Try, try and try again: what analysis of the All Blacks’ chances at the RWC can teach businesses about big data

Another global sporting tournament, another round of predictions about who’s likely to win. At Dot, we find the juxtaposition of sports and stats exhilarating, especially when it comes to rugby. However, what we are seeing emerge in the media around the 2015 Rugby World Cup is typical of many applications of analytics in the business world. Analytics is about actionable insights. Articles that show that the All Blacks are favourites to win the tournament offer insight; but there is little that is actionable, unless you are using the stated probabilities to determine if there is value at the local betting agency …

Any rating system that you run will show that the All Blacks are the best in the world. That happens when a team only loses three games since the last tournament. But, as a spectator, what does that mean for me? And, as a coach, how would I instruct my team to compete?

If we look at this simply, the All Blacks are favourites because they tend to win more than anyone else. The reason for this is because on average the All Blacks score more tries. However, is that attacking prowess relevant in the knock out stages in a Rugby World Cup hosted in the Northern Hemisphere?

Analysing knock out games at all Rugby World Cup tournaments shows teams that scored more tries than the opposition won 71 percent of matches, whereas teams that kicked more penalties only won 46 percent of games. In the Northern Hemisphere, 75 percent of games were won by teams scoring more tries than the opposition and only 42 percent of games were won by teams kicking more penalties. Clearly, the ability to score tries is critical to tournament success.

Now, looking at the average number of tries scored in each game played between the top-flight nations in 2015, the All Blacks sit on top with 3.75, closely followed by the English on 3.13.  

That little piece of analysis may seem overly simplistic, but if we compare a rating of teams based on the metric of average tries scored per game, as described above but over a two year period, with the more complex rating system outlined by Dr Niven Winchester, we get a correlation of 0.94. That means the metrics are very similar in what they measure, but one approach is much easier for an average punter to understand and track at home, and shows that sometimes a simple solution will suffice.

In the week leading up to the start of the tournament, Stuff carried an article titled ‘Rugby analyst predicts semi-final anguish for New Zealand at World Cup‘. From an analytics perspective, the content in that article takes us to the next level by attempting to build an actionable strategy to prevent the All Blacks winning. The concepts of slowing ruck ball, defending the counter attack and winning the battle for territory can be operationalised and measured. For instance, a metric for defending the counter attack could be the number of missed tackles on All Black fullback Ben Smith when he returns a punt. However, that line of investigation neglects one important aspect: you still need to score more points than the All Blacks to win, with history showing that needs to be through tries. That said, there is no need to dredge up painful memories of Campese in ’91 or the French scoring burst after half time in the ’99 semi. 

What it all means is that the All Blacks are the favourite to win because of their attacking process. Jockeying for field position and trying to kick penalties and drop goals won’t defeat the All Blacks in the Northern Hemisphere. Any team that hopes to knock them off will disrupt ruck ball and nullify any counter-attack whilst trying to score tries themselves. That is why England is such a threat to the All Black hopes in this tournament.

  • Dr. Paul Bracewell, founding partner and chief data scientist at Dot Loves Data.

About Author

Comments are closed.