UP519 Live at Sportradar Connect: Soccer Edition, featuring Jamie Carragher

0:00I think in the early 2000s Prozone came in, that almost felt like the first thing that could track the data of players, how far they'd run really. But then you get a situation where certain players were just making runs across the pitch, just so the numbers would go, hi there, that is true. So they were gaming there? Yeah, I remember playing right back and Danny Murphy just made a run across the pitch. I said, where are you going? He said, no, it's good to get me stats up, get me sprints up. That was the unmistakable voice of Jamie Carragher, the former Liverpool and England footballer who today is one of the most watched and admired commentators and analysts working in sports television.

0:39Jamie's one of the faces of Sky Sports award-winning football coverage in Britain and he also appears on the CBS hit show Galazzo in North America. I spoke with Jamie as part of the Sport Radar Soccer Connect event at the London Stadium in Queen Isbeth Olympic Park, home of West Ham United Football Club in Stratford. And this podcast divides into two parts. First, you'll hear Jamie and fellow guests Eric Derm and Niccolo Der Kohl recorded on stage in front of an invited audience. Eric Derm was a star of the Bundesliga for Borussia Dortmund and Eintracht Frankfurt and was a member of Germany's 2014 FIFA World Cup winning team.

1:19And Niccolò de Kohl is Executive Vice President of AI and Technology at Sport Radar. And Niccolò joins me for the second part of this episode, which is a roundtable conversation. We convened at the end of the day with other senior Sport Radar executives discussing the key marketing and technology trends that dominate the sports betting and entertainment industries. So, welcome, I'm Richard Gillis and this is Unofficial Partner.

1:50We're going to talk about data and analytics. That's going to run through this bit of the conversation. Jamie Carragher, I was trying to work out when your first game for Liverpool was and what data analytics meant. What was the conversation then? There was no conversation really at all about data analysis. The only data was, I think, the mark you got the next day in the paper. That was the only data we got. I think I got a seven on my first game. What was the first game? 1997. My full debut was against Aston Villa. So that would have been January 18th, 1997. So a long time ago. But after the game, there was no analysis data-wise provided by the coaches or anyone in the media.

2:33At that time, it was just a case of, through your own experience, you knew if you'd had a good game. or the coaching staff, of course, football like it is now, even though we're inundated with data, it's still about opinions and people going to the pub and thinking who played well and who didn't. So yet there was no data to back up my performance. It was just what the coaching staff and the supporters felt. Did you watch the Kenny Dalglish documentary? Yes. There's a bit in there where it's the Bob Paisley, sort of Ronnie Moran, Joe Fagan era, and they were meticulous note-takers, aren't they?

3:10Yeah. Yeah. I think that came from Shanks and the boot room was together for so long and they used to always go back to the books and the training books. I think a lot of that was more to do with actually the training than the actual games really and how they went about things and I think they'd look back and maybe when the results weren't going well, what happened years before, when the results weren't going well, what they did in training, maybe replicate that as well. So yeah, they were definitely ahead of the time at that stage i think kenny you know mentions that there was nothing like that anywhere then at that stage he'd obviously famously been managed and coached and looked after by jock steen the great jock steen and but yeah he said you know coming to liverpool was a completely different kettle of fish okay right eric there's the same question for you we're moving forward a generation sorry to do this but you know take me back to the the the initial period because there is a Jurgen Klopp moment in your career which is transformational.

4:07Let's go, let's get there. Yeah, it was 2013 when I first came like professional player for the first team for Dortmund and before I was a striker my whole life. So yeah, he came to me after training, I was training with the first team and he was just like, you know, on this position maybe it's a little difficult to reach the highest level because you're maybe a little bit too thin, you know, like not like... You were a forward, you were an attacker? Yeah, I was a striker. So he wanted to show me and make analysis with me for the left and right fullback position because he said you can run a lot, you can sprint a lot, so he really wanted to get this done with me together and he always had my back and I was so delighted and happy that he saw this.

4:59So I didn't really thought about it. I just said, let's start it. And then normally I would have to go to the second team to prove or to learn that I can do it. But at the end of the day, it worked much faster than we both thought. And yeah, then I made my first Bundesliga game in 2013 against Augsburg. And yeah, 2014 I was part of this Germany team who won the World Cup in 2014 and yeah, it was just an unbelievable journey. I never thought about or never dreamed about something like this. So you went from, the speed of which you went from, okay, I'm a center forward, now you're a fullback, and then within a year you're in the World Cup winning German team in 2014.

5:48It's a pretty extraordinary turnaround. Yeah, it was also surreal. so for me it was of course I didn't got the minutes to play in the World Cup but for me it was just everything to be part of that part of that team that feel the team spirit so it was for me it was like a dream come true but as you said it was very very quick so I couldn't really think about it I think it was good at that time and the same question I asked Jamie in terms of what data analytics meant at that point for me we had already the vests with a GPS in it, so they really could track the speed, the recovery time, how many meters you make, how was the tackles, the in-fights and stuff like this.

6:36But for me, it was always, I remember when it was 2013, I always wanted to have it directly on my plate and wanted to check how far, how many kilometers I did and stuff. But that time, the AI was not there, so I sometimes had to wait till the next day. when I was on the recovery bike with the guys. And then we all sat there and looked like, ah, you did this, you did that. And so we compared it. So I think this is a big thing for the AI that you can more or less have it directly live during the game in halftime or after the game. And you can directly tell from the coaches to the players on the pitch, you have to stand a little more on the right.

7:15You have to accelerate a little more. I think it's very, very great. And Jamie, before I go to Niccolo, So what was it like towards the end of your career? I mean, that period, 2013, did you, was the end? So that is a transformational sort of night and day, isn't it? Yeah, I think in the early 2000s, Prozone came in. That was like, it almost felt like the first thing that could track the data of players, you know, how far they'd run, really. But then you get a situation where certain players were just making runs across the pitch, just so the numbers would go, hi there, Daddy, it's true. So they were gaming there?

7:53Yeah, I remember playing right back, and Danny Murphy just made a run across the bay. I'm thinking, where are you going? He said, no, it's good to get me stats up, get me sprints up. So it was at the early stages then. I think it was seen as groundbreaking. I think it was Steve McLaren, certainly in English football, who was the first coach behind Pro Zone, and that sort of pushed him up the ladder in terms of getting a job at Manchester United and getting involved with England as well. But then it just sort of gradually kept just moving the dial more and more. As Eric's mentioned there about sort of, you know, the tracking data, that wouldn't be used for games, that'd be more for training.

8:32We wouldn't put those on for games, but yeah, you could just see it was slowly growing. So Nicola, I want you to build a bridge between the player experience and actually what's happening on the pitch now. So just bring us up to date in terms of what is being taken from a Bundesliga match or a Champions League game. Yes. So it all starts from tracking. As they said, devices, either with devices or with cameras, you're basically getting the video feed and you have this first computer system that can, from the video feed, track players where they are. We got to a point where you can track things very accurately with cameras to the point that you get 20, 30 skeletal position for every person, 60 times a second.

9:19So you know acceleration, speed, how fast you're running, how fast you're reacting. Same for the ball. You know how the ball is spinning. You get to know all of these things and you can basically, at some point, the big step that is happening, already happened in some form is the going from just the numbers. As we're saying, oh, you know, maybe you just game the number running more to how does these things are impacting your game? So like you're running, but when you shoot, like how is the ball going? Was that a good shot or not? How are you in position? How are you fitting the system? These are all things that are happening.

9:55Also for, I guess, health and recovery. Are you landing on the same feet? Are you with two feet? Are you accelerating too much when you land on the pitch after a jump? Is it hurting you? There's many things, even how tired you are. Maybe it's time to pull you out because you risk getting injured. These things you get real time now, like how your speed is decreasing during a game. These are all things that really just deliver better insights to the coaching staff and to scouts and everyone. I think there's a question about has it changed the game in terms of the way in which managers look at the game and look at strategy and the way in which the game on the pitch that we see has evolved?

10:41Oh, yeah, I think it's completely changed football. I think especially within recruitment in terms of buying players. I think the days of managers and coaches traveling the country or Europe watching players in the stadiums, I think it's long gone. I think that will still happen, but I think most of it now will be, first of all, who comes up high on the numbers in terms of what they're looking for, what sorts of data or suits the style of play of the coach. And then they'll identify sort of four or five players and then maybe go and watch the players. So I think it will definitely start with data initially, as I think, as I said, when I started playing, it'll be about watching people in the flesh and getting a feeling there.

11:23Whereas, as you know, the people will say now, you know, the numbers don't lie. It's there for you in black and white. What's it like being a player being analyzed to that level? He's talking about 30 odd skeletal sort of data points How does it impact you? Do you enjoy that process? For me, I think when I look back on it now I would hope they start a little earlier with the AI because I was injured almost my half of my career So they maybe could tell me before like injury risk management like settle down a little bit like Eric is in a red in the red zone or whatever. I was in the red zone all the time, more or less.

12:00But yeah, I think for me, I'm a guy who always like try to everywhere get like support or take something in to learn. So I think it's a great tool to learn. It's great things to know what happens with the game or with the accelerations of the game or with the whole team structure. if you see like my heat map for example is like all the time on the left side what happens if i go a little more to the right side that you can change it in halftime the coaches tell you what it what it means to the game what it can bring to the game when you stand a little little more open and stuff like this and this is what the ai does yeah like it's like a little assistant coach never sleeps it sees everything and um yeah it's but it's up to the to the coaches how they deliver it to the players and how the players take it in.

12:52So that's a good thing that this is, yeah, it's like a, it's like an assistance, you know, it cannot work without human beings. You know, it's like, it's like, um, you have this data as, as you said, Jamie, um, it's black and white. It is there, but you have to deliver it. And that's the nice thing about it. I think lots of things are changing, but one thing doesn't change it, which is like in sport, you always trying to find the most minimal edge to win against your opponent. And whatever you use, this data is just useful. You need to use it in the best possible way to figure out how you get a little advantage to beat your opponent.

13:26Because one of the worries I have, or is generally, is that the big will get bigger. So is the data analytics question just going to, is that the edge? When everyone's got the same data, where is the edge, do you think? I think actually I have the opposite view. We were talking earlier as well. I think this helps democratize some stuff. There's like many examples of many teams that have been able to play from a kind of medium level to really exploit the data to find the best player that fits their system to really leverage analytics in a way that you don't need to spend big monies on the star players to find and really optimize your performance in your Tactics to win this is I think really like democratizing some some aspects of the game Of course, there's always gonna be starts at the end.

14:12It's a human game Jamie, what do you think about that, the balance between where the edge is now and where is it going to be?

14:24When I look at the teams at the top end of the table, namely my own team, Liverpool, who have supposedly been ahead of the curve, if you like, in terms of data analytics, in terms of buying players, the Moneyball project. and even though I say this through gritted teeth it's probably in some ways for the game be nice that a team who have spent 400 million this summer end up running away with the league or beating everyone and it shows it's still a team game it's about being sort of clever it's not about spending the most money it's about getting the right players creating a sort of a unit a connection between what's on the pitch and off the pitch as well and it's really interesting Liverpool this season because as I said, for so long they've been seen as the people who are ahead of the game data-wise but it just hasn't happened for them this season and we're trying to work it out.

15:22I'm not sure they can work it out at the moment as well so it shows, even in this sort of data world where you feel as if you're trying to make less mistakes because the data's there for you, at times it can still go wrong. It's quite hard for a Liverpool or a Man United or to do a Brentford-Brighton type model, isn't it? So buy low, sell high, move the trading of players in that way because there is not data. It's the fan ego. It's the club ego gets in the way. Do you agree with that, Eric? 100%. I think there's an example in Denmark. My wife's from Denmark, so I know a little bit about the Danish football as well, about Mid-Jylland.

16:06I don't know. I hope I spelled it correct. And they won the league, I think, four times in the last 10 years. And this is a club I never heard before. Like, they really bring the young players, the young talent players you never heard of. And all of a sudden, you see them in the Champions League performing and winning the league in Denmark. And there's a big club like Copenhagen competing with them. And this shows, like, that this IA in especially what Jamie also said, that they figure out like with the data, like who is on top, this fits to the team, and okay, we need a left back. Who plays a little bit like Alphonso Davies?

16:43Let's put it that way. And they show you, okay, this and this is the best options, for example. And this saves you time, and you can really, like for a coach, it's really nice to select your team after data as well. But of course, the human being has to fit as well. Like it doesn't bring anything if the data is nice, but you come in and you cannot work with the guy because he thinks he's better than the others, for example. So it's always like a combination, but it's definitely a good thing for a club for Mid-Jill and they showed this Moneyball experience. They showed it is working or it can work, but if it works for every club, you never know.

17:21Then everybody would do it, I guess. Yeah, yeah. There's a television question. So what's it like? I analyze your analysis sometimes. so I've got the you know I've I'm getting I'm looking at data feeds primitive ones whereas you're doing it lie in life what's that like what just take us inside that that world no I I love that especially you know the show I think you turn it on to Monday Night Football is probably the biggest data analysis show around football probably I would I would imagine anywhere in the world really yeah so we pride ourselves on that and all we're trying to give is the viewer is what we think and we know that the players are getting in the clubs.

18:04So that's the sort of data we're trying to give. But as you said now, I think with social media, I think over the last four or five years, I think even supporters now of clubs understand sort of the business side of clubs and what they spend, the type of money that they're bringing in, the type of players they're trying to buy and what they're trying to do and this sort of model that clubs want. Now, every fan wants their club to spend lots of money. But I think there's a lot of supporters now up and down the country you look at teams like a brentford or a brighton and how they go about things uh data wise and probably one club actually in the stadium that we're in right now there's a lot of disgruntled i think west ham fans right now and probably a lot of that is they feel that maybe at times their club isn't run or maybe at the the top of the table in terms of the data analytics and they see teams like brighton uh and brentford who are not as big a club as what as West Ham but overachieving more compared to their club so I think fans are a lot more aware of that right now yeah we try and provide that in terms of the analysis on the pitch as well within the game but I think the level of understanding of the game now as well for I think from supporters I think they're probably a lot more educated maybe because there's more football on TV there's shows like Monday Night Football social media as well so yeah it's a great platform for us and as I said hopefully it gives you know people at home a little bit of understanding of maybe what you know a team meeting would be like on a Monday morning after a game at the weekend someone sometimes the sort of the data and I'm wondering what the most valuable jobs are going to be at a football club sometimes a data analyst job it's a sort of back page story when you know one go one is poached by another club well it always felt when a manager left the club the most important guy he would bring with him would always be his number two as assistant manager that was okay this guy always has to come with them.

19:54Now it almost seems like the data analyst he brings with them. It's almost as important as the guy who coaches with them or is his general manager. So yeah, I think that the data analysis now is a huge part of the game and there's no doubt when a coach leaves a job, goes to a new job, that guy is a big part of that team. Nicolau, why can't other clubs just copy Brentford, Brighton and the other models? I want to push us into the forward, into the future, into this AI question. What does it mean when you're laying that into this question? Well, it's complex. I mean, I'm not sure why. You've got four minutes, 20 seconds.

20:33Okay. I'll try to start with it. No. I think today there's still a lot of human aspects, like AI is still a tool that is enhancing what the humans will do and how you use your data. Of course, there's data and there's insights, and you need to use those insights to really make decisions. What if we're looking ahead into the future? What SportRadar has started to do is building this so-called foundation model. There is a technical name, which is the same technology that you use to build CharGPT or Gemini. And basically, instead of predicting the next word when they're giving you those answers, you use these models to predict what's going to happen next in an action, right, using this tracking data.

21:15And to do that the same way the charge GPT-2 answer, where you need to understand the world, this system needs to understand the game, which is the big next step. And understanding the game really means almost like a chess engine. That's maybe the easiest way to explain it. It's like a chess engine understands chess. Chess is super simple. There's just eight things and moves one at a time. But there, at some point, you can really start evaluating how every move, every step in the field, every pass, everything, like what is the impact? What's going to, what's the, like how it improves this specific action a little bit?

21:53Or even just the mere presence. I can imagine that Messi being in the field, not moving, already attracts attention of three people, right? Just because merely being there, even, I don't know, sitting. And this system eventually will understand all of these things and will probably also start coming up with new tactics. What's the best way to react to this lineup? What's the best way to play given this player? So what is the best player to hire that is now in, I don't know, high school that will fit or will have a career trajectory similar to this other one? But again, this will tell you on average things.

22:31It will never decide how you're gonna score a goal. What magic, what act of crazy dexterity you're gonna invent. That's the beauty of sport. How you're gonna be able to solve a situation that even to that system sounded unsolvable and you do like, you know, or someone makes a bad mistake at the wrong time. And still that's going to be the beauty of, you know, how unexpected the game is, even if you can predict statistically how things are going to go, you never know what the reality is going to be. So what's the role of a manager in that environment? You know, if we sort of push that forward, Klopp was a huge influence on you.

23:12he feels like someone who's got enormous emotional intelligence he feels like someone who's not just the data bit of it but that bit of the emotional part becomes more important when every when everything is being modeled yes absolutely we're looking at a different i can't predict 100 years ahead for sure but for now these models are never going to be able to understand the psych the human psyche and uh i believe that emotional intelligence plus the tool plus the understanding of what these tools are telling you are what's going to make the difference like you still need players to be bond together you know eric was telling a story it's like you know when klopp did the right thing at the right time and that's like still super valuable and it's going to continue to be that it's quite a nice conclusion to to have that human beings might still be relevant you know thank god we still need jamie carriga we still need monday night football um very finally cbs galatso the that program pointed at america it feels different to monday night football what is the difference what brief do you get for that program because there's a question about what america's relationship to football and how that's going to be obviously you've got the world cup next year it's probably the complete opposite end of the scale to sort of monday night football that's why i enjoy doing so it's a one on a monday and i can go to the other on the Tuesday and Wednesday.

24:35You just mentioned Monday Night Football, really data, really about analysis and numbers, whereas this, as I said, is not taking the game too serious, really, and that's what's worked for us. I don't think that would work on English TV. I think it works because it's the Champions League and not every fan watching this team is in the Champions League, so it's not as tribal. It's going out, obviously, in America. it's not really affected by the tone of the tournament is a is europe and uh and yet it just seems to work you know a nice sort of brand of you know different people bringing different things to it but i would say the producers on that show are always taking risks always sort of pushing us as close as we can get to the line and hopefully not crossing out at times and listen whether we like it or not social media plays a big part in in all of our lives and and they really hit home on that in terms of it feels like it's built for social media yes that's exactly what it is so you know everyone in this country knows of it but it's not obviously on a on tv over here but people see it just because of social media clips and as i said whether you like it or not it's very powerful social media in terms of getting messages across or you know getting stuff out there and certainly from our point of view, yeah.

25:55Perfect. Right, listen, we have come to the, you've got to get a plane. I want to thank the group. A round of applause, please, for our panel. Thank you. Thanks, everyone.

26:15I just want to go around the table, just say who you are, what you do all day, what's your job? I'm Darren Small. I'm the Senior Vice President for managed trading services within Sportradar. Okay, I'm going to ask what that means in a minute. Okay, great. My name is Nicky Baia, and I'm the SVP Marketing Services in Sportradar. Okay. Hello, I'm Nicolo Dercole, and I'm EVP for Technology and AI at Sportradar, yes. So, can you just explain what happened today? I was trying to work out whether people were being optimistic or pessimistic about the state of the marketplace. What was your sense? Define marketplace?

26:48Soccer? Well, you define it for me. what are the marketplaces well i mean if we talk about generally i guess soccer and and the popularity about soccer because in the end the event was about it was about soccer it was largely optimistic or or at least certainly not pessimistic at least that was my impression i mean obviously there are things uh that is changing this world as the world generally is changing when we talk about technological developments like ai for instance so i think yeah from that perspective sure there is influences um there's developments but i think generally it's a positive one yeah The AI question, Niccolo, is one that is just coming up all over the place, not just today, but just in every conversation that you have about business, any type of business.

27:30And you're at the sharp end of that. How are you sort of dealing with that? Because we talked about it on our panel with Jamie Carragher and Eric Derm. But there is a sense that we're only scraping. We can only scrape. I mean, I asked you something about the future of AI and I gave you about four minutes to answer. Yeah, yeah. No, I will take less this time. No, kidding. I thought what was really great today was like the two football players, former football players, and now involved in media or fan engagement in the U.S., they really love the impact that technology is having on the sport. They're not opposed to it as like players in this kind of market.

28:10And so this was really great to see. They were super aware of the technology advancements, of how it can help the sport, the performance, the coaching, how it can help the media aspects of it and the fun engagement aspects of it. So this was really great to see. Who is going to win and who's not going to win in the terms of AI, the race that we're in? And I keep coming back to whether or not it's to do with the quality of the data that exists in the first place. because i was talking to someone from the travel industry the other day and they were saying a lot of people are just putting bots on top of what is a pretty crappy data set and just hope think you know fingers crossed and hoping for the best i can sense that that might be there'll be people around the sports industry are saying the ceo wants a really flashy ai you know he's she has heard about ai and wants to have this interface and wants to have the the bells and whistles but we haven't spent the last 20 years getting our data in order i think that's the opposite for sport radar particularly yeah i mean it's been it's core to our product we've spent a lot of time structuring and forming our databases to make sure that the data is good and it can be read and it can be interrogated a lot everything we've done until now has always been about delivering high quality repeatable data to the end user so that they're able to use it in whatever process they want so we're pretty proud and pretty pretty confident with the fact that the data that we have it and the way we continue to collect it will be able to support us in the need around that AI and that pursuit.

29:40Yeah for sure of course there's always improvements in terms we have to do breaking silos making all of the data that we have compatible to be accessible by AI which is things we're working towards do. I will also want to say that many companies just use models the textual models like the chat GPTs of the world to build experiences chatbots or content generation. Of course we also do that but what we're doing that as well of course other companies in other kind of markets are doing is actually also build different kind of models like the ones that learn about the game and these are not they use similar technologies similar architectures but they're completely different and how do you make sure that the bets where the odds we're providing to bookmakers are like ideal to optimize for profits etc or how you really understand the game in a way that almost start to resemble the best humans in understanding of the dynamics of the game that's really what we're trying to achieve and so we're going a little bit beyond just giving some data to an existing model from google or open ai or entropic and then just make it talk about it what's going to happen when this is more i guess for the retail and the sort of betting the sports books question and it's a differentiation question about what the difference is at that end because you guys are sort of back a stage from that as i understand it you're the sort of plumbing between the rights holders that you're collecting data on one side and presenting it to the market on the other but we that's exactly that's a really good question because that's exactly what we're trying to do with everything we're doing like from a from a sports betting and odds perspective from where i sit we are doing a lot of personalization to the to the odds offer through ai services we're changing the way that customers can integrate their products that's because at the moment i mean and you just put me right because again my assumption is i mean i'm using the travel analogy but i probably won't need to go to a travel agent anymore i'll just go to chat or i'll go to some AI.