Exclusive: “She doesn’t win three Grand Slams without going through that” – Shane Liyanage on working with Sabalenka and the world of tennis data analysts

Aryna Sabalenka was named WTA Player of the Year on Monday. The 26-year-old Belarusian enjoyed her best season to date in 2024, capturing two of the four Grand Slams and finishing the year as world No 1 for the first time.

As for any tennis player, it takes a whole team to enable them to perform at their very best and the same is true for Sabalenka.

In an exclusive interview, Shane Liyanage, the founder of Data Driven Sports Analytics, who will be working with the Belarusian for a fifth season in 2025, tells of his personal journey into analytics, the use of data analysts by top players and in particular, his work with Sabalenka.

Can we start with your journey into tennis as a sport, and then how did you make a transition into this kind of career?

Shane Liyanage: So, I played a lot of junior tennis in Australia until I was about 16, 17 and then stopped for a little while. After that, I played Australian money-level tennis, which is maybe the lowest level professional tennis in Australia and this was while I was studying at university. So I’ve always had a little bit of a playing history with tennis. I also did some coaching whilst I was at university to get some income. Then, I ended up going to university and started a few degrees, one of them being data science qualification. Later on, I did a Master’s in Sports Analytics. So I always had that passion for applying data into sports.

I also worked for the government in a number of industries, mining, healthcare, using data, working with the professionals in that industry, and giving them some insights, whether predictive or data visualisation. So I had the passion and I guess I had the skill set.

Eventually, I actually took a role at Cricket Australia. And that was the entry into sport for me to get some experience working with athletes, and to work with a sporting organisation, which is very different to working in a government sector job or with corporates.

Understanding the athletes, understanding the communication, the soft skills, I think I learned a lot during my time at cricket. And then I decided to try it at tennis by myself. Luckily, I had some good connections through my tennis days to coaches and, and a few of them tried me.

I was fortunate enough as an Italian coach, Federico Placidilli, wanted to give me a shot with an ATP player, Thomas Fabiano. He was ranked just over 100 at the time in and then he had a really good season. I guess after that, other coaches reached out. And that’s how my journey started.

Since then you have gone on to work with several elite players as well, including Aryna Sabalenka, Ons Jabeur and Karolina Pliskova. More often that not, do you work with the coaches or directly with the players?

Shane Liyanage: Generally, all the relationships are a little bit different. But with the bigger players, the relationship is almost exclusively with the coach. Of course, we do have some relationship with the player as well, but we’ve got to be very careful that the communication goes via the coach just to make sure that the information is delivered at the right time. With the big players, that’s how it is.

But then I’ve worked directly with some players as well as they don’t have a full-time coach or they’ve preferred to work directly with me. There’s a couple of players that we work with now that have a coach, but the coach may not have very strong English skills, so the player interacts with my team more directly. But I would always say in an ideal world, the relationship should go via the coach.

Everyone talks about how difficult it is to earn a decent living out of tennis when you're ranked outside the top 100 in singles, or top 50 in doubles. At what point, do you think pro tennis players start looking out for a data analyst to add to their team?

Shane Liyanage: I can only speak really from my experience with players, but we’ve worked with players ranked outside the top 500 as well. The thinking is, in my opinion and I’m sure some people would agree and some would disagree with me.. but imagine winning one extra match at a big event. In Grand Slam qualifying, winning one extra round is significant amount. Imagine actually qualifying and making it to the main draw. I guess some of the players have brought into that way of thinking and invested in using us because basically qualifying for a Grand Slam more than pays for our services for a year. So there’s that way of thinking. But yes, it’s true. It is probably, in some people’s ey,e a bit of a luxury, certainly with the money that players earn. And the priority is usually to have a coach and having a physio with you for the whole year. And I absolutely agree that they’re incredibly important things.

But I’d like to think that the investment in having someone looking at the data is still high up there because it’s the way you’re going to get feedback on whether the things you’re doing are working, did a match go the way you expected or not expected, and that sort of an insight into the opponent, which I think translates to earning more income. So, I certainly understand that the reasoning that people have sort of put that to the side, but I think in reality, it’s almost an investment that a player needs to make.

I’ll just end on saying that I do work with some federations as well. So the use of data isn’t just for that sort of top 100 bracket, but they’re using it for junior development, recruiting the right juniors, making sure the right juniors have the right coaches around them. So data is being used at that junior level as well. And then the players ranked outside the top 100. I know that some of the federations that we’ve worked with, we’ve been working really hard to help those players go from ITF to ATP Challenger or WTA 125K level. And then for them to be in the Grand Slam qualification draws where they have the potential to earn much more income. I think 10 years ago, using data was maybe a luxury, but now I feel like almost every player in the top 200 needs to invest in something like this.

From your experience in the industry, what percentage of the players in the top 100 in the world would be using a data analyst at this point?

Shane Liyanage: Look, I’d probably put it around 70%. I’ll try and explain it. So we have the big four federations (the Grand Slam federations), they have an in-house data analytics team.

So all the players in those federations get it to some level. Some of these players may not use the data but they do have it available to them. Then you have some of the other bigger federations like Tennis Canada, Tennis Italy, and they also have a team, either a third party or some in-house team. Then you have, the top ranked players, not from one of these federations, that they’ll go with a company like mine. There’s a number of such companies out there.

And then some teams have an analyst in the team, a little bit like the relationship I have with the Sabalenka team. Some teams have a dedicated analyst and it could be a coach upskilling to do it. And I think all those examples, 70% of the top 100 on both tours have it. If you looked at this three years ago, it would have been about 50%. It’s 70% now. In four or five years, it’ll be almost everyone in the top 100, in my opinion.

When you provide analytics for players, is it a standard set of metrics that you would track for a player? Or is it customized based on the coach’s inputs?

Shane Liyanage: One of the things that we do very early in a relationship with the player is called a deep dive into the player. We look at their strengths, their opportunities and then we work very closely with the team and the coach in particular to identify what are the priorities for the next three years based on data but also based on their experience with the player as well.

Then we try and set up KPIs for them and that’s the broad theme. Even when we’re scouting an opponent, we still put the KPIs that we hold for our player pretty high up there because we have a longer-term goal that we want to achieve.

I think in tennis, the reality is you need to still look at your strengths. Your strengths have to win you the match. Yes, you can look at the opponent’s weaknesses and what to use, but a player can only play within their skillset. So, it’s very much customized. That’s been something, it’s been a priority for me to make sure my team is doing that. With Sabalenka and Jabeur, for a long time, there were two top 10 players, both top five at one stage as well, but they played very differently. So they’ve got very different KPIs that they need to hit to win matches. So, we try and individualize the work we do.

You spoke about scouting opponents for players you're working with. In tennis, it’s a very short turnaround from the time you become aware about your next opponent. How easy is it specific data to your next round opponent?

Shane Liyanage: We have a pretty comprehensive library. We try and update everyone in the top 100 consistently. Even going into Grand Slams, we try and make sure we have the top 200. You could have a lucky loser. It could be a very, very sharp turnaround.

We have a pretty big library already, but then of course, when you see the draw, we’ll try and get some really recent matches as well, similar service, similar type of player. We get the vision quickly to the coach, but we also then produce a data report based on what we’ve agreed on our KPIs. And we get them to team.

And then we’re looking around ahead as well in the draw. Gand Slams are easy because you got that day off, but in a normal event, we’re almost preparing around on both potential opponents. If you have a night match, then you’re maybe playing a day match the next day, you don’t have that turnaround. It’s tough, but it’s how we have to do our work.

Have you had a very different kind of request from the coaches or the team to track as a player, something that you wouldn't really have imagined that people were looking at?

Shane Liyanage: I don’t know if this is strange or anything, but a few coaches have asked for the players’ first step speeds moving to a forehand or something like that. That was a little bit of a unique request.

We have the tracking data so we could do it. Things like when the opponent’s ball crosses the net, the speed of the movement as well. What one of the coaches asked for how many times they missed a split step, for instance. So again, a unique request. A lot of these requests, if we think they’re going to happen a lot, we’ll write some code. So we know we have it ready, but of course there’s some strange, unique ones.

In the recent US Open final between Sabalenka and Jessica Pegula, we thought we saw something. With the new balls, Pegula seemed to be doing really well, with the older balls, Aryna was doing well. That’s what it looked like. We went back and again, we had to write a little bit of custom code, but we were able to pull out and our eyes were correct.

Pegula played a lot better with the new balls, Sabalenka played a lot better with the older balls. And of course, we’re going to take that and think about it and see if there’s actually something we need to do or not.

Talking about KPIs for players, how do you account for the mental side of the game, which is so so important in tennis. Because the third set in a Grand Slam final might be very different from the first set in the early rounds. So how does account for that when you're doing the data analysis?

Shange Liyanage: I think the reality is there’s no way to do it perfectly, but we definitely use scoreboard pressure. Things like facing a break point or in a tie break, we deem it as a pressure point.

We definitely look at things like that. We also look at things like you may have just played a really long rally that you lost. So we think, okay, immediately there’s some negative emotion because it was an exhaustive point.

So we have a way of tracking that and looking at your behavior after that. There’s a bit of work that we’ve started doing with AI, looking at facial recognition. And again, it’s early because the video quality can be very different between the tournaments and the streaming sites.

So using your facial expressions to go, okay, you just had a positive emotion. How do you behave after that? You had a negative emotion. How do you behave after that? Looking at the outcome of points.

Of course, there’s now more third-party data coming in. So we can look that your heart rate was really high on this point. So we can assume there was some pressure and then we can combine that with the normal shot data as well to see, were you hitting shorter? Did you miss more? Were you more conservative? Things like that. We try and bring in a few different ways, whether through another third-party data device or we will derive something based on scoreboard pressure. But admittedly, there’s a lot more that can be done in that space. And certainly we’re only just scratching the surface with how we deal with that.

Talking about Aryna Sabalenka. A couple of years back, she was struggling a lot with the serve and double faults. She worked with a biomechanic coach at that time to improve that. The last couple of seasons, she's been really turned around that aspect of her game. Were there any specific data points related to her serve that helped to give feedback to the coach as to what was going wrong and how it has improved?

Shane Liyanage: Definitely. Again, I can mention some things and not everything, but definitely we saw, the outputs of certain serves, the ball toss where she was hitting it from in terms of the racket head speed, wasn’t optimal. We also had data points to show that in pressure situations, the whole swing service motion was even more sub-optimal. We had all of those things.

And then of course, the work that the biomechanics coach did, I think improved her understanding of her own service motion, but also fixed some of the technical flaws, which again, under pressure, was breaking down more, but I think the flaw was there to begin with. And the work that Gavin MacMillan did was exceptional. Of course, Anton Dubrov, who was traveling with her, had to make sure it was applied every week.

So again, his work was instrumental as well in it. And then we had the use of data afterwards. You almost had that before and after. So we could see, right, this is the service motion now. You can see the speed of the racket. We can see the percentage of balls in play.

You break it down by in play, T serve, wide serve, kick serve. So we could go another layer in terms of seeing how the serve performed. And I want to say that that year was a blessing in my opinion, because it allowed her to work on other parts of her game as well.

There was that one Australian Open where Aryna was averaging 20 double faults and she still made the fourth round. So she was giving 20 points to her opponent and still winning lots of matches. And it meant other parts of her game improved. And I think she opened up to working on the technical side because that happened. And again, Gavin not only did work on the serve, he worked on the forehand and backhand as well. So for me, that year was a real blessing and she doesn’t win three Grand Slams without going through that.

In the last couple of seasons and before she won her first Grand Slam, are there any specific things that you can share as to where the improvement was, in which aspects of a game specifically supported by data?

Shane Liyanage: Again, I can probably be a bit general here. I think everyone knew that she was a big ball striker, but certainly in the last two years, it’s finding that balance between a big ball and a big ball with spin.

We’ve found an optimal number in terms of forehand speed and spin rate that we’d like, backhand speed and spin rate, even the serve as well. And internally, I know Gavin, Anton and I, we talk about Aryna winning matches with 85%. And what we mean by that is 85% of the power spin.

She finds that range where that number there is bigger than the WTA Tour averages for both speed and spin, but it’s a reliable ball that we are confident she can make a lot of balls.

It’s actually a challenging ball because it’s still fast, but it’s spinning a lot. And on the WTA tour, if you can get the ball out of the hitting zone of players, it makes it really hard for them. So I think definitely that’s been one of the big improvements.

And she’s also got more variety in her game. I think anyone that’s watched her in the last 18 months can see that she’s using the drop shot a lot more. She’s trying to get to the net a little bit more and being more comfortable with finesse shots and slices and things that I think at the start of her career, it was very foreign to her to play those shots.

And it was only in very extreme circumstances that she would do it. But now she trusts herself even in a pressure moment to use those. And then movement, her movements have improved as well. And something that we’re going to continue to work on and to get even better going forward.

One other thing I want to check is about the differences in terms of working with men and women on the ATP and WTA tour. Besides the obvious differences of serve speeds and the speed of the shots, are there any kind of differences that come out of the data, which surprise you??

Shane Liyanage: It’s surprising but the speed is something that there’s not actually a significant difference. You take the serve away and the first serve away, the speed is not so different. The spin is significantly higher on the men’s side than women.

But the ball speed, like Aryna, you would have seen some statistics during the US Open, she was hitting the forehand bigger in terms of speed than the men. But of course, if you look at the spin, it should be significantly lower than the average sort of male player. But the big thing is, I think the women’s game is more a linear game, whereas the men’s game can be a bit more lateral.

The men will open up angles a lot more. Again, they’ll vary the heights a lot more than a female player. I guess working over a number of years, it’s not surprising to me now. But when I first started working with an ATP and WTA player, these were things that I quickly realised in terms of strategy that a male player would utilise the width of the court a lot more than a female player.

If somebody's out there reading this who wants to get into the profession of a data analyst in tennis, what would your advice be to them in terms of the path they should take?

Shane Liyanage: My tip is, firstly, to obviously try and get some sort of formal sort of qualification in data, because I think that gives you a good grounding in terms of the methods, whether it’s Python, whether it’s SQL, whether it’s data visualisation tools, you get a little bit of an introduction into those things. Then what worked for me really well was actually working outside of sport. As I said, I worked in areas where I was not the subject matter expert at all, I would have worked in the healthcare sector, where I had to talk to doctors and nurses about things that I had no idea.

So it really helped me to learn from the subject matter experts and that sort of listening skills. I had the data skills, but they’ve got the knowledge. So that helped me. I think for anyone that’s learning, they should go get a data set on and then work with an expert in that area, and then try and build something that’s useful for them. I think that’s a really useful skill. And then I think in terms of sports, there’s lots of really good sports public data sets out there.

I remember when I started, I was playing around with tennis data sets that were publicly available and trying to come up with something that was insightful. So I think play around with it, get your skills up to a certain level. And then go volunteer. If an opportunity doesn’t immediately come and there’s no job there, just volunteer and try and solve a problem. I’ve spoken to a number of people and I said, go to your local football club. And they wouldn’t have the luxury to have an analyst, but I’m sure they have problems that can be solved with data.

So you go there and you prove yourself there, and you’ve got a reference there, and maybe there’ll be a job there. But even if there’s not, you’ve done something at that level. And eventually an opportunity will present yourself and you’ll be ready because you’ve done it in a low risk environment first, and then you have an opportunity.