Rio 2016 Review: Conversions and Centering

Injera for the win... © World Rugby
Injera for the win… © World Rugby

This is the first installment of an Olympic review series, in which I will analyze team performance in different facets of the game. This particular post examines conversion kicking and how well teams centered their tries. I feel these skills are often overlooked or at least not well measured. To better measure these skills I will attempt to isolate them from each other and, when evaluating conversion kicking, do so relative to the difficulty of the kick.

Conversion Success
There were 175 tries scored across the 35 games in Rio. All were accompanied by conversion attempts, with the exception of Great Britain’s sudden death try over Argentina in the quarter-finals. Of the 174 attempted conversions, 113 were converted for an overall success rate of 64.9%. This rate is nearly identical to recent World Series averages.

Though this conversation rate is interesting, it tells us little about how difficult the conversions actually were. We know centered kicks are easier than wide kicks, but we don’t know what an expected success rate is at each location. To better describe the varying difficulties of conversions, I sorted kicks into 5m increments by width.

Width (m) 0-4 5-9 10-14 15-19 20-24 25-29 30-34
Attempted 54 18 13 14 22 19 34
Made 53 18 12 5 9 7 9
Success % 98.1% 100% 92.3% 35.7% 40.9% 36.8% 26.5%

We see that well-centered tries are converted nearly every time while wide tries are converted roughly a quarter of the time. We also see that the smallish sample size in a few of the sections creates some counterintuitive results. Though conversions in the 15-19m range were converted less frequently than conversions in the 20-24m and 25-29m ranges, for example, the closer conversions are surely easier to make. This discrepancy is almost surely due to sample size and we would expect success in the 15-19m range to rise with more data. To smooth out this curve I completed a logistic regression and adjusted to match how automatic centered conversions are. This provided a rough model of expected success based on width.

Conversion Points Added
I used this model for expected conversion success to more precisely evaluate how well teams converted their tries. Knowing a team made only half of their conversions, you might think that they kicked quite poorly, considering the tournament average was 65%. But their success rate may be low because they scored many tries near the sideline. Adjusting for the width of the attempt, we can express a team’s conversions success relative to what an “average kicker” would have done with the same kicks. My model of conversion success based on attempt width serves as this “average kicker”.

Since conversions are all worth the same amount of points, we can easily use our expected success rates based on width to calculate an expected points value. For example, if roughly 25% of kicks in the 30-34m range are successful, then an attempt from that distance is worth 0.5 points (.25×2=0.5). If a kicker misses from this distance, he cost his team 0.5 points on average. But if successful, he gained his team 1.5 points. Calculating all of the conversions this way, we get a total number of points each team gained or lost from their conversion kicking. This is shown in the chart below along with this calculation for only pool games where point differential mattered most.

Team Points Added
(all games)
Points Added
(pool play)
Tries w/
Con. Attempt
Points Added
Per Try
(all games)
France 3.4 2.1 14 0.24
Great Britain 3.3 1.8 13 0.26
Argentina 2.6 0.2 16 0.16
Fiji 0.6 3.7 26 0.02
New Zealand -0.2 0.8 17 -0.01
Spain -0.9 -0.5 7 -0.13
Brazil -1 -0.5 4 -0.25
Japan -1.2 0.5 15 -0.08
South Africa -1.5 -2.4 22 -0.07
Kenya -1.9 0.1 9 -0.21
USA -2.1 -0.2 17 -0.12
Australia -2.2 -0.3 14 -0.16

So a tip of the hat to Terry Bouhraoua of France and Tom Mitchell of Great Britain for helping their sides to the best marks. Conversely, Australia and the USA lost a few points when compared to the average. Seeing the USA at -2.1 we are reminded that they only needed one more conversion to advance to the quarter-finals at the expense of New Zealand. However, their kicking was near average in pool play and the majority of their lost points came from their two consolation matches. What may have sunk the Americans was Fiji’s excellent kicking, which included a difficult conversion from 29m wide against the USA in the final pool play game. The USA failed to connect on a similarly distant kick a few minutes later.

When dealing with one conversion and the Olympic quarter-finals, you remember moments like the above. Another was New Zealand’s first try in their opening match against Japan. Gillies Kaka, already in the try zone, completed a long pass over a defender to Scott Curry who dotted down under the posts. This eliminated the risk of a conversion from 20m wide and gained his team 1.2 expected points. The easily-converted two points were crucial to their point differential at the end of pool play.

Few teams opt to make passes like the one that Kaka made; most tries are scored where available. Thus, it’s difficult to get a sense for which teams excel at creating easier conversion kicks. In addition, much of try-centering is a byproduct of offense, rather than risky in-goal passes. The centering may occur before the try zone and be hardly distinguishable from open play rugby. Currently I am unable to separate how well teams did in those situations compared to actions within the try zone itself. No matter the cause, our data can provide insight as to how well teams improved their conversion chances (and subsequently their point total).

Using this model, we can evaluate each try against the average conversion success rate. Evaluating against the average success rate separates centering from conversion kicking in our analysis. If we estimate that 99% of centered tries are converted, the offensive and centering actions earned the team 1.98 expected points in addition to the five points for the try itself. Any points gained or lost in comparison to the 1.98 expected points would be accounted for in the section above dealing with conversion kicking.

It should be noted that for teams with few tries (Brazil only scored four) the numbers below are not much of a predictor of future results. However, the table does provide a measure of their performance at the Olympics.

Team Points Added
(all games)
Points Added
(pool play)
Tries w/
Con. Attempt
Points Added
Per Try
(all games)
USA 2.1 2.5 17 0.12
Australia 2.0 1.9 14 0.14
Great Britain 1.8 1.9 13 0.14
Japan 1.7 0.5 15 0.11
France 0.4 1.5 14 0.03
Kenya 0.2 0 9 0.02
New Zealand 0.1 1.5 17 0.01
Brazil -0.2 -0.1 4 -0.05
Spain -0.2 -1.4 7 -0.02
South Africa -1.1 0.7 22 -0.05
Argentina -1.3 -1.2 16 -0.08
Fiji -4.4 -0.6 26 -0.17

The USA and Australia, both poor converters, were best at centering tries. Fiji scored often in difficult conversion positions, and Great Britain both centered well and converted well throughout the tournament.

The centering values above add further clarity to the USA’s point differential after pool play. Though their conversion kicking was below average, they centered their tries quite well. Perhaps their excellent centering provided little opportunity for their conversion kicking to add value above the average, contributing to a poor conversion metric.

Possible Conversion Points Added
To help analyze what chances a kicker had to add value, it would be interesting to know the percentage of possible additional points a kicker added. If all of a kicker’s attempts are centered, he could never miss and still add practically no points for his team since the centering of the try was the main reason the try was converted. Meanwhile, a kicker with only the widest attempts could add significant points from opportunities the kicker with the centered tries never had.

I have totaled the maximum possible additional points each team could have added when compared to the average. I’ve also calculated how much each team actually added with their made conversions. This is only a positive statistic and any misses are counted simply as zero added points.

Team Max. Possible
Points Added by Conversions
Actual Points
Added by Conversions
% of Possible
Added by Conversions
France 9.4 6 64%
Great Britain 7.3 4.4 60%
Argentina 12.6 5.8 46%
Japan 8.8 3.7 42%
South Africa 16.5 6.4 39%
New Zealand 11.8 4.3 36%
Fiji 22.6 7 31%
Kenya 6.1 1.6 25%
Australia 7.8 1.8 23%
USA 9.9 2 20%
Spain 5.1 0.9 18%
Brazil 3 0 1%

This list is much like the first conversion metric; Australia and the USA are still towards the bottom and Great Britain and France are still at the top. Apparently Australia and the USA had below-average conversion kicking in this tournament. Compared to our Conversion Points Added metric, Japan has risen from eighth to fourth. They have an odd profile of missing some mid range conversions while being able to hit some from out wide, gaining large value on the wide kicks.

Considering the competition-wide success on close conversions, the best kicking teams set themselves apart by making a higher percentage of their wide kicks. Australia and the USA had few wide kicks because they centered well, but they missed everything beyond 20m wide.

Combining the Skills
Finally, I have combined each team’s added points from conversions and centering. Whether teams are adept at conversion kicking or try-centering, this stat will give us a complete look at how well teams converted tries.

Team Combined Added Points from
Conversions and Centering
Great Britain 5.1
France 3.8
Argentina 1.2
Japan 0.5
New Zealand -0.1
USA -0.1
Australia -0.2
Spain -1.1
Brazil -1.2
Kenya -1.7
South Africa -2.6
Fiji -3.7

The combination of conversion and centering stats reveals greater disparity between the best and worst teams. Surprisingly, the gold and silver medal winners are at opposite ends of the table. This chart shows just how strong Fiji was in restarts, set pieces, and loose play. They dominated the tournament even though they were the worst team in the field in terms of centering and making their conversions. Great Britain, meanwhile, added a whole try’s worth of points simply with how they managed “the extras”. If a try doesn’t seem like a lot of points to gain, just remember that Great Britain beat New Zealand, South Africa, and Japan by two points each. A lone conversion saw them win their pool and advance to the gold medal match.

While this model will need to be refined with additional data from future games, it does provide a glimpse of each team’s Olympic performance. Both areas of the game, conversion kicking and try centering, provide teams with a quantifiable advantage. More analysis is required to determine just how predictive these metrics are and how often they affect game outcomes.

Great Britain’s unlikely run to the silver would not have been possible without their centering and conversion work. This might have been an edge that they exploited to secure a medal or perhaps they got lucky. Whatever the reason, Great Britain’s success demonstrates that conversions, and the spots on the field from which they are kicked, are valuable enough to decide games and even tournaments. These facets of sevens can not be ignored.