Attacking evolution is a mirage

Zelím Nel

Tries are being scored at an unprecedented rate in Super Rugby history, but what if I told you that high-volume attack is, as a tactic, the least efficient it has been in five years.

Now might be a good time to stop reading if, 1) you don’t want anything to challenge what you know about the game or, 2) you don’t have at least 10 minutes to concentrate, like a hungry bear using a twig to fish for honey in a swarming beehive.

I see that you have chosen wisely!

Over the past five seasons, I’ve tracked Super Rugby results based on what tactical preferences, or “trends”, are used by winning teams. The method has been to log the tactical decision-making of the winning team, in every league match during each of the Super Rugby seasons spanning 2013-2017.

The “decision-making” bit is important to remember. I have not measured the quality of the winning teams’ execution; I’ve measured whether they’ve chosen to Kick, Run or Tackle, while Penalties Conceded reflects just how disciplined and/or aggressive they’ve been in applying the combination of those tactics.

So, in each match, I tracked whether the winning team recorded a number of Kicks, Tackles, Carries and Penalties Conceded that was equal to, or more than, their opponents. The answer to that question is either Yes or No (Y or N) in each of the four categories.

At the end of each season, that collated data reveals the percentage of league matches that were won by the team that, for example, kicked more than their rivals. There were 120 league matches in 2014 – if 80 of those were won by the team that kicked more in each of those matches, Kicking would have a success rate of 66%.

The same calculation delivers a result for each of the remaining three categories.

Next, I tracked the combination of those trends in each match. Let’s say the Stormers beat the Bulls 25-17 and put up the following stats line:

In the above example, the Stormers are the winning team. They Kicked more, Tackled more, Carried less and conceded more Penalties than the Bulls. They would therefore register a collated “trend” for that match that reads as follows: YYNY

Obviously, contingent on variable match stats, there are 16 possible combinations of those four categories in matches that don’t end in a draw – such as YNYN, NNNN, YYYY.

Congrats on making it this far, soldier – you must really love your rugby!

At the end of each season, the total number of matches credited to each of the trends delivers a Win Rate for that trend. In 2014, 14 of 120 league matches were won by a team that adopted YNYN tactics (that’s Kick more, Tackle less, Run more, fewer Penalties Conceded) for a 12% Win Rate.

The purpose of this exercise was to identify which combination of tactical decisions is the most successful and, along the way, it showed how the expansion and dilution of Super Rugby in 2016 has created a tactical mirage.

Let’s start in 2013.

There were three draws in 2013 and, of the remaining 117 league matches, 88 were won by the team that kicked more (or at least as much) as their opponents.

That’s a 73% win rate (88/117), which made ‘Kicking More’ a bigger determiner of success than playing at home – 2013 homefield advantage: 67%.

Kicking More was also more effective, on its own, than Tackling More (56%), Running More (44%) and More Penalties Conceded (54%).

But how did Kicking More feature when viewed as part of a winning trend? Very well.

The shortest route to victory in 2013 was YYNY, or Kick More, Tackle More, Run Less, Concede More Penalties.

That trend accounted for 34 victories (highlighted in green), or 28% of all wins – twice as efficient as the next-best trend, YNYN (in yellow).

In other words, a team was twice as likely to win by Kicking More, Tackling More, Running Less and More Penalties Conceded, than by Kicking More, Tackling Less, Running More, Fewer Penalties Conceded.

The Kick and Tackle More combination remained dominant until this season:

As you can see below, while YYNY has remained the most successful trend over the past five years, it is this season at its lowest percentile (21%).

Conversely, the trend that Johnny Pundit prescribes as the way to go, namely Kick Less, Tackle Less, Run More, Fewer Penalties Conceded (or NNYN), is peaking in 2017:

This would seem to indicate that the experts who have been declaring that “the game has changed” are correct. Attack has overtaken defence and winning is now all about who is most committed to “playing ball-in-hand rugby”.

Here’s the kicker, if you’ll excuse the pun: as a tactical decision, Running More in 2017 is the weakest indicator of success it has been over the past five seasons.

As the SuperBru guys, and betting agencies will attest to, the results of Super Rugby matches are more predictable now than ever before. In my opinion, that’s mostly down to the expansion teams, added to the Super 12 since 2006, not having the quality player stocks to be taken seriously.

It’s relevant to this report because the gap in quality between the competition’s best and worst teams has grown rapidly over the past five seasons.

Let’s use an extreme example to showcase the gap in quality between a team like the Sunwolves and a New Zealand heavyweight, such as the Crusaders.

If the Stormers played the Gold Cup champions Rustenburg Impala, they would get away with playing any kind of rugby they like, such is the variance in quality. Even if the Stormers didn’t kick the ball once in 80 minutes, they would beat the club side almost without exception.

To negate this dilution effect on the statistics, I filtered the match data to exclude that which is attributed to contests between teams of equal quality. The best way to do this was to only factor in the trends from league matches between the teams that went on to reach the playoffs – for 2017, I’ve included the top eight teams (on log points) after Round 14.

In 2014, six teams made the playoffs: the Waratahs, Crusaders, Sharks, Brumbies, Chiefs, Highlanders. During the league phase, combinations of those teams met 18 times with the following results:

In those strength-versus-strength matches, Kick More hit 83% success rate (winning 15 of 18 matches), Tackle More won two thirds of those matches (66%), Run More won half of the time (50%), and More Penalties Conceded won 56% of matches.

Now let’s compare the unfiltered data with the strength-versus-strength numbers:

With the exception of 2013, the effectiveness of Tackle More increased significantly when the data was filtered for Strength-versus-Strength.

This year, Tackling More increased in effectiveness by 12% after the filter, from 46% to 58%. Meanwhile, Running More dropped 11% from 53% to 42%.

At 42% in 2017, Running More (against a team of equal ability) is the least efficient tactical decision it has been in five seasons. This flies in the face of the common belief that rugby has “evolved”, and that possession-hungry teams are rewarded while high-volume defence is an old-fashioned idea that doesn’t win games.

The trends data also rubbishes popular opinion. In strength-versus-strength matches, YYNY (Kick More, Tackle More, Run Less, More Penalties Conceded) jumps from 21% effectiveness to 32%.

Meanwhile, Johnny Pundit’s version of how the game should be played drops from 14% effectiveness – when teams like the Sunwolves and Cheetahs are in play – to 5% when it’s strength-versus-strength.

The conclusion is that, while teams are constantly finding new ways to kick, run and tackle, pro rugby matches are won, in the most part, by teams that limit their exposure to risk and capitalise on poor risk management by their rivals. This includes prioritising territory above possession, and defensive imperatives above attacking imperatives.

The increasing popularity of amateur rugby ideals in recent years, especially in Australia and South Africa, has only ramped up the rewards for those who are savvy enough not to play to the media, but to play to win.

- Zelim Nel

Let's chat

  • Quent Still

    Would be interested to see another category for ‘passing’. How many passes are completed.per team in a match. I think it was Brendan Venter who wrote an article post last world cup showing the the top 3 teams passed almost twice as much as the other teams. Not sure if this has any affect or is measurable. It does shed a slightly different light on ‘ball in hand’ (SA style hit up vs. Maybe Kiwi passing into space)

    • Freddie MacFarlane

      I have always been intrigued as to why so many passes go behind players during matches, this normally holds up the attacking play by a split second sometimes killing the move.

      I cant imagine that teams practice that way so I would assume that it has something to do with the pressures levels of the match itself.

      Building on that passing stat you were talking about, I would like to see the quality of the pass, looking at if it was behind, low, high or out in front, I would be interested to see how the quality of a team’s passing affects say
      their line break stats.

  • Lou Beukman

    Great article. I do agree Quent and would also like to see off-loads brought into these analysis. What I take from this article mostly is that if your fly-half has great skills but can’t manage a game, your going to struggle to win against the strong teams!

  • Max

    Look like the Stormers and Bulls of 2007-11 used the correct gameplan to beat the NZ teams. but we had to play like the ” perfect teams of NZ”. Their gameplan would have put them on top of the tables if they had stuck to it. The Lions tour to NZ is the perfect example. The bottom line is that there isnt a place for 15 man running game in rugby. The balance of attack and defense is what rugby is all about. I enjoyed the Lions rugbygameplan against ABs because they used the basic rules of rugby(PRESSURE, SUPPORT AND PLAY) to counter ABs on attack and defense, where the ABs to their frustations, came off second best.

  • humblepie

    Dear Zelim,
    many years ago someone published a book with the title “How to lie with Statistics”. Since then I became sensitised how this tool is employed. I lately observe an obsessive trend to report rugby stats such as “how many meters each player ran” as if it is gospel! First of all, the poor person that collected this stat didn’t have a hope in hell to reflect on this accurately and it results in a “garbage in – garbage out” scenario with meaningless conclusions. The same applies to collecting a single, collective “Kicking” stat. This single stat will include vastly different scenarios such as:
    1. So called exit plan kicking employed by conventional teams when in their own 22
    2. Kick passes such as employed by the more modern teams
    3. Kicking with the intent to score a try
    Clearly the first scenario should not be grouped in the same stat as scenario 2 and 3. The intent is vastly different. The first is employed by teams that are only willing to play in certain areas of the field. The latter is employed by teams with an aggressive, fast paced mindset.
    I would rather like to see stats such as
    1. %possession
    2. teams that follow exit plan strategies
    correlated with winning success. I say this very hesitantly due to the limitations on collecting these stats accurately.
    For now, I unfortunately do not attach any value to your conclusions.
    My penny’s worth

    • Andrew

      Great, point Humble!

      Although I don’t disagree with Zelim’s assertion that top teams kick more. It’s, as you point out, the type of kicks they use. The type of offloads employed and most importantly the accuracy of attacking play.
      Without having stats in front of me I am willing to bet that the Lions, Crusaders and Hurricanes all average less phases per try than the other teams. So often the teams they play tend to do less defending but miss more tackles. By the same token when the top three teams kick or turn over possession they have better defenses so opposing teams will go through more phases on attack.

  • Amien

    Its like nfl stats show so much bt game time means so little. Morne steyn stats will show you he is brilliant bt as a flyhalf he was average bar his kicking ability. Stats will show carlos spencer wasnt so good bt we all know what a skillfull player he was.

  • Iamwhy

    What was the % of successful kicks for points in 2017 relative to the other seasons?

  • Fanie

    One can only laugh if you attach any value to this article.

  • Johan

    Interesting read.

    Stats is always interesting and most scientific research are based on generalization and stats.

    However OBSERVER BIAS is a big issue when using stats.

    Let me prove it.

    kick, tackle, run, pen
    In keeping with the above it would be interesting to see:
    Kicking meters, tackle% completed, meters run, turnovers won

    Penalties’ stats is always close, in the 50% area, lets see successful turnovers won vs win rate

    I don’t have access to the database but just comparing a few teams for this year that would be:
    Lions vs Bulls (best of SA vs worst of SA)
    kick meters: 7573 vs 8130
    tackle%: 85% vs 85%
    meters run: 12576 vs 10208
    turnovers won: 65 vs 40

    Crusaders vs Blues (best of NZ vs worst of NZ)
    kick meters: 7122 vs 7655
    tackle%: 87% vs 86%
    meters run: 11800 vs 11385
    turnovers won: 39 vs 47

    From this you can see that for these 4 teams (SA and NZ) kicking less and running more wins, completely the opposite of what this article says.

    • Freddie

      I like this point, you can always

      This is the frustration i am finding with the stats we have available, there is very little context. The stuff that would allow you to better analyse team and player effectiveness would be measurable context data:

      Was the player kicking under Pressure?
      Who is kicking and when or from where?
      Was the player running against broken or set defense?
      Was the tackle attempt in broken or set defense?
      Was the pass behind, above, low or in the sweet spot?

      Also for me the outcome of different actions would help identify the efficacy of the actions of the teams, like a kick made a Gain of 40 meters, or a loss of 20 meters due to a counter attack or return kick. Its all good and well to say a player ran more or kicked more but the important info is why and what happened as a result.

      This type of context, that can be easily measured, will give greater accuracy in analysing the efficacy of different team’s overall strategies.

      As your example above shows that without that context you can find stats that could favour either side of this discussion.

    • Johan

      Also note that the Bulls have kicked further and run less meters when compared to the Blues.

      Which shows that the Blues have kicked less and run more and have won more.

  • Cobus Brits

    Very insightful. Rugby has become so proffesional that lines are seldom broken anymore during phase one except for kicks and off-loads which means teams are looking for cracks in the game to slip through. Tiresome multiphase attacking play might bring some success but it is still the higher kicking teams that defend well that reap the best results.

Comments are closed.