Marc Sophoulis
Thanks for joining us once again here on Crunching the Numbers. You have Marc Sophoulis and Shane Liyanage from Data Driven Sports Analytics. Shane, thank you once again for joining us.
Shane Liyanage
Hey everyone, hopefully you've had a great week this week and look forward to talking to you on this podcast today.
Marc Sophoulis
Hopefully everyone's doing well in these tough times around the world and obviously, we just want to bring you the latest tennis and keep you up to date with everything tennis has been doing from a data perspective and matching the data with the art of coaching as well and being able to implement that. Today's topic, we're going to talk a little bit around the tournament analysis and planning side of things. Obviously Shane, yourself and I have been doing for a period of time together and looking at not only our own player, which we touched on in our last episode, but looking at how we utilize what our player does and be able to break down opponents. Shane you're a big part of what I've been doing over the last few years. And we had great success at the Australian Open and have had great success with a few other players in the last few months as well.
What you do is really critical to what I do. And it complement each other. And it's the art and the science merging together to be to create some success. Shane, can you talk us through...let's say we have a player that is about to play a tournament. What are the steps that go into place from your side of things to be able to then help the coach or the player?
Shane Liyanage
Yeah, so today we're going to talk about when our athlete's in a tournament and the work that way we do before the tournament starts. Once the draw's known, what happens during the match and post match, what do we do? You will notice that at all of these stages, there's something that's quite fundamental and that's communication. So the communication obviously between Marc and myself has to be clear and we have to make sure that we understand each other. And then Marc has all the information he needs from me, so he can deliver the message he needs to deliver to the athlete as clear and articulate as possible so there's no misunderstanding.
The other important thing that I want to do right at the start is I want to distinguish this work to the sort of pre-tournament work that we were discussing in last week's podcast, where I think the focus is almost 100% on the athlete, with the exception of maybe a few benchmarks of other players that we might try and compare our player to. Really the focus is on our athlete, their patterns, their training. But when we go into a tournament, I think particularly in my mind as the analyst or the data scientist, I switch immediately to thinking about the opponent a bit more. While Marc and the coaches and the players, they're chilling in their cafe, sipping their pina coladas, I'm trying to get my hands on the draw as fast as I can.
As soon as I know who they're playing, the first thing I'll do is have a look at my video database to see if I've actually got some data on this player or this opponent already. And in particular, what I'm trying to look for is not just the opponent, but when that opponent has played someone similar, a similar style, height tendencies on court as our player. And then also the next layer is really to look at; what are the conditions the match will be in? So what surface are they playing on? Is it indoor? Is it outdoor? Is the court set up in a way that wind might be a big factor. Are they playing during the day? Are they playing at night? These are all things I try and pull out matches that are similar. So the information takes out that scenario variability. But that's not always possible.
I think I mentioned last week the matches in my database...there's a pre-built dashboard that sits on top of that. So when the matches are in there it will already capture some key area information like serve directions, return positions, behaviour under pressure. This list is not exhaustive, but there's a lot of things that automatically pre-populates when I select a player. The video itself can actually be filtered for the coach. Probably important as well for me, even if the match does already exist in the database, that I actually still watch a few of these matches. So I'll use the filters and try and look at key moments and key things on the forehand, backhand positions and take notes. One of the problems when you're working with some of the future level players, you don't have the luxury of all the footage. So you have to scour the internet. You have to tap into your coaching network. I've got a great analyst sort of network that I communicate often with, and we share information. So you going to need to tap into that. Almost as a last resort, you can look at things on YouTube, Dailymotion, some of the video sites online to try and get some footage.
My aim is to try and get at least five matches worth of data, but if there's no footage in the database and the player is playing the following day and often these draws are 6:00pm... so there's not a lot of time, then it's probably two, at best three matches, depending on what footage is available. I love Grand Slams because as the analysts in the team, you have that extra day. But I still want to stress the importance of getting that information in a timely manner is incredibly important because you want the coach to be able to use some of that in practice with the player.
Marc Sophoulis
You get the opponent's name, you find them, you get a match with this. You cut up the match to make sure that there's no length of match that we have to watch. So basically, you take away all the space in-between points and games. You send me a video of the player, which is condensed from like two hours to 20 minutes, which is super. And I get to watch it in shorter time frames. And you also then tag it so then I can actually research just when the opponent's serving or just when they're returning. So I can get a really good feel for what the player does.
Shane Liyanage
Yeah, absolutely. So I think that's fundamental like - Marc mentioned reducing the length of time for the coach and the player. They've got so much on their mind and so many things from a training perspective, from gym and eating, that they don't have the time or the luxury to watch 10 matches. So condensing it to key parts that the opponent does well and key parts that your player does well, or the opponent doesn't do so well. That's really important. The other thing I try and get to the coaches, I've got a pre-built dashboard that spits out a report. So I'm trying to get a report with some of the key statistics to the coach. But in an email I'll just put my observations to the coach. Because I've watched probably more of the footage, it just gives a quick summary for the coach on how I'm seeing the game.
Marc Sophoulis
It's actually important to have the vision and the numbers and the interpretation of your side. And then I read it, watch it and give my interpretation. We bring it together to be able to create a game plan. And I think that is really, really critical. To make sure we grab everything and then we put it in the funnel, and we filter out all the stuff we don't need. We just bring out that little bit of gold information that we do need to be able to plan against the opponent.
Shane Liyanage
Yeah absolutely. And I'm reaching at the analysts that might be listening - I don't think all that data ever gets wasted. So in this situation, we might only use two or three things for this scenario. But down the track we could be looking for another scenario. So I think doing the analysis and doing it so it's mathematically correct. All of that is really important and that work still needs to happen. But absolutely, it does need to go through a panel, because as Marc touched on in a previous podcast, the player can't take in a lot of information before on that.
Marc Sophoulis
Yeah, so it's about keeping it simple, especially from a coaching perspective. I want to keep it simple for my athlete to understand. When you do your analysis of the opponent...we know what our player is going to do in any particular moment, we've got their game plan sorted, their patterns of play. We know where our player likes to serve on the big moments and how they like to play the patterns, etc. What are you looking for specifically? And we talk about this as being condensed into certain elements of the opponent's game. But what do you give me? And can you give the listeners a bit of an idea from a data perspective, how much do you give me?
Shane Liyanage
Yeah. So I think it's really important with the key parts of tennis, the serve, the return, the volley game, movement. I try and give a summary for each of those points to the coach. But you touched on patterns, so I would look at patterns that work well for the opponent, patterns that don't work so well and then patterns our player does, and try and marry up - "okay, if we play our good patterns, then that's actually exposing the opponent". So trying to do that sort of analysis and give that to the coach.
Marc Sophoulis
And we have a real big focus - I like to have three areas that I work with in my game planning, and that is: when my player is serving, when my player is returning and when the ball is live and it's in play after those first four shots. So basically there's three scenarios you can play. It's the serve +1, return +1, and then live play. You break it down for me, which I think is really good, to be able to then have a game plan in each of those three categories. And I've got in front of me one of the ones you've broken up from a particular player, you show me how many errors he had etc. But my real focus point is - we can only control basically the serve and the +1, probably the return on the +1. And then when the ball's in play, how do I want to play and where do I need to put the ball to beat the opponent? So the way you break it down is really important for the coach and making sure that you break it down to the coach's philosophy as well.
Shane Liyanage
And often I'm trying to also, particular in the report, give some information under pressure. Maybe there's one or two little gems in there that you can pick out just to keep an eye on. But again, it's trying to keep it simple in the report as well.
Marc Sophoulis
That general analysis for me, the 'under pressure' analysis is really important. I remember there was a match that you gave me an analysis for a player we were working with in Mildura. You came up with - on the big points the deuces are 30-30, the opponent is always going to serve to this spot. And then we had our player basically sit on that return on the big moments, and she was able to combat the opponent's best play the whole time. And that just took away the strength of the opponent, the opponent started to then change their patterns. It didn't work for them. They made more errors. And I think that was a huge moment because I only got the player to sit in that spot on those big moments and not every single point.
Shane Liyanage
I suppose just one example. I want to stress here that I'll give a lot of information. It has to go through that filter with the coach and communication is big and there are situations where the athlete will try and fish for information and go to the analyst. So I think with Marc's athletes, it hasn't been that bad. But with a number of coaches that I've worked with, the athlete will try and get more and more information. And it's really important as the analysts to check in with the head coach before you give anything. There's a plan and you want to stick to that plan.
Marc Sophoulis
That's a great point, because the reasoning behind that is there's obviously the trust element. There's the coach knows the player better than the analyst. The analysts will know the numbers. And just knowing how each individual will accept that information is critical. So what you may say to one of my athletes might be 100% correct, which I'm sure will be, but it might be delivered in a way that the athlete receives it differently to what the coach would give it. And I think that's the critical part of amalgamating the science and the art together, because athletes are in a mindset of they just want to win. So they will get as much as they can, but think that's the right thing. But generally, it's not. It might only be one or two things that will change the match. And that's all you really need. And the coach knows the athlete much better than anybody because of the trust and relationship built previously.
Shane Liyanage
The next thing that I want to move our discussion to is what happens during the match. For me, it will go back to where the relationship and the role has been defined with me and the coach. And sometimes the coach really wants me to do live tagging. If that's the case, then my preference is to sit either in my house, in the hotel room and get a stream of the footage and then I can do it with the computer. Because often you can't go court side with the PC. And the mobile devices in terms of tagging panels, they're not up to that high performance level that you need. So that's my preference if I'm doing live tagging. If I'm not doing live tagging then I'm usually trying to be court side in the box and just take a notepad and just jot down a few notes from the match. Some critical moments, key observations, behavioral. Just take those notes down, because I think that will help me when I do the comprehensive match annotation tag later on to focus my attention on certain things and will also help in terms of the email summary of my view of the match to the coach.
And if I'm not doing the live tagging, usually I'll do the match tag at some point on the same day after the match is finished. But noting that the priority is given to tagging the next opponent if they're still in the draw. The Data Driven Sports Analytics we've certainly reduced the tagging time and we've starting working with some great partners on trialing some computer vision solutions which are reducing the tagging time. So it's allowing us to get more of the tagged data done quickly. I think the other point I want to stress here, is some coaches want to do a review within two hours of the match. If that's the case, then you do need to do your player's match quickly. Others just want notes. So if you've taken notes during the match and your observations, that's all they want. And some coaches (I think I'd put Marc in this bucket) have a very forward looking view and they don't really want to look in too much detail on the match that's taken place, particularly if they're in the draw, until the end of the tournament.
Marc Sophoulis
My philosophy is if my player is continuing to win through a tournament, as we had David do at the Australian Open or Katherine in Mildura. If they're winning, you don't want to look back as to what they've just done. You want to keep continuing to look forward. OK, what can we do now? OK, we've got to recover really well. We've got to stretch. We've got to look forward, we're going to get a good night's sleep. We've got an opponent coming up that does a, b, and c. Let's practice for that. Let's get into the mind frame of looking forward to stepping into the ball rather than waiting back in this match and you try and push their mindset to what's about to happen. And to stay present. If you continue to look back, we're just focusing on things that have already been done. For me, I want to make sure that my player celebrates the wins and celebrates the little things and enjoys that, but don't analyse that. That's been done. The only thing I take from the previous matches - "I thought you did this, this and this really well, we're going to take that to the next match". So I'll take the positives from that previous match. And then at the end of the tournament Shane, I ask for all the vision and then we do the reviewing at the end of the tournament.
Shane Liyanage
Yeah, I want to stress to the all the performance analysts and data scientists working in teams and not just tennis, in other sports as well. But you make sure that you keep that data that you've collected, you label things in your database nice and efficiently, because at some stage, you're going to call upon that analysis to do a report or to do some prediction modelling. And you want to make sure that the work that you did during that tournament is appropriately labeled and you can sort of make sense of it down the track. The more data you can collect the better that subsequent analysis can be.
Marc Sophoulis
Do you want to talk a little bit about post tournament. From your perspective, what happens post event?
Shane Liyanage
The other thing that's sort of going on in the background, it doesn't necessarily go into that sort of tournament review, but the data from the tournament that the player just played, it goes into the database and every 25 or so matches, I try and do a report for the coach and the athlete can say it as well, which just highlight some of the things that have happened in that sort of block of time. So you look at key areas; serve, return, directional numbers and patterns that have worked well, haven't worked well. And then you compare it to perhaps the twenty five game block before. And that's something I do periodically.
Marc Sophoulis
Yeah, I think it's really important. I think after a tournament, basically my role is getting all the information, the data from you and then seeing the commonalities that happen. A lot of coaches take the data. Say for example, the player doesn't serve well in one match, but they've been serving really well previously, the coach sometimes jumps on the, "oh you didn't make enough first serves, so we need to go practice your first serve". But it's not a commonality. I think we can't jump to conclusions that once off events don't necessarily need your attention all the time. That then gives the player too much of a worry about, "oh my God, I'm not serving well". And I feel like we've got to make sure from data we look at common denominators, common problems, common issues, common positive things as well, and utilize the commonalities that apply to those rather than once off occurrences. And if we get good at that as coaches or analysts or players, we become more successful and consistently successful, because we're not jumping on one moment of "my didn't work in that match. You know, it doesn't matter if it doesn't happen once. It's a matter of the common things. That's development. Is correcting and analyzing the commonalities and not just the once off occurrence.
Shane Liyanage
Yeah, that's key Marc. So with the stuff that we go out with, we've got to make sure that we've got a sufficient amount of data in our sample size before coming up with a conclusion based on one match or part of a match. Certainly with the tournament, we have the entire tournament's worth of data when we sit down with the player and to discuss trends. And that's perhaps one of the big reasons why I like to do those 25 match blocks before I come out with a report. You can have one match that's a complete outlier compared to any other match.
Marc Sophoulis
Absolutely. In concluding this podcast, its important for our listeners out there to really take into account that the communication between your data analyst or your player and yourself as a coach, or you're a player out there and you've got a coach - communication is critical to devising a game plan. Whoever else you work within your team, everybody needs to be on that same page to ensure that you amalgamate the art and the science between what you're doing. So get the raw numbers or the vision and the opinion and then also make sure that you utilize the art to give that to your athlete. Or if you're a parent and you've got a child out there, it's the art of how much I give someone that is critical, not just giving everything that I see and everything that I know.
Utilize the little bits of gold when you're planning for matches. So, if you see an opponent has ten different things you can isolate, make sure you only use one or two against the opponent. I've got a little rule when I do a game plan, it's the 80-20 rule. 80% of the time focused on what you can control and 20% the opponent. There's no point, vice versa, because if you're focusing too much on the opponent, you've got too much out of your element of control, out of your space. So allow your athlete to play to their strengths and then identify two or three little bits of gold in the opponent that we can expose.
The other thing is also review once the tournament is done. So continue to look forward, don't review every single match. The player doesn't want to hear all about what they did wrong in the previous match. Take the little bits of positivity and move forward, take that into the next match. At the end of the tournament do a review together and then train for what you need to work on for the next tournament. And also, if you're a little bit more like me, you need a little bit of structure to your planning. So think about planning three scenarios: When you're serving, when you're returning, and once the ball is live after those first four shots, are the three areas that you can really identify for your game planning and game structure.
Hopefully we've been able to give you a little bit of an insight into what happens on game day for your athlete, for your analysis, through your coaching. We haven't touched on being a parent on game day, but if you're a parent, just make sure you don't pull your hair out too much, because it's quite a nerve wracking time sitting in the stands. And we definitely feel it and the player also feels it as well. And please allow the player to play and do their job and we'll do our jobs. And hopefully we get some success in the future and hopefully you're enjoying our podcasts. Obviously you've been with Shane Liyanage from Data Driven Sports Analytics. Shane, thank you so much for joining us again.
Shane Liyanage
Thanks, everyone. Thanks Marc.
Marc Sophoulis
My pleasure. And I'm Marc Sophoulis from The Tennis Menu. Please join us on all our social media avenues. Stay in touch. We want to keep you sane in your own households at the moment during your isolation periods and hopefully you're trying to learn as much as you possibly can and become the best athlete, parent, coach or data analyst you could possibly be. So enjoy the next couple of days. Couple of weeks, couple months. And hopefully we'll continue to bring some great information to you all in your homes and look forward to speaking to you all next week. Thank you!