Since the turn of the century, Clermont Auvergne have been one of the most prestigious and successful clubs in France’s Top 14.
Historically, they are a high tempo, attack focused and fiercely passionate team. Their home stadium, “Parc des Sports Marcel Michelin,” is one of the most formidable fortresses in all of rugby – famously between 2009 and 2014 the “Les Jaunards” were unbeaten at home in 77 games.
The team have been French Top 14 champions twice (2010 & 2017) and made it to the final thirteen times. They have been Champions Cup runners up three times (2013, 2015, 2017) and won the Challenge Cup on three occasions (1999, 2007, 2019).
Data and analysis are integral to Clermont’s success, helping to inform their desired playing style, the recruitment of new players, competitor analysis before and after game weeks, and planning for the future of the club. Stats Perform spoke with Joe Larkin, the Head Performance Analyst at Clermont, to explore how data and analysis inform decision-making throughout the club and how Stats Perform products are powering these decisions.
The Bigger Picture
The game we love is changing. Constantly. Spotting trends in the game early and capitalising/preparing for them is integral to long-term success. Clermont utilise Stats Perform’s RugbyHub and Data Engine software, which is powered by Opta data and integrated with video, to spot trends from other clubs and competitions without having to sift through 1000’s of hours of game footage.
Q: As a club, what are your key longer-term objectives and how does your use of data support that?
“We want to see where trends in the game are, where the style of play was, where it is now, and where it’s going. We use the trends to inform and build our squad for the future,” explains Larkin.
Two such trends that Larkin has recently observed in Top 14 is the increase in ball in-play time and an increase in kicking, a trend Larkin believes has filtered down from the style of the French national team, on the back of their recent success.
“I use RugbyHub and Data Engine to learn from non-direct competitors,” he says.
“The products give me the ability to mine through the specifics of team performances, for example I’m able to watch how a team in the Premiership exits from a line out in their own 22, I can watch every single lineout from every team”.
Larkin also recognises the benefits of being able to look at examples of set piece prowess from Super Rugby teams – “Super Rugby offers examples of good set play structure. We can look at how they attack off clean ball, even though northern hemisphere rugby doesn’t always have as much clean ball generally”.
Larkin’s Performance Analysis team relish the ability to be able to analyse both northern and southern hemisphere rugby, neatly processed and clipped up, so they can make strategic decisions on Clermont’s long, medium and short-term strategy.
Competitor Analysis
With games in the Top 14 coming thick and fast, it is important to understand and analyse your opponent at speed. Opta data integrated with video, allows the analysis department to package up clips for the players on an individual basis, which they can edit and present using Clermont’s “rugby language” and align it with their calls, structures, and areas of focus.
Q: How do you use Stats Perform products to help aid your competitor analysis?
“As a team we will create an analysis pack for the lineout callers, guys like Sebastian Vahaamahina and Julien Cancoriet. Typically we will send them a package on a Sunday night for the team we are playing the next week, so they can come in on Monday ready to talk it through with the coach and figure out a plan of attack”.
When it comes to lineouts you can’t argue with Larkin’s approach. His team have the most effective lineout in the league so far, winning over 85.4% (216/253) of their own lineout ball.
With margins so close in Rugby, these insights have the potential to influence key tactical decisions, which could be the difference between winning and losing.
Data-Led Recruitment
The game is constantly evolving and as a result, teams must change and mould to fit new playing styles and trends in the game.
Larkin explains how data can help them predict the direction of the sport early and recruit players to fit a future playing style or ideology.
Q: How do Stats Perform products and data help inform your recruitment decisions?
“We use Data Engine to find certain types of players that we might currently be missing in the squad due to injury or contract changes, for example recruiting a back rower who frequently steals the ball (rather than one who carries a lot). We can target turnovers won and jackal success in the breakdown specifically, to find which players are rising to the top in that position with those particular attributes”.
Larkin also uses the data when faced with an injury in the squad. He can do a deep dive into the available players in that position and watch clips of all their positive contributions, but also look at all their weaknesses and errors. The analyst finds this especially useful as an agent can “make almost anyone look good in a highlight reel” – meaning it is important for Joe and his team to see unbiased and comprehensive clips of all the players available, giving them all the necessary tools to make well-informed recruitment decisions.
With the increased tempo of the game in recent seasons and ball in play time on the rise, it is important to recruit mobile front rowers who can play large periods of the game without fading out.
Q: Are there any players you have recruited using Data Engine that have been particularly successful?
“After analysing Data Engine statistics from France’s Pro D2 league, we identified a young Moldovan prop called Christian Ojovan. He topped the league in all important metrics for us at the time and was conveniently playing down the road at Stade Aurillacois, so he was a great option for us. More importantly we identified his strengths were the same ones that made a successful prop in the Top 14. Since being in the team he has improved by playing in a better league, but still has the important base attributes we identified and led to his signing”.
The data didn’t lie and Christian has gone on to be a real success story for the team this season, ranking fourth in the league for minutes played by props, fifth for prop carry metres and seventh for attacking ruck hits in his position.
A Data-Led Future
Staying agile and ahead of the curve is as important off the field as on it. In addition to RugbyHub and DataEngine, Clermont utilise raw Opta data , which enables their data analysts to analyse the data and manipulate it in Power Bi for more specific team requirements like “injury statuses, recruitment or future trends in the game”.
Larkin sees the future of rugby analytics imitating the path of football, which in recent years has started to evolve from descriptive analysis to predictive analysis, informed by advances in AI and Machine Learning such as Stats Perform’s Qwinn. Examples of these predictive models are already being used in rugby, including Stats Perform’s Kick Predictor model which was released ahead of this year’s Six Nations. Larkin believes that similar predictive insights will become key to analysis, both in set pieces and open play, as the decade evolves.
Q: Where do you see the future of Rugby analytics going?
“In a certain period of a game we need to be able to spot game swinging moments more easily and during the game so they can be capitalised on, these moments can often be the difference between winning and losing.
“Looking ahead, I see us being able to quantify things such as expected line break opportunities for a certain player from a certain part of the pitch, or potentially an expected try after a line break. These types of insight would be really valuable in informing key aspects of our analysis.”
Clermont’s forward-thinking approach to rugby analytics is providing the club with solid foundations to sustain and evolve their on-field performance, in line with wider trends across the global game, as well as inform key recruitment decisions which can help them challenge for future honours.
Clermont use Opta data, Data Engine and Rugby Hub. To find out more about these products, take a look at our Pro Rugby Services.
