DENVER – Analytics and big data are changing every industry, and you can bet sports aren’t the exception. Enter SciSports, a Dutch-based sports analytics startup that is quite literally changing the game of soccer.
SciSports has a plethora of offerings that range from Datascouting and Intelligence-as-a-Service to Insight reports and their latest 3D data capturing technology, BallJames, and at SAS Global Forum in Denver last week, we interviewed CEO and founder Giels Brouwer to talk about how analytics are changing the world of sports for the Tech in Sports Podcast.
You can listen to the interview on Tech in Sports Daily by just clicking right below, or even check out the full episode on Tech in Sports right here. But if you’d rather read the interview, check out the transcript below.
The following transcript has been edited for clarity.
Tech in Sports: Alright everyone, I’m really excited about who I’m speaking to right now, because right here in front of me I have Giels Brouwer. He’s the founder and CEO of SciSports. Giels, thank you so much for taking the time to speak with me today.
Giels Brouwer: Yeah, lovely to be with you.
TIS: We’re in Denver at the SAS Global Forum. You were up on stage talking and presenting in front of the whole conference about what SciSports does. So I know you guys are a Dutch company. So if you can first explain just what SciSports is and what you guys do.
Brouwer: SciSports is basically helps football or soccer clubs … I’ll keep it simple, but I’m gonna call it football. We help them to make better decisions by using data. We developed two things. The first one is an index in which we are able to evaluate equality and the potential for every player in the world. And the second one is BallJames. That’s a new tracking solution where we can do real time 3D tracking of every player, and the referee and the ball are 3D.
TIS: Just looking at your website, I looked at the expertise tab, and I saw some things that really, really fascinated me. There’s datascouting, and then Intelligence-as-a-Service. Can you just explain what you mean by those?
Brouwer: Yeah, datascouting is basically when you have a lot of clubs in Europe, they don’t have the capacity to hire an entire analyst team or analytics team. So we are their data scouting department. So we sit down with the technical director of the club, and we help him by giving him advice on which player he should sign or sell. In the end, data is just a tool. We will never say you should sign, but we say data suggests that these are right players to sign. And then he makes a decision. But they are really part of the scouting team.
The Intelligence-as-a-Service is basically an enrichment of that. So we built the data intelligence for the clubs. We get data in and we give advice out.
TIS: Got it. I also saw … just to go through the product list, I saw that there’s Insight, SciSkill, SciSkill Performance Indicator, and then BallJames. Do you want to just give the quick pitch on each of them?
Brouwer: Yeah that’s fine. The SciSkill is really a tool to evaluate the quality of every player in the world. I used to play a lot of Football Manager. I was addicted to it. And there was actually the need to found a company, because I wanted to rebuild Football Manager with real data. And right now the SciSkill is able to find which is the brightest talent in Chile.
With the SciSkill Performance Indicator we go one step deeper. Then we can say, “Well, this player, he has a creativity of 97.” And then we see that he’s world class. So that’s one step deeper.
When we go to the BallJames project, we can even see the ball speed, or the fielding angle of a player, or how far the ball dribbles from the foot when he’s running, whether certain miles per hour or something.
 |
 |
 |
TIS: What’s the most specific stat that you can get, and what’s your favorite one that you can pick up?
Brouwer: That’s a good question. I think one of the most favorite things we developed was to … we call it the passing value. With that, we can calculate the chance of being scored within 15 seconds from every pass. So we can see the contribution of every pass to the goal. We train over 100 million passes to get this model done, so that was pretty cool.
And the second cool thing, I guess, is … We have a horrible name for it, but basically what we calculate is where a player dribbles inside the final third of the opponent, is he able to give good action to his teammate in the final … in the box area. And that’s just one metric that’s combined with a lot of different variables.
TIS: I imagine … so you’re only doing soccer, football for right now. I imagine this could be applied to many other sports as well.
Brouwer: Yeah totally. The company is called SciSports and not SciSoccer. But we’re getting there. But first we want to make sure that we change the football industry and then you will see us in other sports.
TIS: So we are here at the SAS Global Forum. So if you want to kind of explain how you’re using SAS there. For our listeners who don’t know, they’re a huge analytics company. One of the biggest in the world. If you want to explain how you’re using SAS in your technology.
Brouwer: Yeah, basically what we … we’re a small company. We’re now have 50 employees. But we built a solution that was applicable in one stadium. We used SAS together as a deep learning partnership. We combined joint R&D efforts to build a solution that’s able to be installed in entire leagues. So we used their Viya platform and their events stream processing unit to make sure that we can stream the data in real time from the pitch to the media and to every sports fan in the world.
And together with them, we managed to scale down. We used to generate 56 terabytes of data per match. And you cannot stream that in real time. So we developed a system that’s able to put a GPU on the camera and do the calculations on the camera, and stream the outcome to the club.
TIS: So with fans … it’s easy to see how clubs can get excited about this technology. How can fans get excited about this tech?
Brouwer: Well it’s, again, a gaming thing. I’m actually a big gamer, really fanatic fan of many games. But also the game FIFA. I used to play the Dutch National Championships once. And what we can to do is bring the experience that you get in FIFA to the real football industry so that you can see in live radar on the screen when the match is being played. That you can see the speed of a player, that you can measure the agility and all those things. It’s the kind of things I want to know as a fan. And now we can finally make it happen.
TIS: Would it almost be at some point just like an app that they have on their phone at the same time while they’re watching? And then they’re just like, “Okay, this pass had a 50% chance of leading into a goal” type thing.
Brouwer: Yeah, yeah those kind of things. Even if you take it one step further. I want to be able to stream a match in hologram-style in real time. That’s what you can do if you track everything that happens on the page.
TIS: Okay, so would this hologram kind of just be on a table type thing?
Brouwer: Yeah, just on a table or maybe stream it in the stadium if… imagine all the things you can build with it. If you take VR, AR holographic explorations, there will be new users at the World Cup in 2022 I guess.
TIS: Yeah for sure. Actually, speaking of the World Cup, that makes me ask, are you doing anything with any national teams?
Brouwer: Actually we are. I cannot say the name of the country now. They’re a European country, and I think now they have a good chance of making it to the finals.
TIS: Alright, that’s awesome. So up on stage yesterday, I believe, or earlier, you spoke about a Dutch player who you helped or you gave data advice to. And he’s now on the national team, the Dutch national team. You want to go through that story?
Brouwer: Yeah, that was incredible. His name is Wout Weghorst. Try to pronounce it. He was playing at the Dutch secondary level. He was a striker and his contract was not renewed. But according to the data, we saw that he was always in the right position, he always scored the right goals at the right moments, and scored a lot of points for his team. And then we said to Heracles Almelo, “Take a look at this player because he’s for free. You can try to sign him now.” They weren’t scouting him and they said, “Well, he looks kind of clumsy, but still he scores the right goals.” They signed him as a first striker, and within one year he ended up on a Dutch national youth team. He became the top goal scorer. They sold him for $2 million to Dutch top club. He became the top goal scorer there. And was offered the Dutch national team, and now he was a value of approximately $10-15 million.
TIS: You see, that’s incredible. So this is all … you mentioned it earlier, this is all using data as a tool. Not the final answer. You’re just kind of adding another weapon in these players’ and in these clubs’ arsenals, kind of.
Brouwer: Yeah, that’s true. It’s not that you’re gonna buy hours based on the recommendations that they give you. You’re gonna see the actual field, actually talk with the people around. And I think that’s important for football as well. But it can really help you to find new things in the industry or get you some extra percentages. And in the end, data is getting really important in decision-making process.
TIS: I just now have a few other questions about sports analytics that my co-host Mandy and I, we talk about all the time. Have you noticed any push back on this technology? Do you ever go to a club and they’re like, “Ah, this sounds like it’s too much for us?”
Brouwer: Yeah, especially in the beginning. The football industry is really conservative. I think every sports industry is conservative. Maybe American sports are a little bit less conservative. But yes, sometimes when you go and talk to someone for the first time, you really have to convince them. Or they say, “Well, we just don’t believe in data.” They don’t even want to listen then. But fortunately, you see the last years, people are trying… understanding better that it can be a support tool instead of data being the goal or something. But I think in the end, the industry is changing so fast and they see that other clubs are using it and they are successful with it. So the rest will follow.
|
|
|
TIS: Right now, are you just in European clubs?
Brouwer: One club in South America, and currently we have three clubs in the MLS on trial. But no contract signed yet. And all the rest of the clubs are in Europe.
TIS: Got it, okay. Just off the top of my head, just speaking about it now I’m like, well the MLS in the U.S. And Canada, or TFC, at least, the Toronto Football Club where we’re based in, they’ve been winning. But they could really … I imagine they could really use it. Have you seen any interest at all in North America from this?
Brouwer: Yeah, it started with the last two months. We just opened our version 2 outside of Europe because we wanted to launch when we were ready. We validated our system in Europe, and now we’re ready to go also to the MLS. Yeah, as I said, there are three clubs already testing our platform right now. So I think there will be many more to come.
TIS: Mandy and I once had this ridiculous debate about AI and data. And we were wondering … we both didn’t think it was possible, but it was something a general manager of some team said, that AI could replace them. Do you think AI could replace a coach or a manager in sports?
Brouwer: That’s an interesting topic. I cannot give a definite answer. It will probably be something we can discuss for hours. I don’t think you should want an AI to take control, as the players will all have their own personality and you also give them … treat them well. You have to talk with them, and I don’t think they will listen to an AI. But if they would listen to an AI, that could be a solution. But I think it would be odd. I hope the human will always be the ones who decide.
TIS: Yeah, that was kind of the conclusion we had drawn as well. We were like, “Well, I guess it would be cool to see it happen.” But then half of a coaching job is to manage your players’ personality, and I don’t know if Siri or Alexa is gonna be able to do that.
Brouwer: No. It would be tough, but it would be really cool if the AI could really advise the coach on, “Do this now.” And then he just says yes or no. Because then it can also learn from the coach’s decisions and it can really become the assistant. So we should come up with a name and then let’s try it out.
TIS: One of the things that, just off the top, where you talk about the pass percentage. It’d be cool if AI was almost like an assistant coach, you know? Where it’s just on the side and was like, “This guy keeps making passes that are not… talk to him,” type thing.
Brouwer: Yeah, like that. Or when a substitute player comes in, and you see that his speed is … his acceleration is three meters per second faster than the player that you have against him. So you should replace your own player with a player on the bench, and then you can make sure that the player isn’t dangerous anymore. So those kind of things… as an assistant coach, I think AI would be perfect.
TIS: Okay, awesome. Alright, well Giels, thank you so much. This is awesome.
Brouwer: Thank you for having me.