Wallabies 2011 Involvement Rates

This year in my analysis of games involving the Wallabies I’ve included a new way of measuring the level of work by players in the game – the “Involvement Rate”.   

The Involvement Rate is calculated as follows:

(Number Of Carries + Number Of Tackles Made + Number Of Breakdown Involvements) / Minutes Played

I don’t measure the Involvement Rate as a way of determining who is the best player for the Wallabies or even who is the most effective.

Any attempt to use statistics to show who was the “best” would include too many subjective decisions.  Statistics can’t be used to answer those questions unless someone sets the parameters of how the “best” is defined.

When we consider who was the “best” we take into account factors such as who scored the most tries, who scored the most points, who scored the winning try or points and who made the key plays in the game.  These are just some of the factors we consider and the list would be different for each person.  If a winger scores three tries in the game by catching the ball and falling over the line on each occasion, is it the winger or the player inside of him that drew the last defender and made the pass that deserves the credit? If the goal kicker kicks a penalty goal from in front to win the game, is it the goal kicker or the tight head prop that won the scrum penalty that deserves the credit?  If we could all agree that a game was won on defence, was the best player the player who made the most number of tackles or the player who made the most number of dominant tackles?  Or was it the player who made three try saving tackles?

You’ve probably all heard about the movie “Moneyball” starring Brad Pitt – I saw it last night and really enjoyed it.  I’ve also just started reading the book and it’s very interesting to learn how the measurement of baseball statistics has evolved over time.  The key thing behind the whole Moneyball concept was the different way the Oakland A’s team valued players by placing more importance on the way batters were judged.  The data was there for anyone to use – the A’s simply decided that using the statistic that had always been used to measure the effectiveness of batters was not as important as another statistic.  Their choice made a lot of sense to me but in making the choice they broke with the way that every other major league baseball team measured effectiveness.

When the A’s looked at players judged on what they believed was the key statistic they found that the players who performed best in that statistic happened to be among the poorer performers in the traditional statistic and as a result were amongst the players with the lowest salaries in baseball.  They went out and recruited and drafted as many of those players as they could which meant they could keep their payroll well below the top teams in the game. 

Essentially they decided that it was more important that a batter got onto first base rather than how they got there.  They didn’t care if the batter ignored as many pitches as possible to earn more balls and a walk to first base or whether they hit it a magnificent groundball to the fence and achieved the same result – they just had to make it to first base.  Their fundamental view was that if they had a high percentage of batters making it to first base, they would be advanced around to home by the other batters making first base and this was just as valuable as hitting home runs, which occurred much less frequently and it was much more expensive to buy players with good home run statistics.  That may sound basic but viewing things that way was considered radical.  Ultimately the A’s got results using their method and other teams adopted a lot of what they did.

The factors included in the Involvement Rate are there because I see them as the most important factors in determining how much work players are doing.  The fact that I don’t include factors such as passes, line breaks, scrums packed or involvement in lineouts doesn’t mean that they are any less important. 

The basics of rugby are – you catch the ball and run at the defensive line, you tackle any player with the ball who comes near you and you hit the breakdown wherever necessary to make sure your team retains the ball or in an attempt to disrupt the opposition’s possession.  And of course you win your set pieces but I measure that in another statistic. The more of each of those activities you can do, the more important you are to a team. 

As the Oakland A’s did I’m using things we’ve measured in rugby for a long time but looking at them in a different way, or more correctly in a different combination but don’t expect an “Involvement Rate” movie.

When I measure involvements I only measure what I believe are effective involvements in each category.

When it comes to a carry I’ll only record it if the ball carrier commits a defender, so no credit is given to a back three player who catches the ball, runs 20 metres and then passes the ball to a supporter when they’re still 10 metres away from the defensive line.  Similarly no credit is normally given to a player that catches the ball and passes it on straight away.  The definition I’ve set allows me to include as a carry a situation where a player catches and passes straight away under pressure as the defence rushes forward – they may not go forward but they have committed the defender trying to crunch them and in my view deserve credit. 

Tackles are obvious – I only record completed tackles and there is no deduction for missed tackles.  The question of missed tackles is covered separately in another statistic. 

Trying to give you a clear definition in words of what I consider a breakdown involvement is a little more tricky – the best way I can explain it is that if a player makes a hit on the breakdown with their shoulder, gets in position over the ball to protect it even if the other players have gone to ground or been cleaned out or in defence a player makes an attempt to play the ball, whether they are the tackler or not, I’ll record a breakdown involvement.  I don’t record an involvement for any player that stands at the side of the ruck acting as a pillar or any player who leans against the ruck when there was no threat that required them to even be there.  This is a subjective area but I’m the only one that codes the games, so you get a reasonably consistent view.

The Involvement Rate measure is weighted in favour of forwards because across the 14 games in the Wallabies 2011 season, breakdown involvements represented 52% of total involvements and 82% of those breakdown involvements were by forwards.  The backs led in the carry category with 53% whilst the forwards completed 67% of all tackles.  As a result, whilst the average Involvement Rate for the team across 2011 was 1.27, the forwards average was 1.64, nearly double the backs average of 0.84.  That sort of split fits with the obvious thing about rugby – the forwards set the platform for the backs to work with.  The Involvement Rate is still relevant for backs but primarily when measured against other backs.

Of course there are also differences between players in different positions.  You’d hope you’re your halfback has a very low Involvement Rate – you don’t want them having to get involved in breakdowns and they’re usually in a covering role in defence so you don’t want them having to make too many tackles.  In fact in a perfect game you’d only want them to have carries. 

You also wouldn’t expect to see a front rower having an Involvement Rate close to a back rower.  If you do, it’s an indicator that something is out of the ordinary – either the front rower did a lot of work in the game or the back rower didn’t!  That’s where I see statistics as being important – they alert you to areas that need to be looked at more closely to see where improvements can be made or opportunities can be taken advantage of.

Another factor that is relevant is whether a player starts a game and plays the entire game or is replaced at some point.  Obviously when that replacement is made and whether it is the result of an injury will also have some effect.  You’d also expect that players coming off the bench would have a higher Involvement Rate for the period they’re on the field as they’re fresh.  The results looking at some of these factors are interesting.  The Involvement Rate of players who started in a game but were replaced was 20% higher than for those who played the full game.  The Involvement Rate of players who started on the bench and came on as replacements was 24% higher than for those who played the full game.

Of course the results are different for different players.  An interesting example is Radike Samo who had an Involvement Rate of 1.27 for the season but an Involvement Rate of 1.01 in the games where he played the entire game, an Involvement Rate of 1.47 in the games where he started and was replaced but an Involvement Rate of only  1.19 in the games where he came off the bench.

I’ve talked about the way I measure each of the involvements but I also need to tell you about the other key factor – the minutes.  The Involvement Rates I’ve published prior to today have been based on an 80 minute game.  In hindsight this involves some inaccuracy because the average time played in the Wallabies 2011 season was 81 minutes 22 seconds per game. 

In addition, after the Wallabies game against Ireland at the World Cup one of our contributors, ‘Rucking Good Stats’ pointed out that there were some major variances between the amount of time the ball was actually in play during each game and this would also affect the Involvement Rate.  He was quite correct – the average time the ball was in play in all the Wallabies games in 2011 was 30 minutes 22 seconds or just 37% of the time on the clock.  In the game against Ireland the ball was in play for only 22 minutes 58 seconds or 28% of the time on the clock. That meant that the involvements in the Ireland game were achieved in only 63% of the time in other games. 

That factor was significant enough for me to know I had to make changes to the way I was measuring the Involvement Rate but the difficulty was in working out how long the ball was in play not just for the whole game but also how long it was in play for each player in the period that they were on the field, whether they started and were replaced or whether they came off the bench.  Fortunately the raw data I collect is time based and I was therefore able to extract that data.  To make sure we’re comparing apples to apples I’ve gone back over the entire 2011 season and extracted that data and re-calculated the Involvement Rate for each game based on the minutes the ball was in play so that what you see here today is as accurate as I can get it.

Whilst the results in any one game are interesting, there are other factors that make the results in any one game hard to compare to results in another game.  Those include the weather, the type of game played, what the opposition does, possession shares etc.  The results over a full season are even more interesting.  The table below shows the Involvement Rate for all Wallabies in 2011.

I’ve attached a sheet that shows the Involvement Rate for each player in each game during 2011 so you can see how the measure varies from game to game.  Click here to download.

During the Wallabies bronze medal play off match against Wales in the World Cup I tweeted that I thought Scott Higginbotham was having a quiet game.  I had a few people tweet back who didn’t see it that way and thought he had done a lot of work in the game.  I made a mental note to look at that again when I analysed the match.  Unfortunately life got busy and it’s taken me until now to make comment.  Higginbotham’s Involvement Rate in that game was 1.23 and only Radike Samo and Tatafu Polota-Nau were lower in the forward pack.   Why did I see it differently to others?  Higginbotham carried the ball seven times in that game – equal highest in the forward pack with James Horwill (although Horwill was replaced earlier than Higginbotham).  We all know that Higginbotham is a very effective ball runner and he was again in that game.  I suspect the fact that he was prominent with his ball carries meant that the lack of work at the breakdown and in defence was easier to overlook.

For the entire 2011 season Higginbotham’s Involvement Rate was 1.55, a little below the average for forwards.  It was 1.48 when he played the entire game – again a little below average, 1.33 when he started and was replaced – again below average and 1.98 when he came off the bench – above the average for other forwards.  Does this mean he is best suited to a bench role?  Not necessarily – you’d have to look at the circumstances of each game to see if there were other factors having an influence.  He made 0.34 tackles per minute the ball was in play when he was on the field compared to the average for all forwards of 0.31 but there were 13 forwards who achieved a better result with Matt Hodgson topping that measure at 0.69.   Higginbotham was involved in 0.90 breakdowns per minute the ball was in play compared to the average for all forwards of 1.02 and there were 14 forwards who achieved a better result with Beau Robinson topping that measure at 2.44 in the short time he was on the field against Samoa.   Higginbotham carried the ball 0.31 times per minute the ball was in play compared to the average for all forwards of 0.23 and there were only three forwards with better results topped by Stephen Moore at 0.35 followed by James Horwill at 0.34 and Nathan Sharpe at 0.32.

I believe Higginbotham is at a point where he is the number one candidate to start at number 6 for the Wallabies (subject of course to form in Super Rugby).  I think the primary role of a number six is to be a ball runner and I believe that Higginbotham is the best ball runner the Wallabies can call on as whilst someone like Wycliff Palu is also a good ball runner, Higginbotham is so agile and quick he can get into space much better than many of the other back rowers.  But to fulfil his potential he needs to lift his work rate and that means he needs to play a little tighter.

I could comment on just about every player in the 2011 list but there are a few that stand out.  Tatafu Polota-Nau stands out for the wrong reasons – his season Involvement Rate of 1.00 is on par with many of the backs and is so low in comparison to the other forwards that I struggle to understand why he was even selected on the short End of Year tour and unless Stephen Moore was carrying an injury, I’m amazed that he started over Moore in the last game against Wales.  His Involvement Rates in the last four games of the Wallabies season have been very poor.  I can’t believe that if he’s being selected if he’s still carrying injuries but I can think of no other reason to explain what’s happened to his form.  And it’s not just that he’s not doing the work – he’s also not having the impact we all used to get so excited about.

Others that I should comment on in the forwards are Dan Vickerman who topped the 2011 Involvement Rate of any player with any substantial game time at 2.77.  He was only bettered by Beau Robinson at 2.78 in his short time against Samoa.  David Pocock came in at 2.20 but given the amount of time he played, that was in itself a remarkable effort.  The recent lift in performance around the park from Salesi Ma’afu compared to the other props is pleasing to see as his scrummaging has also come a long way.  Rob Simmons is another who needs to lift his work around the park – apart from the semi-final  against New Zealand he’s been consistently outperformed by the other locks.

In the backs Berrick Barnes and Anthony Faingaa have worked hard in recent games. Adam Ashley-Cooper’s numbers are not far away but the fact that he ended up slightly below the average for the other backs in 2011 reflects that it wasn’t his best season.  Rob Horne beat the backs average in each game – it would be nice to see what he can do if he can stay on the park.

My 2011 MOTS (Man Of The Stats) is a pretty obvious choice – Dan Vickerman.  To come back from such a long time out of the game at top level and do the amount of work he has done is a real achievement.

With the end of the Wallabies season I’ve also finalised a review of the key statistics for the Wallabies across the entire season.  Over the next few weeks I’ll publish articles analysing those statistics.

Merry Christmas to all of you and a G&GR New Year.