Collision Course: Sports Betting + Data Science

Key Takeaways

Hello, everyone. Ken here, back with a new video for you. Today, I'm talking about the impending collision between uh data science and the sports betting market. This past weekend, I watched a really interesting docu-series called Action on Showtime, which profiles a lot of the big players or some of the big players in the sports betting market in Las Vegas and in Atlantic City. Now, I think that from a data science perspective, this is going to be a fascinating market. There's a lot of opportunities there, especially after the federal government removed the ban, which allows states to actually monitor or manage their own sports betting activities. The individual states can determine if sports betting should be legal or not. And in the next couple years, we're going to see a lot more states, including Illinois, where I live, legalize this activity. Please hit that like button, because it really helps me with the YouTube algorithm. And if you want to see more content at the intersection of data science and sports analytics, please consider subscribing to my channel. I've been involved in sports analytics for some time now, and I've always been fascinated with the sports betting market. Through my work, I'm privy to a couple different private data sources, so I'm unable to actually bet on these outcomes, but I still think that there's a lot of really cool opportunities, and I'd like to talk a bit about the market and how the data science skill set can integrate with it going forward. As a data scientist, sports gambling is one of the purest types of problems that you can attempt to solve. In theory, you really only have to beat the general public opinion about a bet 52.38% of the time, roughly. Now, that's a lot easier said than done, and you don't always see people being profitable. A lot of people have lost a lot of money doing this. Still, trying to hit that mark, trying to get as much edge as you can, is very fascinating to me. Sports books individually are also very interesting. Now, the reason why you have to beat the general public around 52% or 53% of the time is because sports book, they take a little bit of commission on each bet you place. This is called the juice by some people. Now, the juice can be different on different bets and across different sports books, but generally it means that you need a 2% or 3% edge. Because of this edge that the sports books create, they want to have roughly equal people betting on either side of the odds, so that they almost always take that commission cut. They're not trying to necessarily beat the bettors, they're just trying to take their cut every single time. And with enough volume, they will definitely make a lot of money. This is a slight oversimplification. Some casinos and some sports books use a little bit higher math and machine learning to optimize and to potentially take a greater amount by, you know, ef- effectively betting against the individuals, but that's probably some logic in a story for another time. Because sports books are trying to optimize locally and maximize their their earnings, this means that globally different sports books can have different odds for the same bet. That's a really good thing potentially for sports bettors. That means that you have options, and based on what your model says, you can choose the line or the odds that favors how you would like to bet. Now, I'm interested to see how this changes as more states legalize sports betting and there are more online options for this. I think that if there are a ton of online options, the marketplace will likely see odds kind of converge, but that's to be determined. Additionally, with the sports books, you can see the lines or the odds change and move over time, even within a sports book. So, if you You understand consumer sentiment, you can look at tweets, things like that. You may be able to predict when your odds will either improve or get worse before the actual event that you're betting on starts. From what I've seen, most sports bettors, even the ones that are really making a lot of money, aren't using overly advanced models. And maybe they're not needed. But, at the same time, I think that there is a great opportunity to create an edge using some advanced machine learning and from actually expanding and scaling and collecting data, etc. So, there's a great fit for the data science skillset here. Now, there have been companies that that have tried this. There's a company called Jambo's that has got millions and millions of dollars of venture funding to attack this market with a machine learning solution. Unfortunately, it looks like they haven't produced the type of results that they were guaranteeing, which is a little bit concerning and is obviously a risk. But, regardless, people are willing to pay money for these solutions, and there's a lot of opportunity, potential opportunity, and even arbitrage here. I wouldn't recommend anyone to jump right into sports betting. I probably wouldn't even recommend that you start sports betting in the first place. I just think that this is a fascinating problem that is a great test of your data science ability. You can still test to see if your models would have beat the odds. The odds are generally very public, and that's good enough if you're just trying to improve your skills and evaluate how you can tweak different models. If you're interested in learning more, getting some data on this, you can go to my website, link below, playingnumbers.com. I have a bunch of free data that I've collected across the internet. I'm trying to make a really good starting place for anyone who's interested in data science, sports analytics, sports betting, any of these markets. So, go ahead and check that out, and if you create a cool solution, and you know, please submit it to us. We'd be happy to share it through our website. Thank you so much for watching. Again, I think that this is a fascinating market that you should perhaps learn more about, especially if you're interested in sports. Until next time, good luck and enjoy.