief, TPR and WARG are derived through the following six steps:

  1. Run linear regressions to confirm which statistics correlate with winning percentage. The stats get distributed in to buckets and we review the success of teams achieving those statistics. One crucial change was to exclude any buckets from the regression with fewer than ten games in it. We end up with tries, running metres, kick return metres, post-contact metres, line breaks, line break assists, try assists, tackle busts, hit ups, dummy half run metres, missed tackles (negative), kick metres, forced drop outs, errors (negative) and, in Queensland only, penalties (negative) as having significant correlations out of the data provided by the NRL.
  2. Take the slope of the trendline calculated in the regression and weight it by its correlation (higher the correlation, the higher the weighting). Through this weighting, we develop a series of equivalences between stats. The below is shows the quantities required of each stat to be equivalent to one try in 2020:
    equivalences
  3. Players who accumulate these statistics are said to be generating production, which is now measured in Taylors, and is the product of the weighting/slope multiplied by the quantity of stats accumulated multiplied by 1000. However, due to the limitations of the statistics, some positions on the field generate significantly more Taylors than others.
    Average Taylors per game by position (1)
  4. To combat this, the production generated each game is then compared to the average production generated at that position (averaging previous 5 seasons of data in NRL, 3 seasons for State Cup). We make the same adjustments for time on field as in PPG and then divide by 10 for aesthetic purposes. The resulting number is the Taylor Player Rating, or TPR.
  5. We derive a formula for estimating win probability based on production for each competition and then substitute in a winning percentage of .083 (or two wins in twenty-four games, per the previous definition of a replacement-level team) and estimate the amount of production created by a team of fringe players against the competition average. This gives us a TPR that we can set replacement level at. The Taylors created over and above replacement level is added to the notional replacement level team’s production and the increase in winning probability is attributed to that player as a Win Above Reserve Grade, or WARG. Replacement level in TPR for the NRL is .057, Queensland is .072 and NSW is .070. The career WARG leaders are currently:
    career warg
  6. Finally, we go back and check that it all makes sense by confirming that TPR has some predictive power (~61% successful tipping rate, head-to-head) and there’s a correlation with team performance (~0.60 r-squared for team season production against team winning percentage).

For a more in-depth explanation, you can refer back to the original PPG primer. The differences between last year’s system and this year’s are slight and, for most intents and purposes, PPG and TPR are equivalent. Some of the changes are small in impact but important.

The most obvious change is the addition of NSW Cup data to the Queensland Cup and NRL datasets. This was driven by my interest in assessing the farm systems of each NRL club and you can’t make a decent fist of that if you’re missing twelve feeder clubs from the picture. It will also allow me to better test talent identification in the lower levels if I have more talents to identify and to better set expectations of players as they move between competitions.

For the most recent seasons, TPR only uses past data to calculate its variables, whereas PPG used all of the data available and created a false sense of success. A system that uses 2018 data to create after-the-fact predictions for the 2018 season isn’t going to give you an accurate view of how it will perform in 2019.

Finally, projecting player performance into the future is a pretty powerful concept, even if the tools for doing so are limited. I went back and re-derived all of the reversion-to-mean formulas used in The Art of Projection. It turns out that the constants for the projection formula don’t change much between seasons, so this is fixed across the datasets for now. It also turns out adjustments for age and experience are different and largely useless under the TPR system, such is the ephemera of statistical analysis.

One application for projections is that I’ll be able to run season simulations using the winning probability formula and team production that will be able to measure the impact of including or excluding a player on the outcome of a team’s season. It may not be super-accurate (the projections have large average errors) but it will be interesting. I also like the idea that out- or under-performance of projections as an assessment of coaching.

Finally, to reiterate things that I think are important caveats: TPR is a value-over-average rate statistic, while WARG is a volume statistic. No, statistics don’t tell the whole story and even these ones don’t measure effectiveness. Yes, any player rating system is going to have a certain level of arbitrariness to it because the system designer has to make decisions about what they consider important and unimportant. I’m fully aware of these things and wrote 1500 words accordingly at the end of the PPG primer.

A thing I’m trying to do this season is publish all of my rating systems on Google Sheets so anyone can have a look. You can see match-by-match ratings for NRL and the two State Cups if that’s your jam.

« Older Entries