Sportlogiq is a company with big plans and a product it believes can fundamentally change the way sport is consumed.
Architects of an artificial intelligence (AI) analytics platform that tracks live video in real time, the Canadian firm has been developing its software in partnership with a number of National Hockey League (NHL) teams over the past couple of years and has just signed its first league-wide deal with the Swedish Hockey League (SHL).
Now, Sportlogiq is turning its gaze beyond the ice. Not only hoping to take statistical measurement out of the hands of humans and put it into machines, Sportlogiq wants to provide deeper insight into sport as it happens. SportsPro spoke to the company’s chief executive and former Olympian Craig Buntin (below) about his ambitions for the future.
How did Sportlogiq come to be?
We originally started this company with the intention of building technologies for self-driving vehicles.
As with any startup you push as hard as you can push, and when things don’t work you shift and move. The more we started working, the more I realised that we’re sportsmen who knew very little about cars.
So, in the early days we were just taking whatever it is we could get our hands on and coming up with pieces of technology that would basically identify human motion and interaction. Applying it to sport was just where we had a customer who was willing to pay, so that kind of moved it in this direction.
When did you realise that this could be something that could be picked up widely?
Immediately. We wouldn’t have started the company without that goal. The way that the company got off the ground, I’d basically shopped the idea around probably to six or seven universities, basically walked through the hallways of engineering departments, looking at posters and meeting PhDs and just talking about where the future of AI and computer vision is going, and in that process I met our co-founder Mehrsan Javan. He was just defending his PhD thesis in computer vision and technology design for security surveillance, and so it did anomaly detection in video.
We had this sport video and you could detect all the different types of shots – shots that were really good, or really bad, and those being anomalies. So, we started exploring that and we realised very quickly that the industry itself was just right for transformation. There are three companies in the world who are generating over a billion dollars of revenue, and they’ve got thousands of people manually tagging video.
The Optas of the world, there’s Sportradar - a number of companies who have hundreds of thousands of employees tagging videos. Now, tagging videos is fine, it does give you that human context to it for what’s happening, but it doesn’t give you the entire picture.
So, let’s use a pass as an example, if you had a player running up the pitch, receiving a pass, and then running with the ball making a shot, if you could tag every single one of those over five years you can begin building a profile of that player, and maybe heatmaps and some insights. What you don’t get is all the context: if a player received a pass, was the player under pressure? Were there other passing lines open? Was that the ultimate play that that player should have made? Were there other defensive players on the other team drawing opponents and closing down space? All of the contextual pieces of that pass are just simply not being captured in the industry right now.
What we wanted to do really was build a machine that could see, understand, and describe the game the same way that a person can. And unless you can really understand contextually everything that’s happening on the pitch, and every motion of every player, you just can’t do that.
Sportlogiq has been working with a number of NHL franchises over the last couple of years
How did you develop the product?
Mehrsan and I basically wrote that on a wall, and we said, ‘What does the future look like? What will sport look like five years from now if all these different technologies exist?’ We then split that vision for where we were going into eight different areas, eight specific types of technologies that needed to be built, and we brought each one of those to a different university lab across the country. We got some government grants, we funded all of the research at each one of those labs, and we basically brought on eight of the top researchers in the country. This was about building a technology that would fundamentally change sport. That’s where we started, and that value has not changed.
So the Swedish Hockey League is your first competition-wide deal, have you tested it in other markets?
Hockey is the perfect testbed with the way the sport moves, there’s the flow element to this sport; similar to basketball, similar to football, you really see the similarities in the technology itself, or in the way the game is played. We’re a Canadian company who of course know hockey very well, but it was never about starting a hockey company, it was about building a technology that we could then scale across multiple sports.
In our first two and a half years we signed on 25 NHL teams, so we now are really the primary product used across NHL, with five major broadcast networks [using us] in terms of our storytelling ability. We began predicting game outcomes with over 67 per cent accuracy, and so we really cut our teeth.
We looked over to Sweden as this sort of perfect [league-wide] starting point. What’s really incredible about the SHL and about Sweden as a country is they say, ‘Okay, they’re our broadcast partners, they’re our sports spending partners, they’re our teams and we’re the league, let’s all get into the same room and talk about how we can use your technology to benefit everybody, every partner across every vertical that we have’. To really transform the way that the sport is played and engaged with, you have to have that buy-in.
So the aim beyond Sweden is to move into soccer, is your product scaleable to that goal?
There’s so much that is yet to be done in soccer, it’s just a huge, huge sport, period. So, we saw that as a perfect area to really, really be the best sport tech company in the world.
If you look at countries in the world where football and hockey are both big, Sweden is one of the few areas in the world where fans are huge fans in both of those sports. So, it was a perfect starting point for both hockey and soccer there, but we are very, very much engaged with multiple players in the UK right now. Even the Premier League teams that we’re working with right now, the interest, the enthusiasm, the excitement is exactly where it was when we first started in the NHL a few years ago.
We’re setting up an office in London quite literally right now. We are hiring, we’re looking to bring on the best people in the world, and we’re looking at London as really a long-term player for us.
Soccer, is just a huge, huge sport, period. So, we saw that as a perfect area to really, really be the best sport tech company in the world
What can Sportlogiq change and bring to soccer that’s not been seen before?
Firstly, our tracking system is currently the only fully automated real-time player tracking system in the world. We’re operating on a latency of milliseconds, when our competitors right now are coming at this delivering data within hours. So, just being able to flat-out acquire the data and deliver the x-y positions of the ball and the players in real-time, for us that’s the foundation on which the company was built. Now, what we can do with that, if you look at a fast-dynamic player, somebody who has run any number of kilometres in a game, that might be a stat that a fan might get right now, and you have some sense of who’s running the most. Data is only data, what really matters to fans is stories, so understanding context and having it broken down as to why it’s important, or how one player is using running.
The second thing, I mentioned earlier a video tagging company might tag all of the passes, well we can tell you not only where the passes were, but how many times where a few players will have put pressure on the passes. So, is a player a strong passer under pressure? Is a player making better decisions, i.e. making the passes through areas where he’s got more space? Is the player making successful passes when there is no space to make those passes? Those are the kind of metrics that we can now deliver, and the type of stories that we can now tell.
Those insights, are they going to be focused for broadcast, or for teams?
There is a fine line to walk when you’re working with both teams and broadcasters. If you’re watching ice hockey in North America right now, you’re seeing our stories, they’re coming out pretty much everywhere. There’s a strong chance that we’re working with both the teams that are playing in that game. Now, what we’ve found is that teams will dig really deep, they’ll want to know details of where the spacing is on their opponents etc.
What we found when telling stories with media, is we find the things that really matter to fans. Take goals as an example, let’s say there’s three different types of ways that these teams will generate their goals, and maybe one team might have higher expected goals tonight, that means they’re taking higher quality shots. Now, those are the types of things that are going to be obvious to teams but will be really interesting to most fans. So, being selective in terms of how deep you can go down, is really that line that we do walk in hockey. We’ve been doing it successfully in hockey and our assumption is that we’re going to do it successfully in soccer.
Buntin says Sportlogiq has to walk a fine line in providing data insights for fans without giving too much away from its partner teams