Does the week off help?

With the Rugby League play off series, or finals series for those down under, now in full swing I thought it would be interesting to see how much the rest up next week might be of benefit for the teams that fought hard in week 1 to win that right.

Super League has had play offs since 1998 and the NRL has been the premier Southern Hemisphere competition since that year too, following the end of the 'Super League War' down there. In both competitions the play off format has changed around a bit in the 15 seasons that have been completed since 1998. However, one constant is that at some point one of the rounds has seen one or more teams earn a week off.

The number of teams enjoying a rest and the point in the series this comes has varied, particularly in Super League, but this rest at the 'business end' of the season is perceived to be a benefit. However, there is the counter argument that a team benefits from a run of games at this point in the season and form may be better than a interrupting this with a week off.

The table below shows how teams have responded to the week off since 1998 in Super League and the NRL when up against unrested opponents. I've also added the numbers for the NFL as a point of comparison - obviously this is a different sport, but I think the physicality and intensity is comparable.

In the NFL the week off is awarded purely on the regular season performance - the week missed is the first week of the play offs. In the NRL the rest week comes as reward for league position and a win in play off week 1 - in 1998 there was a 10 team play off with two teams earning a week 1 rest, but since the rest has been week 2. In Super League, 1998 to 2001 saw a week 1 and week 3 rest for one team, 2002-2008 saw two teams miss week 1 then play each other in week 2 (so no advantage over opponent) with the winner of that getting a rest in week 3 too, 2009 onwards sees the two best ranked week 1 winners earn a week 2 rest.

As you'll see from the above table, a week off in Rugby League doesn't really offer a great advantage over the following week's opponents who haven't had a rest. In Super League more teams have actually been beaten after the rest than have won, and in the NRL rested teams are only just ahead in wins when playing an unrested opponents.

Conversely, in the NFL a victory by the rested team is more than twice as regular than a defeat. What is interesting about this is the rest is always week 1, before the teams have played a knock out game. Looking at Super League, when the rest was in week 1 between 1998-2001, three of four times the rested team won (although none of two week 1 rested teams won in 1998 NRL). This does suggest that the most beneficial rest in terms of the next game you'll play is week 1 - across the three leagues in the two sports the total for week 1 rested v unrested is 44-22 in their week 2 match-ups (67% win rate). Week 2 rested teams are 20-16 in week 3 match-ups on unrested opponents (56% win rate). Week 3 rested teams are 5-8 in week 4 (the Grand Final for Super League 1998-2008 - 38% win rate in this small sample).

Where the rest is week 1, the rested teams have been consistently the better teams all season. When the rest isn't this week, there is a chance the best team doesn't get the rest. This is a likely explanation to why more week 1 rested teams win against an unrested opponent the next week than other rest schedules. The numbers also suggest you don't want to be interrupted once you start playing knock out play off games - you still want a rest, but you're better off getting your rest before you get into play off game mode.

All this has shown though is what you're likely to do in the week after you've rested, when you come up against an unrested opponent at this time of the year. Surely the bigger question is what impact a rest has on your overall chances of ultimately winning the title? The following table shows the same leagues over the same years indicating if the Grand Final / Super Bowl winner enjoyed a play off rest or played every week.

Here, things look different. 10 of the last 15 Super League and NRL Grand Final winners enjoyed a rest week during the play off run (67%). 8 Super Bowl winners enjoyed a rest, but this is only 53%, so the eventual winner is almost as likely to have carried momentum through.

Rest is more beneficial in Rugby League than American Football. Play off momentum seems to be less important in the long-run, probably because in Rugby League it isn't truly knock-out all the way - better regular season teams get second chances, whereas in NFL after the rest you have no second chance.

In the week 2 rest week systems, those played in Super League and NRL now, it switches to 10-8 (56%).

So, hard fought wins for the Warriors, Wolves, Rabbitohs and Roosters means what exactly? Well, they are slightly more likely to win in week 3 after their week 2 rest, and having rested, they are slightly more likely to go on and win the the Grand Final than if they hadn't rested at all. The week off helps, but not that much.

Based on the very recent history, the week 1 wins are likely to put the teams from Sydney's South and East in a better position in the NRL than Wigan and Warrington are in the Super League.

One thing is for sure, when it comes to the play offs, things aren't as predictable as you might think in Rugby League and the current system - that is under much criticism from fans - enhances that unpredictability. The top teams that earn a rest are still more likely to come out on top, but being the best in the regular season doesn't add that much weight to your play off winning chances - that's a story for another day in itself though. For now, enjoy some exciting and intense knock out Rugby League with the best teams, rested or with momentum, going head-to-head throughout. It should be good!

Competitive balance in Rugby League - NRL compared to Super League

We all know that the NRL is better than the Super League - its more competitive, more intense and games are of a higher quality...so they say. However, after reading plenty of books recently where the accomplished facts of other sports have been widely challenged, I've been wondering just how much that holds true when we play the numbers game.

Personally, I don't dispute that it is a true generalisation based on hours of viewing games from both sides of the world each and every weekend, but I thought it was worth having look. I was especially spurred on after listening the BBC Manchester's Rugby League phone-in a few weeks ago when a caller came on and challenged this way of thinking using a very small data set. He basically referred to how many draws we've had in Super League this year and gave the results from Round 10 of this year's NRL campaign where there were three big one-sided results (there was also a draw that I think went unmentioned). Yes the average winning margin that week was 24.5 points, but I would note the 54 and 62 put up by Souths and Penrith respectively are two of only four 50+ scores seen in the opening 12 NRL rounds this year - compare that to Super League which has seen two scores over 80 this year (more than a point a minute!) and in Round 10 of its own 2013 season saw three 50+ scores, plus a 48, to give an average winning margin of 33.9 points.

These are clearly two freak weekly rounds and far too small a sample to draw conclusions from, so I've looked at the outcome of all games during the last ten seasons (2003-2012) and used the normal measures of competitive balance I've used in previous blogs to draw comparisons and conclusions from - match winning margins as a short term measure of how close games are, win percent ratio as a medium term measure of how close regular seasons have been, and a simple count of how many winners there have been to indicate the long term competitive balance.

I've put together a few tables to summarise the data. The first two show the numbers as they relate to each season for each league, and the second two show how each team has done when averaging out their performances over the ten years. Gold Coast have only played in six of the ten years, and I've only included Super League teams who've played in at least six seasons.

The numbers above show that the average regular season game is notably closer in the NRL, with the Super League average 17.8 being 23% more than the the NRL's 14.5. Every respective season in NRL saw closer games than its Super League counterpart and no Super League season had an average winning margin as low as the NRL 10 year average, with no NRL season as big an average as the Super League overall average. Clearly, the general position is that the outcomes of individual games are closer in the NRL than in the Super League. NRL games are therefore, in general, more competitive. However, it is worth noting that when it comes to the playoffs, where the best teams that year face each other in a (sort of) knock-out competition, NRL games aren't as close as their regular season games and are basically as close as Super League playoff games. This suggests Super League playoff games are a step up in competition, whereas NRL playoffs aren't.

The win percent ratio for the NRL is also notably lower than Super League. The closer to zero this figure is, the more competitive the season has been - i.e. the closer all the teams were to each other in terms of games won/points earned. Only one Super League season was closer than the NRL tended to be at the end of the regular season, and only two were closer than the least competitive NRL season. Particularly early on in the sample, the Super League was not close between the better and the worse teams. The NRL figures actually compare very favourably when compared to other sports too when I've looked at similar research I did a few years ago for the English Premier League, the NFL and Major League Baseball. It is a very competitive league season in, season out.

In the longer run, when you consider the amount of different teams winning titles and getting to Grand Finals, the NRL again looks more competitive. In terms of 'Minor Premierships' the numbers are close - five in Super League and only one more in the NRL. The share of wooden spoons was also marginally higher in the NRL. However, nearly twice as many teams have made it to the Grand Final down under as have managed the feat in Super League. Melbourne have been a fairly dominant force, although a big astrix is next to their name certainly before 2010, but only Sydney in 2003 and 2004 played in consecutive NRL Grand Finals, whereas every Super League final has featured a team who appeared the previous year other than in 2006. Long term dominance by a small group of teams has certainly been easier in Super League, which the following tables will also show.

 In looking at the above tables its worth bearing mind that the NRL had 15 teams 2003-2006 and 16 from 2007-2012, Super League had 12 teams for 2003-2008 and 14 in 2009-2012. NRL has always had 8 teams making the playoffs, Super League had 6 2003-2008 and 8 2009-2012, although overall teams have had an almost equal opportunity to make the playoffs regardless of which competition they were in - all other things being equal teams would make the playoffs around 50% of the time.

Clearly all other things aren't equal, but they are closer to being equal in the NRL, where every team has made the playoffs at least twice and no team have been in the playoffs every season (Melbourne would have achieved this but for the penalties of cheating leaving them bottom without escape in 2010). Seven of the NRL teams make the playoffs between 40-60% of the time - the figures you would expect for all teams in an closely competitive league structured this way. In Super League this figure is three. Two teams have only made the end of season party once, both having played in 8 of the 10 seasons.

On average, Leeds, St Helens and Wigan will finish in the top four, Warrington and Bradford in the top six. Only three NRL teams have averaged a top six place (Melbourne, Brisbane and St George) and no team a top four place.

During the ten years, five NRL teams have 'won' both the minor premiership and the wooden spoon, which has never happened in Super League - Catalan Dragons are the only team to finish in the top three and also finish bottom of the table. Every NRL team has finished as high as fourth during the regular season, whereas four of the regular Super League clubs haven't managed this. Every NRL team has finished in the bottom half of the table, but Leeds and St Helens never have in Super League. Taking things further, on 8 occasions an NRL Grand Final team has failed to make the playoffs the following year - this has never happened in Super League.

All this backs up the generalisation that the NRL is more competitive than the Super League and that the teams are more equal. Games generally have tighter results, seasons tend to be closer between top and bottom teams, titles are distributed more evenly, and previous performance is less of an indicator on how well a team will do the following year.

Why is it more competitive? Well I would argue its because of the way the league is structured, with consistent membership and interdependence between the member clubs for continued growth and sustainability. Of course it also doesn't hurt its chances that it is one of the biggest sports in the biggest TV markets in Australia, compared to unfortunately being a minority sport in the UK and France.

....

That clears up the NRL being more competitive, but I've been thinking how I would analyse whether the games are more intense in the NRL than Super League using numbers. This I feel is a more difficult task. What you would be expecting is better defensive statistics across the board in the NRL to demonstrate this, but then that league also features the supposed best attacking players as well, so these factors may counteract each other. Would you expect less breaks because defence is better, or more because attack is better? More missed tackles or less? I've had a quick look at some numbers from this season so far and the missed tackle numbers on the face of it are very similar (about 8% of all tackles), NRL attacks tend to make more yards per carry and more tackle busts per carry, but less clean breaks per carry. A problem I face with this is I can't be sure the statisticians for the NRL competition use the same processes and criteria as the company who anaylse Super League games. Based on eye-test alone NRL seems more intense generally, but then Wigan v Warrington or Leeds v Huddersfield have been as intense as anything I've seen in the NRL this year. I've also seen some awful NRL games and some terrible Super League games - but the main difference is the bad NRL games are low scoring and the bad Super League games can be high scoring - bad NRL games tend to have poor attacking play, whereas bad Super League games are often due to poor defensive play. I'll go away and have a think at how I could look at this in more detail, I may need to dig deeper into my old toolkit that I pretty much left behind when I finished Uni...