Steve Kuhn
Subject: Predicting Sport Outcomes
Bio: Entrepreneur, Founder of sportspredict.com
Larry Bernstein:
Welcome to What Happens Next. My name is Larry Bernstein. What Happens Next is a podcast which covers economics, politics, and culture.
Today’s topic is Predicting Sport Outcomes.
Our speaker will be my buddy Steve Kuhn an entrepreneur who started Sportspredict.com that allows the public to predict sporting events for fun to give them a predictor ranking.
Steve Kuhn:
Very delighted to be back to talk about Sportspredict.com, which is a non-gambling website where you can make predictions about various sports like the NFL, English Soccer, Indian Cricket and Formula One. You name the sport, we are creating markets that you can build a track record and prove that you are a better predictor than your friends and one of the best predictors in the world.
Since 2018 when sports gambling was legalized by the Supreme Court, we’ve seen a big rise in sports gambling in the United States. I don’t think that sports gambling should be illegal, for some people it’s probably not healthy, but other people want to predict sports but don’t want to gamble for ethical or religious reasons, it might be because they’re not old enough.
We don’t want to be Draft Kings or FanDuel. Our model is Chess.com. There are 600 million chess players in the world and over 200 million of them have an account on Chess.com.
Chess.com users pay a premium to get lessons, analysis or gain access to tournaments. And that revenue stream plus eyeballs is worth a couple billion dollars as a platform. We are trying to find ways to get people to make sports predictions to build their track record.
Larry Bernstein:
Let’s say it’s football, the Chicago Bears, what predictions can I make?
Steve Kuhn:
Anybody listening to this can go to Sportspredict.com’s NFL channel and see what predictions we’re making for the Bears game. For example, which team will win, what will the total score be over or under a certain number. But other questions would be will Caleb Williams throw for over 200 yards in this game. For each question, there’s an open market trading price from zero to a hundred, and you can either buy or sell the contract.
Will Caleb Williams throw for over 200 yards? Let’s say that it was trading at 45. You buy that contract at 45 and if it does happen, the contract settles at a hundred you would win 55. Or if he does not throw for over 200 yards, the contract would settle at zero and you’d lose 45.
Larry Bernstein:
Now, as you mentioned, I can’t make money, but I can build a track record. Tell us about how that evaluation occurs.
Steve Kuhn:
So, we’re making about 10 markets for every NFL game and there’s 16 games a weekend. So, over a weekend, you could make up to 160 different predictions. Some players like to predict every market. More selective players maybe make five predictions on a weekend. Our formula for calculating the best player is a combination of two factors. What is your average win per trade? We rank everybody, and what is your T-statistic or what is your likelihood that you’re a winning player?
Let’s put it this way for two different players, both of whom won five points per wager. They would have the same score per trade, but the one who’s made more trades would have a higher T-statistic. Therefore, they would have a higher SMART rating. SMART stands for sports markets’ accuracy rating. So, the one who’s made more predictions will have the higher SMART rating.
Larry Bernstein:
Is that a function of if we make one bet, it could be chance?
We make a hundred bets, it’s more likely than not that we have something to add.
Steve Kuhn:
That is exactly correct. So let’s say you make 10 wagers and on average you’ve made four points per trade. You’re going to have a T-statistic about 0.4 may means that you have a 60 percent chance of winning. If you do a thousand predictions, at that point, your T-statistic goes way up. You would be T-statistic close to four, and that would mean that you’re almost certainly a good player.
Larry Bernstein:
When I start making bets, do I effect the markets?
Steve Kuhn:
Yes. It’s like we are echoing the stock market that if the market is at 52, if you buy it, it moves up to 53. If you sell it, it moves down to 51. So markets are dynamic. And then during the game, we’re also moving the markets around using AI. So if you say the Bears are going to win and they started off the day trading at 52, if in the first quarter they score a touchdown, maybe that moves to 60.
You would have a live leaderboard that would show how you’re doing in the game.
Larry Bernstein:
Now there is a real-world betting sites. How can people use the predict markets to help them either in betting or in fantasy football?
Steve Kuhn:
We are scratching a different itch than FanDuel and DraftKings. If you’re motivated to make money, those are the places you want to go. It is devilishly difficult to make money on FanDuel and DraftKings because they take such a high percentage in vigorish or rake. The estimate is over a year, only about 3% of players make money on those websites. And if you are a winning player, they’ll reduce the amount of money you are allowed to wager or ban you completely. So, it’s very difficult to make a lot of money. It’s very difficult to make any money on those sites because the rake is so high.
Larry Bernstein:
What’s the bid/offer?
Steve Kuhn:
For FanDuel, standard bets the bid/offer is about 5%, but they strongly encourage parlay bets.
You make four different bets where you might have the possibility of having a very large payoff on those bets, the rate can be as high as 30%.
Larry Bernstein:
Let me explain what this means for the layman. The bid/offer for a typical bet might be 5% so in the stock market equivalent it is where you can buy a stock at 100 and sell it at 95. And for parlay bet where you need to be correct on each of 4 outcomes like the Bears win, the total points scored in the game exceed 50, the Bears score first, and the Bears quarterback throws for more than 200 yards. If you were to bet correctly on these four items, then if you bet a dollar, then the sports gambling sites will pay you back $11. Unfortunately, the theoretical fair value based on the true odds should be $14.
In every casino, the game is stacked against the gambler. But these odds are terrible in the sports betting sites. With a typical single bet whether the Bears will win with a 5% bid/offer has an expected loss of 2.5%. This implies that the bettor is expected to lose half his wager size after 28 bets.
The parlay bet with four different wagers embedded in it has much lower payoff compared to is true odds and thus with parlays bet, the gambler is expected to lose half his money after two such bets.
By comparison, playing blackjack at the casino has an expected loss of 0.50% per hand. So, sports betting on a single game is 5x worse and the parlay in the example would be 60x worse.
Steve, how many people make money on sports gambling websites?
Steve Kuhn:
Historically from all the data to be a winner on DraftKings or FanDuel, you have to be in the top 3% of people to break even or make money on gambling websites.
Larry Bernstein:
Going back to your website where there is no money gambling but instead betting for honor and prestige, what happens when you trade on your website?
Steve Kuhn:
When we open a market, we set the initial price using an AI algorithm, our algorithm makes mistakes. Now, if we were FanDuel or Draft King, that’s a big problem because we would be losing money. But in our case, that’s just part of the game. It’s who can find and act on it quickly. So, our markets are decidedly less accurate at any one point in time than the FanDuel and DraftKings.
But don’t know if that’s a bug or a feature, that’s part of our game. What that means for us is our cost to create a new market is very, very low. If FanDuel have great algorithms, great programming talent to know that they’re making perfectly accurate markets at all times. Over time, ours are getting better. I’d say right now there are ways to beat our markets, but that’ll get harder over time for two reasons. One, our AI will get better at setting the initial prices. And two, inaccurate markets are a problem if you have 30 players. But it’s less of a problem if you have 3000 or 3 million players because the correction is done by a small number and then it’s sufficient for everyone else.
Larry Bernstein:
What we’re consuming by using sportspredict.com is joy to know that we’re better at making predictions than our friends and the rest of the participants. How will Sportspredict improve your enjoyment of engaging with sports?
Steve Kuhn:
We’ve seen at Formula One events where we get everyone in the bar to play. There’s a live leaderboard, and everyone’s really excited about it. When their name gets on the leaderboard, they’re taking selfies, and when something happens in the race that affects one of our predictions, the whole bar gets excited about it. Everyone knows whether they got it right or wrong. And fans are watching the race a lot more closely and enjoying it more. I know that that’s true.
The English Premier League created something called Fantasy Premier League a decade ago where you can make predictions on their matches. And this product came out in a market that already had a lot of established gambling options. The UK has been a very active sports gambling market for decades, yet it has been very successful. With 11 million passionate players, the English Premier League has now been able to monetize the eyeballs playing that site with advertisers for over a $100 million a year. Plus, players watch more of the games because all 10 games on a weekend matter for their predictions, and they estimate that’s another $100 million in value. They estimate that this has made the English Premier League more valuable by $2 billion in enterprise value.
Larry Bernstein:
Describe the app itself.
Steve Kuhn:
The challenge always in designing any game or app is making it easy to use and making it understandable for people. And so the easiest way to test this for yourself is to go to Sportspredict.com and sign up. It’s very easy to sign up and then go into one of our markets. We have NFL markets. Golf fans, we are starting to make markets already on the Ryder Cup.
The Green Bay Packers versus Cleveland Browns game, the first question you’ll see is will Jordan Love, Quarterback of the Packers pass for 280 more yards? And currently that’s trading at 42 and you’ll see two buttons. You can either buy it at 42 or sell it at 42. We do have a video of how that works, and six months ago it was absolutely critical that people watch the video to be able to understand how to play our game. Today people log in and they know it right away because prediction markets have become mainstream.
Larry Bernstein:
Why do you think that Sportspredict.com will catch on?
Steve Kuhn:
Sports are unique in that there’s a tremendous amount of very detailed data available for free. You can go to various websites and find every play in the NBA for the last two decades You have access to LLMs
Larry Bernstein:
Large language models
Steve Kuhn:
where you can do analysis of those databases. So you could take a file of the last NFL’s last five years, play-by-play, download that into ChatGPT do analysis to find mistakes in markets. And that’s about equally true for a Stanford computer science major or a teenager in Nairobi, Kenya.
You can’t do that in finance. You’re probably not going to have the same access to data. Sports fans are so passionate, they’ve compiled this data and put it out to the world for nothing.
Larry Bernstein:
How in future releases are you going to provide information so that people can get better at their predictions?
Steve Kuhn:
I’m a huge fan of the work of Professor Tetlock and his book, Super Forecasting where he talks about people who are good predictors? I would love to bring that knowledge and set of experiences to the world more broadly.
Larry Bernstein:
For the benefit of the audience, Phil Tetlock currently a Professor of Political Science at the University of Pennsylvania and the Wharton Business School did experiments. Let me give a couple of examples, he asked people to make predictions like will Trump survive as president until his term ends in 2029? Will the Ukrainian government surrender in its war by the end of the year?
Some people were better at making predictions in international relations and politics than others. Tetlock noticed that the people who excelled at these predictions were not public intellectuals or eacademics.
And then when he put together the best predictors in teams, the predictions got even much better.
Steve Kuhn:
The most famous group of predictors from the Tetlock contest was a group called
Larry Bernstein:
Steve Kuhn:
They won almost every contest and the prize money were fairly small, but these were incredibly talented people who were obsessed with trying to prove their status as the best, which is of our point about Sportspredict. But on the back of this, they have formed a consulting group to help companies make better predictions.
The question is, can I prove through Sportspredict that I am a great predictor. I can do it in one sport. Can I do it in multiple sports? I think that would get the attention of places that need great predictors such as hedge funds, insurance companies, et cetera. Those companies should use this as a recruiting tool.
We’re broadening the number of people in the world who could compete on a fair and level playing field to demonstrate that they too are great predictors.
Larry Bernstein:
Going back to the bet whether Caleb Williams, the Chicago Bears quarterback, will pass for more than 200 yards?
You need to consider a lot of things. Who are the Bears playing this week and what is their opponent’s defensive rating? Is it going to snow and will the ball be wet? Is the star tight end injured?
You must think through all the issues and it broadens the understanding of the game, it focuses the mind and hopefully that makes the interaction with the sport even more enjoyable.
Steve Kuhn:
Yeah, I couldn’t say it better myself. And that has been the experience of the people who play our games. One thing we talk about is prize money is valuable and people like making money.
Carlsen is considered the best chess player. He’s won $10 million in chess prize money throughout his career, but he’s earned 10 times that amount through his status as the best chess player.
Another fun Magnus Carlsen story is he plays Fantasy Premier League and one year, very famously, he was for a while, number one out of 11 million players, he ended the season in the top 10 and he was ecstatic about this. Every interview with the world’s number one chess player was talking about how great he is at Fantasy Premier League.
When I was a kid, I was a chess player and playing in tournaments, not just with other kids but with adults. I loved competing with adults on a fair playing field and I was taken seriously. This is a platform for really talented kids to show that talent and to have fun with it, compete against their brothers and sisters, compete against their friends, compete against their parents. Maybe the best predictor of NFL is a 12-year-old girl in Indiana that no one’s listening to right now, and she’s screaming into the wilderness about how much she knows about the NFL and not anybody’s paying attention, but if she plays Sportspredict has a track record that’s vastly better than everyone, people will take her seriously.
Larry Bernstein:
You have become a serial entrepreneur and started a professional pickleball league. Tell us about how you founded Sportspredict.com.
Steve Kuhn:
My career is all about trying to predict things that’s what you do at hedge funds. You’re trying to understand mistakes in the market. In the spring of 2009 when subprime bonds traded down to 25 cents on the dollar, everyone talks about the big short as a way to make money, but actually the biggest money was the big long, that was a vast overreaction, and I acted on it at the time.
Later, I invested in pickleball because I fell in love with the sport and saw that this was a sport that I thought was undervalued. It had the reputation for being a sport only for old people and it had a weird name. And those two things meant that people weren’t taking it seriously. But yet everyone I saw that played it loved it with a passion.
Larry Bernstein:
And so how did you react to that?
Steve Kuhn:
I started a pickleball league. I started a pickleball rating system, again, a rating system is important. I have had a chess rating since I was six years old and I always knew what my rating was and I knew that if I studied harder and worked at it, I could probably get my rating up.
I did the same thing with pickleball. I created a rating system, it’s called DUPR, it has over a million users worldwide. You can go anywhere in the world where people play pickleball and they’ll talk about their DUPR.
If you’re a 3.5 DUPR player, you want to get lessons and you want to get better and you can measure your results. Chess.com wouldn’t have 200 million members if they didn’t have a chess rating system.
I did that in pickleball, created a rating system for that. And then I looked at sports gambling and said, “What’s something that would allow more people to do sports prediction but also give them a rating.” And I have been working on it now to make it a reality.
We are looking to raise money for sportspredict.com and we have a significant interest in it already. But if you are an accredited investor and this is something you would be interested in hearing more, feel free to reach out. I’m steve@sportspredict.com.
Larry Bernstein:
Tell us about the future of Sportspredict.com.
Steve Kuhn:
We have several different paths to make this bigger and better. First of all, we are looking to find corporate partners who we could create a prediction challenge. We’ve already partnered with one sportswriter Ethan Strauss, and we’re doing a prediction challenge with him and his users.
Then I’d like to explore this again as a way of a talent search. It’s harder to find what other objective measure can tell you about analytical talent. In the US we have the SAT, but there’s nothing that is an analytical test that crosses international borders, that is an objective test to measure prediction talent and data analytics skill.
Larry Bernstein:
Fantasy football is a big deal. Can you participate in that in any way?
Steve Kuhn:
I love fantasy football. Fantasy football requires a decent amount of effort to do reasonably well. You have to do a lot of research on the draft each week. You must make sure which players are playing. And for a lot of players they love that aspect of the game.
But what we’re doing here is easy to play. You log on a website, you see some questions and simply answer whether you think it’s more or less likely than the odds presented. The greatest games in the world are where it’s very easy to understand the rules but incredibly hard to play at the highest level.
For example, the game Go, all you’re doing is either putting a white dot on a certain square or you’re putting a black one. It’s easy conceptually to understand the rules. But playing well is incredibly complex and difficult. Sportspredict.com is an easier barrier to entry than fantasy football. It requires less knowledge base to begin.
Larry Bernstein:
I end each podcast with a note of optimism. What are you excited about as it relates to Sportspredict.com?
Steve Kuhn:
The last few weeks we’ve started to connect with people on this. There was skepticism that there would be interest in something that’s not gambling. There is a growing feeling that sports betting in the United States has gone too far. We’ve seen in the last two months, podcasts from Andrew Yang, Scott Galloway, and Tucker Carlson talking about sports gambling. So a bipartisan feeling that an alternative to sports gambling, not banning it, but giving something else that’s fun.You can play with your kids that’s teaching them valuable life skills that might be good for them in a future career at hedge funds. Larry, that might be a good way to get your kids involved in that. So, I think what has been exciting for me in the optimism is people are starting to understand that there is a use case for this, and this is scratching a different itch than sports gambling.
Larry Bernstein:
When my daughter was applying for kindergarten in New York City, an admissions officer said, “I know it’s early, your daughter’s only five, but have you thought about what career she would choose that would make you proud? I said, well, my wife and I both worked at Salomon Brothers. Obviously, we hope my daughter takes a career in fixed income, but we would also be supportive if she went into equity trading.
Steve Kuhn:
She didn’t get in, that was not the right answer.
Larry Bernstein:
It was. She just didn’t know it.
Steve Kuhn:
That’s fantastic.
Larry Bernstein:
Thanks to Steve for joining us.
If you missed the last podcast, the topic was Little Debate in College Courses.
Our speaker was Jon Shields who is a Professor of Political Science at Claremont McKenna. Jon recently finished a study of nearly 30,000 college syllabi where he found political bias in the reading materials that are assigned and hardly any conservative texts.
I would now like to make a plug for next week’s podcast. The topic will be the 10/7 Failure. Our speaker will be Yaakov Katz who is a journalist for the Jerusalem Post and the author of the new book entitled While Israel Slept: How Hamas Surprised the Most Powerful Military in the Middle East.
I want to hear why the Israelis screwed up, and what Hamas thought would be the Israeli response to the 10/7 surprise attack.
You can find our previous episodes and transcripts on our website
whathappensnextin6minutes.com. Please follow us on Apple Podcasts or Spotify. Thank you for joining us today, goodbye.
Check out our previous episode, No Debate in College Courses, here.
