AFL WOMEN’S Rankings: Round 4 2021

Yes, I will acknowledge Fremantle…

In last week’s post I said that even though Fremantle had won ten matches straight, they were ‘beatable’ and arguably hadn’t been as dominant in their wins as the Kangaroos. The Kangaroos have then proceeded to lose their second straight match, didn’t kick a goal, and the Dockers now sit on top of the rankings.

So yes, OK, the rankings say that Fremantle are the best team now. Though watch out for those Crows lurking just behind … they did have as many scoring shots as the Dockers when they played them, you know – which brings me to my next topic…

Points or scoring shots – what’s a better indicator of form?

My rankings are based on the final points margin. On the weekend though, Melbourne had double the scoring shots and inside 50s of the Bulldogs, and lost by two goals. The Kangaroos had as many scoring shots and more inside 50s than Collingwood, and lost by 20 points. As I mentioned above, Adelaide had as many scoring shots and inside 50s as Fremantle last week. Should we really be looking at the number of scoring shots instead, at least for AFLW where there’s less opportunities for goal accuracy to even out over the game?

Maybe. If we compared teams’ percentage by points this season with percentage by scoring shots, they are (probably unsurprisingly) fairly similar (see table below). The most significant changes are that Adelaide has the second best percentage, while Brisbane and Collingwood fall down the order – Brisbane in particular goes from a giant percentage of over 250 to a more ‘mortal’ 150 mark. At the other end, the improving Tigers look even a bit better.

Unlike the rankings the table above is not adjusted for opponent strength, but the scoring shot relativities don’t seem too far off where teams would be commonly thought to ‘sit in the pecking order’ so far this season. Adelaide and Brisbane seem to be more in line with where they ‘should’ be – although there could be some recency bias in that assessment given the weekend’s result – though perhaps the undefeated Magpies may seem on the low side.

AFL WOMEN’S Rankings: Round 3 2021

Fremantle’s ten wins in a row is very impressive, but they still seem to be ‘beatable’.

It may seem like being a ‘hater’ to say it – I assure you I am not –  but as great an achievement as the Dockers’ ten straight wins is, don’t declare the premiership race over yet.

Yes, Fremantle flattened two-time premiers the Adelaide Crows by 30 points on the weekend, but they did it with the same number of scoring shots, less disposals, and about the same number of inside 50s (although the Dockers do tend to get higher impact per disposal). In fairness, they could well have beaten West Coast by much more than nine points if they had higher accuracy in front of goal the week before, in a game where they had 35 to 14 inside 50s.

Nevertheless, I would still say they are not a nigh-unstoppable force, just a leading contender. The majority of their ten wins have been by three goals or less. The Kangaroos have arguably been slightly more dominant overall in their wins, though they keep falling short against Melbourne and snuck home in their final against Colingwood. But another undefeated home-and-away season (assuming the season finishes), and then home finals, would certainly help the Dockers in their quest for a first flag.

Aside from the Dockers and Kangaroos, the other teams that the rankings like to contend for the final six are Adelaide, Brisbane, Carlton, Collingwood, and Melbourne, keeping in mind we do not the rest of the fixture. Carlton has only one win out of three matches, but does play Richmond next week.

Or at least they do at the moment. Here are my tips until the next set of state border closures changes the fixture…

AFL WOMEN'S Rankings: Round 2 2021

AFL WOMEN’S Rankings: Round 1 2021

Collingwood’s win against Carlton was perhaps not that much of an upset.

Welcome back to my AFLW Rankings, for season 2021.

In the season opener, Collingwood recorded a one goal win against Carlton (it would have been two goals if not for a goal after the siren to the Blues), which was largely proclaimed as an ‘upset’. This would likely be because last season, before it was prematurely halted, Carlton finished second in their conference while Collingwood finished fourth, and because the Blues easily won their first final while the Magpies narrowly lost theirs.

Collingwood though finished off 2020 really well. Their narrow finals lost was against the mighty Kangaroos, by a mere two points. Prior to that they beat Brisbane and the Bulldogs by around five goals each, and lost to the other top team Fremantle by only three points (they also lost to Melbourne – note though they also beat Carlton prior to that, but that isn’t counted in their ranking points). After adjusting margins for opponent strength and home ground advantage, that gives them an average adjusted net margin of over 20 points (see chart below).

Matter of Stats, which has just started predicting AFLW results, also have Collingwood rated highly, and forecast a Collingwood win.

[Note: if anyone is comparing the ranking points and changes with those from the end of last year, I made a slight change to the formula at the start of this year. I had been using a team’s past seven matches to calculate ranking points, but I have reduced that to six matches given the premature end to last season, and that it removes the last ‘proxy’ result for the expansion clubs. I may expand this out to nine matches as the season progresses, given the number of rounds this year… but let’s see how far we get through the season first.]

AFL Statistics Series #4: Gaining Territory to Score

Back – way back – in my first post in this series, I said that one of the most important measures for an AFL team to win, outside of points for, was its metres gained.

As with points for, there is something obvious about this: a team that scores more is usually going to have more metres gained than its opponent. It is difficult to directly break down a player’s ‘points for’ though, outside of the goals and behinds they directly kicked, and most people would not think that is how to measure a player’s entire impact on their team scoring. Unless the player took it out of the centre themselves someone had to get the ball down to the player that scored.

We do though have a measure for the metres each player gains, at least recently, and with gaining metres being a necessary condition to a team scoring it may help us to better break down and estimate what a player contributes to their team’s score. I expect this to be a main component of my system of player value that I outlined last month.

Metres gained, as I said, are only a recent measure, available on Footywire back to 2015. However, another part of what I will examine below is that it is more or less an extension of older measures that have been with us longer and are more accepted and understood, like inside 50s. That is, valuing inside 50s is similar to valuing metres gained, and vice versa.

I stress here though that I am not heading towards the conclusion that the most valuable players are those that gain the most metres. For a player to gain metres they either need to win the ball themselves, or someone needs to win it for them. When we get to examining the value of winning the ball though, we should be able to express that too in terms of how it contributes to a team gaining metres, and ultimately generating scoring shots.  

Gaining metres leads to scoring – unless you don’t keep the ball

First off: how much does gaining metres tend to lead to more scoring? There is clearly a positive relationship (see chart below), but not an acute one. For example, the Tigers gained by far the most metres in the 2020 home and away season, but were fourth in scoring shots (and also in goals).

Gaining metres by itself of course does not necessarily result in scoring – some teams may play ‘aerial ping pong’, kicking it back and forth to each other without anything crossing the goal line. The next relationship to consider then is that between metres gained per turnover and scoring shots (see chart below – we could also add in team possessions that result in stoppages, but it doesn’t change a lot the point I’m making for now).

Now that looks a bit stronger. The Tigers had more metres gained, but also more turnovers than other top sides, which stifled their scoring capacity relative to the territory they gained. (They also create more opposition turnovers as well, but we’ll get into that when we look at the ‘points against’ side of things.)

Compare this to, say, the relationship between scoring shots and contested possessions, given that contested possessions are often cited as an important factor in winning (see chart below). The relationship is, basically by definition, not as direct as metres gained – teams can advance the ball towards goal through either contested or uncontested possessions, and they differ in the number of possessions they take to get downfield. That is not to say that contested possessions do not matter, but it is within the context of how they ultimately help a team to gain territory that we may get a more accurate overall picture of how they influence a result.

Again, while from a team perspective gaining metres without turning the ball over is what leads to scoring, that doesn’t necessarily mean it’s that simple to measure a player’s value. The players that have high average metres gained per game and high metres gained per turnover tend to be general rebounding defenders like Daniel Rich, Tom Stewart, Shannon Hurn, Jordan Ridley, and Nick Vlastuin. These would still be among the better players in the game, for their ability to kick long and find targets, but they would also often be assisted by teammates, and we’d also need to consider this to get a better picture of their value.

Approximating metres gained with other ‘territory’ statistics

As said above, metres gained are only available to us (I think) back to 2015, but there are other statistics that have been around longer that also capture the concept of gaining territory. Inside 50s and rebound 50s – which measure how many times a player moves the ball int their forward 50 arc, and out of their defensive 50 arc respectively, are available on AFL Tables from the 1998 AFL season onwards.

Between inside 50s, rebound 50s, and scoring shots, we have three measures of when a player has moved the ball past specific points on the field. How well do these relate to metres gained? Fairly well, with players that have high average metres gained tending to have high average totals if we combine their inside 50s, rebound 50s and scoring shots (see chart below). For AFL players who had a minimum of five games in 2020, their estimated average metres gained per game rose on average by:

-          about 60 metres for each increase of one in their average inside 50s;

-          about 50 metres for rebound 50s,

-        about 20 metres for goals (noting that a player can record both a score and an inside 50 in the same kick), and

-          about 30 metres for behinds, also keeping in mind the opportunity cost of missing the goals. 

This approximate estimate of metres gained is within plus or minus 30 average metres per game for about 80 per cent of players. (Of course, I don’t know exactly how metres gained are calculated, so one possibility is that what I’ve just done is re-create something close to the method by which they are worked out.) Without looking in detail at the position of players, it does seem to underestimate metres gained more for ‘outside’ players and wings, and overestimate them more for ‘inside’ midfielders and rucks. However, that doesn’t necessarily count too much against this method of approximating metres – the estimate may actually be a better measure of player value than the ‘raw’ measure, since getting the ball inside the 50 metre arc could carry slightly more value than just the ‘raw’ metres gained would suggest.

What this method does miss though is movement of the ball by players between the 50 metre arcs. Can metres gained also be fairly well estimated from the ways of moving the ball – i.e. kicks, handballs, marks, and hit outs? Yes (see chart below), though as it turns out, at a player average level at least, only kicks contribute positively to the estimate at about 30 average metres gained for each kick. Marks in particular contribute negatively (minus 10), perhaps indicating that players that take more marks are more likely to do sideways or even backwards kicks, that is, chip the ball around the backline.

What this means though is we have come full circle from the more ‘wonky’ measure of metres gained all the way back to a player’s ‘basic’ statistics. A player’s kicks were always a way of indicating to us how much territory a player was gaining. Metres gained may (I believe) measure this more precisely, but as a concept it’s just taking us back to how often and by how much a player is responsible for moving the ball towards the goal posts.

Probably the next part will be to work out which players assist in or set up the gaining of metres. For that, we’ll look more closely at ‘possession chains’ to see which types of possessions are more likely to lead to certain other types.

AFL Rankings: Finals 2020

Richmond may not generally be as dominant as other recent ‘dynasties’, but they are when it counts most.

For the second year in a row, and the third year out of the past four, Richmond are premiers. The Tigers beat Geelong by 31 points in the first ever AFL Grand Final played at night, and the first ever played in Queensland. Directly after their win, much talk turned to Richmond’s place among the very best teams of this (and last) century.

First off, I’ll note that Richmond didn’t even quite end up as the #1 team on my rankings. The Cats held that positon after their huge wins in the finals against Collingwood (68 points) and Brisbane (40 points). However Geelong cannot seem to get past Richmond, having now lost to them in the finals in each of the Tigers’ premiership years.

This suggests that Richmond may not be quite as dominant as other recent ‘dynasties’. The three teams that have won three flags in quick succession in the 2000s were the Brisbane Lions in 2001-03, Geelong in 2007, 2009, and 2011, and Hawthorn in 2013-15. Each of those teams had a higher winning percentage, and higher percentage of points for compared with against, than the Tigers in their premiership year (see table below) – particularly the Cats and Hawks. The Cats and Hawks also perhaps had tougher opposition in contention for the flag, based on the records of other top four sides in their premiership years. Indeed, the ‘all-conquering’ Tigers are not that far above the team that got eliminated in the first week of the finals for three straight years in 2013-15.

Where these Tigers match the other successful sides, and separate themselves from their one-and-done counterparts, is their record in finals. With the exception of Port Adelaide in their preliminary final this year (they also lost their qualifying final to Brisbane), and arguably Geelong last year, they have generally gone through teams like a wrecking ball, with an overall percentage of 164.6%. That is slightly below the Cats, but better than the Hawks and Lions, and of course way better than the Tiger sides that could not make it into the finals’ second week.

So it seems that Richmond’s dominance comes from playing at or near their best when it matters most. In that regard, they are perhaps most similar to the Brisbane Lions teams that were held in such awe as they rattled off three straight flags early this century. The Lions never finished on top, as the Tigers have not done either – at least during the years they won premierships – but they crushed opponents in the finals, with six wins of 40 or more points. Like the Tigers, they seemed to learn how to time their runs just right.

Of course, Richmond’s stepping it up a gear at the business end of the season is emblematic of the performances of their best player, Dustin Martin, who has been the best performer in both the finals and Grand Final of all three of their premiership years. Martin has kicked an astonishing 24 goals in the Tigers’ ten finals in those years, whereas he has averaged closer to a goal per game during the home and away rounds. In the two of those finals he kicked the least goals he racked up 19 inside 50s. He led the team in contested possessions in all three of their Grand Final wins. There has probably never been an attacking force in big games quite like him.

So another season over, and we’ve ended up where we began – the Tigers will remain the reigning premiers next year. As a Richmond supporter, this pleases me greatly; still I hope we never see another year like this, and it is a happier season overall for players and fans (except perhaps this year’s lucky fans in Queensland) in 2021.

AFL Statistics Series #3: Building a System of Player Value

Almost two years ago, I started a series of posts about statistics in the Australian Football League. The aim of the posts was to set out how much I thought the various statistics in the AFL mattered, and how they connected, in telling you how a team was able to score and stop its opponents from scoring.

Part of my interest was also to come up with a system for rating players’ contribution to winning. The AFL Player Ratings system is too complex for an amateur analyst like me to replicate. Other systems are simpler, but are less intuitive in how they measure value. If possible, I would like my system to be in terms of the player’s contribution to the team’s and opposition’s actual scores. While that would be the most elegant way of quantifying value, I’ll go into further below the potential challenges that presents.

After months of false starts and ends, here is where my thinking has led to. This isn’t the system, but it’s how after much consideration I now plan on building it.

Get the ball, move the ball, aim for the goals

As I said in my first post, the aim of Australian football is to get the ball towards and through your goals, and to prevent your opponents from doing the same. All AFL statistics can be classified in terms of these objectives.

The simplest part to capture is the result of a player’s shots on goal. Goals and behinds are easily available, and therefore we can easily see a player’s points per scoring shot compared with the league average. I went through in my second post in this series how this could be measured, and I don’t expect to talk about it too much more in the rest of these posts. The measure can be improved by using expected score, and accounting for kicks out of bounds, but since I’m aiming for a system that can mostly work with simple, historical statistics I’m probably going to leave this part alone and focus most of my exploration elsewhere.

Most of the work will be in the part that is often used to measure the impact of players – how often a player gets the ball and gets it down the field to help their team score (see table above). In other words, how much does a player do to generate shots on goal for their team? Relevant to this will be measures of:

- how often a player gets the ball – i.e. possessions;

- how the player got the ball – did they win it themselves in a contest, or were they part of a chain of possessions for their team; and

- how far they get the ball down the field – which can be measured by metres gained, but also by inside and rebound 50s, and also even just by kicks and handballs. 

The other part of ‘attack’ is how many points a player costs their team by not retaining possession. Turnovers are the most commonly cited statistic in relation to ‘wasting’ possession nowadays. However, many turnovers happen not through a player kicking the ball directly to an opponent, but by kicking to a contest which their team loses, so perhaps we will need to be a bit more nuanced in distributing the ‘blame’ for losing the ball.

Measuring a player’s contribution to defending, at least conceptually, should be the reverse – what does a player do to prevent or not prevent the opposition generating scoring shots? The common defensive measures – tackles, spoils, and intercepts – capture the former. Harder to capture is the latter, i.e. what does a player not do to prevent the opposition from scoring? Potentially this is part of why midfielders tend to occupy many of the top positions on rating systems – they are in more contests, and we can relatively easily measure the contests they win, but it is harder to measure the contests they lose.

Scoring is more variable than possession

As I said above, I would like my system to be in terms of the player’s contribution to the team’s and opposition’s actual scores. Sounds appropriate, right? However, many player rating systems do not have those relativities.

Consider the 2019 Grand Final (see table below). Richmond’s score was over four times greater than Greater Western Sydney’s. On the SuperCoach rating system though, Richmond’s collective rating points were less than 20 per cent higher. That is more similar to the difference in metres gained, and the difference in disposals was even closer – the Tigers had only one more disposal.   

Getting the ball towards goal for a score requires not only having possession, but also having the right sequence of possessions. Usually, it will not be just three consecutive kicks towards goal – it may be two kicks forward, one kick the other way, another kick back, a handball back the other way … until eventually a team gets the ball far enough forward that it scores.

What is the value of what the team’s players did then? Is it how often they got the right sequence of possessions to score? Or is it just how often they got the ball, and got it forward? What is the ‘true’ value of high possession-getters in ‘bad’ teams? That will be one of the questions I wrestle with in future posts.

AFL Rankings: Round 18 2020

Port Adelaide and Brisbane have the home finals, but Richmond and Geelong look to be slightly ahead as the leading contenders for this year’s premiership.

After a home-and-away season unlike any other, we have our eight finals teams. Port Adelaide finished on top of the ladder, where they had spent the entirety of the season. The Brisbane Lions took advantage of a season played largely in their home state of Queensland to finish second. The last three premiership teams – Richmond, West Coast, and the Western Bulldogs – are back for another run, along with recent finals stalwarts Geelong and Collingwood. Finally, a much improved St. Kilda team have made it back for the first time in nine years.

Who do the rankings like for the premiership? In what may be good news for Cats fans, or bad news given how last year turned out, the rankings are picking Geelong for the second straight year (see table below). It is not a pick made with much confidence however, with predictions for several very close matches.

Heading into the finals series the favourite for the premiership is Richmond, followed by Geelong and Brisbane. In response, the ‘hip’ pick for the flag this week seems to be Port Adelaide, with people arguing they were being underrated given they had spent the whole season on top.

While this year’s fixture first seemed like it would be more even than usual, with each team playing each other once and many games at neutral venues, in reality it gave some teams a massive net home ground advantage. (Not the greater home ground neutrality previously suggested on this blog.) As Max Barry at Squiggle has pointed out, with home ground advantage based on state, Port Adelaide played six games with home ground advantage this year, compared with two with a disadvantage. Things were even better for Brisbane, although the exact count of how many games they had an advantage in depends on how one treats their away matches on the Gold Coast.

In contrast, teams like Richmond and Collingwood had no games against non-Victorian sides before they had to leave Melbourne and move into their hubs. Some opposition supporters (and media) rarely miss a chance to point out how many games the Tigers and the Magpies usually get at the Melbourne Cricket Ground each year and therefore how little travelling they do, but really the majority of those games are against cross-town teams that have had to do about as little travel as they have. The ‘real’ home ground advantage comes when playing teams from out of town.

Port Adelaide is definitely a decent chance to win the flag. The 2020 Grand Final however will not be played in Adelaide, but in Brisbane. Of the top four sides the rankings currently do not favour Port in a Grand Final match-up with any of the others – Geelong and Richmond because the rankings consider them stronger teams, and Brisbane because of the Lions’ home ground advantage. The Power are still considered a good chance to make it to the Grand Final though, with potentially two home finals.

Overall though, even Richmond is less than a 30 per cent chance with the bookmakers, and it would not greatly surprise to see any of the top four teams hoist the cup at year’s end. Really, would a winner from outside the top four be a surprise either? Perhaps the main surprise of this year’s AFL finals series – albeit a pleasant one – is that we got to this point at all.