SimpleBet talks sports data and its machine-learning pricing model

A look into SimpleBet’s AI-powered approach to pricing US sports’ in-play bets, as the company continues to poach industry talent from Europe and the US

Having poached a handful of well-known betting industry names in Europe, New York-headquartered B2B betting start-up SimpleBet is pledging to bring competitive in-play pricing to the US market through its machine learning-powered trading platform. Built by in-house engineers in the US and Sofia, Bulgaria, the platform aims to offer accurate and efficient in-play odds for US sports.

Executive vice-president of sportsbook product and revenue Leo Gaspar joined the firm from Romanian operator Superbet in September 2018 to drive product development and grow the team, which enticed the likes of Robert Simmons, a former principal software architect for The Stars Group; ex-VP of game operations and development at DraftKings Mark Nerenberg; and Nedda Katcheva, who also worked at The Stars Group as head of sports technology and then engineering.

SimpleBet is closing in on its first commercial deals in the US and is looking further afield to Asia for future business. It is currently licensed to operate a B2B service in Nevada and New Jersey (as well as the UK and Malta) and is not putting a great deal of energy into regulating beyond that for now. Gaspar and CPO Nerenberg sat down with EGR North America at ICE in Boston to present the company’s business model and plans to influence the as of yet almost non-existent in-play and live betting opportunities in the States.

EGR North America (EGR NA): How has the process of building the team been?

Leo Gaspar (LG): We have around 80 people now and we’re split between New York and Sofia where most of our engineering team is based. We have 30-plus data scientists in the US and 40-plus software engineers in Sofia. It’s been a hectic couple of months for us growing from our very early stage. Our data science team is made up of very talented people mostly from Harvard and the Massachusetts Institute of Technology and they are led by Francesco Borgosano, who joined us from Kambi as chief analytics officer last December.

Our chief revenue officer is Stephen Master and he’s from the board of the American Gaming Association and head of the association’s betting task force. We are a good mix of people from a sports betting background combined with those from a US consumer product focus. We’re trying to combine the best points of view and recently hired some product people from Bloomberg and Google. We want to disrupt the industry and to do that you need to bring in people from outside sports betting.

Mark Nerenberg (MN): We’ve got top-tier data scientists that are really talented and are learning how to do more product development. They’re traditionally very academic. We have some people with general product backgrounds, others with sports betting industry experience and other talented engineers.

EGR NA: What stage are you at in the product development journey?

MN: Right now we’re focused on machine learning-based pricing for live betting and we’re going to have pricing for MLB specifically predicting the outcome of any at-bat with a couple of different markets. I think we’ll be live after the NBA’s All-Star break. We’ll have it ready on our end by June. We’re looking at NFL, soccer and NBA, but going to market with MLB and going from there. That’s our best chance to show off what we can do with machine learning.

“Our vision of the company is to price up every next occurrence within any sport or event, and in order to do that you have to use machine learning” 

LG: That’s all done in-house. It’s fully automated and there is no human intervention whatsoever. It took quite a long time to develop the engineering behind it. You need to use modern technology and there are not many engineers that know how to deliver on that. We’re happy to be going live with that soon. Our vision of the company is to price up every next occurrence within any sport or event, and in order to do that you have to use machine learning. You would need hundreds of traders to do it all manually and we wanted to make betting more intuitive for even occasional fans. Will also have standard markets priced using AI, but our focus is next occurrence, which is something that hasn’t been done yet.

MN: Users are going to be able to bet on all these markets and get a decision shortly after. Right now you can place a live bet but it will usually last for the rest of the game, which is popular, but this will add further engagement and allow fans to bet on the next bat or drive in football. Our long-term vision is anything you can imagine betting on in a sports game, we’re going to have pricing for that. That’s where the opportunity to increase the market and attract a new type of user is. The use of AI is faster and more accurate and more automated.

EGR NA: Is the US sports bettor ready for this level of choice?

MN: Consumers are ready but everything takes a little time to overcome, including data and latency. The data that is available in real-time will improve but there is enough right now to move into this space and try new things that are not already being done. A year or two from now the opportunities will be even more exciting.

EGR NA: Are you looking further afield than the States?

LG: We’re looking to launch anywhere, Asia and Europe included. With MLB we will aim for Asia and the US market mostly, but as we roll out our other products we will extend our reach. We are working on providing our product to B2C operators, data providers and B2B suppliers.

EGR NA: What enticed you to join SimpleBet from Superbet?

LG: When the US market opened up I wanted to do something different because coming from the European space all the sportsbooks looked the same and the traditional bookie was boring to me. I wanted to add a gamified experience. That will be the second stage for us. In order to deliver this gamified experience, we need to develop the pricing algorithm first and the UX part will come later. That’s the ultimate goal for me coming from the industry.

EGR NA: How did you poach others from the industry?

LG: With the opportunity to work on a greenfield project. The technology we’re working on and the stack we’re using has been built entirely from scratch, which means no legacy technology. The hype around the US market is so big and everyone is excited.

EGR NA: Why the decision to be headquartered in the US with such a big development team in Sofia?

MN: Part of it was the PASPA repeal, and a part of it is the live bet types we think will most popularly tie into US sports and the US consumer, and what they want. These in-game discreet occurrences fit very well with US sports where there is more scoring, more player stats and fans are used to the fantasy model.

EGR NA: What are the main differences between European and US bettors?

LG: Our belief is that players in the US lean more towards player betting markets whereas in Europe people tend to bet more on a team’s outcome. We are going to offer more on the player prop side of the market and hopefully that will drive more fan engagement and work better for leagues as well.

Baseball Game

SimpleBet’s machine learning-powered trading platform will feature pricing for MLB. Photo: iStock

MN: I think it’s two things: the maturity aspect and the fact the sports are so different. Those two things create totally new opportunities because the sports will create different types of bets, and because the industry is new, there will be more motivation for change and a new start. There is a lot more motivation to innovate here, whether it’s creating social opportunities or gamifying an offering through personalisation. Some of that is happening in Europe too but this is an entirely new opportunity.

In the US we are getting new official data providers and the feed is becoming faster. In DFS we used the media feed and that has all the player stats play-by-play and what parts of that they can get into the faster betting feed is what they are working on now. I’m hoping we will catch up. When is the pitch type and pitch location going to be in the betting feed? And when will we have substitution data about who’s on the court? We’re maybe a year away from that. We’re working with the data providers and they want to know what we need from them to have these engaging experiences. I’m hopeful that it will be fast. These providers are motivated to make this happen so it should be sooner rather than later.

EGR NA: Have you signed up any big operator partners yet?

LG: We’ve been talking to most of the tier-one and tier-two operators and we will soon be starting to sign commercial agreements. People in the space are excited about our team and what we’re working on. We just want to ensure when we deliver our product that is first class.

MN: We’re talking to some partners and will hopefully get the product live after the All-Star break and we have other products going on in parallel. We’re getting ready for Soccer and NFL. We’ve done a lot of work on NBA but it’s nearly the end of the season and it’s mostly internal testing, but we will be ready for the next season. There have been some interesting delays based on data availability. All the leagues are still figuring out their official stats providers and exactly what is available is still a little unclear. I understand it from their side too, but we would have been up and live a little bit sooner.

EGR NA: What is the answer to the sports leagues’ involvement in the betting industry?

LG: We need to use official data and it will be the only data we will use as a product company. We think products like ours will increase fan engagement and leagues will love it. We also understand the market here isn’t fully developed, but once it is all these questions will be answered. I’m sure everyone will soon understand the value of sports betting, including the leagues and the government. œ

Artificial intelligence | Interview | Leo Gaspar | Machine learning | SimpleBet | Sports betting