Project Introduction

99 billion dollars (www.businesswire.com). That is the amount of money wagered on sports in the past year (2022). That is a massive number and it’s expected to keep growing at an annual rate of 9.8%. The past few years have seen tremendous growth in the sports betting industry. This is largely due to sports betting on pro and college sports being made legal at the federal level in 2019. However, this had to be put into effect on a state by state basis. Currently there are 36 states that have legalized sports betting. Although the specifics differ by state, like Washington where online sports betting is illegal, but in person sports betting at tribal casinos is legal (www.forbes.com).

The passing of this law prompted easy access to sports betting at a level never seen before. In person sportsbooks were being created in states other than Nevada for the first time. What really changed the industry was the creation of online sports betting platforms like DraftKings and Fanduel. This allowed anyone to sports bet from their phone or other device without ever having to leave their house. This had to approximately 1 in 5 Americans betting on sports in the past 12 months, with approximately 50% of Americans saying they’re interested in sports betting (www.pewresearch.org). With this growth in popularity many new sports bettors or people interested in sports betting likely have the question: Is it easy to be a profitable sports bettor? The simple answer is no. Assuming bets are only placed on totals and spreads with the typical -110 odds, a win rate of 52.4% is needed to be profitable. This doesn’t sound too difficult. However, sports books are very good at setting their lines to make it difficult on the bettor. To put it into perspective professional sports bettors will consider a 55% win rate a good year (www.actionnetwork.com). Is it possible to gain an edge on the sportsbooks? That’s what this project aims to do. It looks at if a machine learning model can be created that can predict a total bet or spread bet with at least 52.4% accuracy.

Definition of Terms

To understand sports betting and in turn how this model works a few terms need to be defined. The two most common types of bets are total line bets and spread line bets. In total bets the bettor has to choose whether the total points in a game will go over or under the total line (points) set by the sports books. When the total points in a game is higher than the total line set before the game, it’s referred to as the over. When the total points in a game is lower than the total line set before the game, it’s referred to as the under. In general the odds will be -110 when selecting either the over or under. In spread bets the bettor has to choose whether one team will win by a certain number of points or the other team won’t lose by a certain number of points. The two terms used here are the favorite and the underdog. The favorite has to win by certain number of points, while the underdog has to not lose by a certain number of points. For example, let’s consider an NFL game between the Kansas City Chiefs and the Philadelphia Eagles. The spread line is -2.5 Eagles and +2.5 Chiefs. The favorite will always have a negative sign in front. So, in this case the Eagles are the favorite and must win by 2.5 points for the bet to push and more than 2.5 points for the bet to win. Any other result is a losing bet. The underdog will always have positive sign in front. So, in this case the Chiefs are the underdog and must not lose by 2.5 for a push and less than 2.5 for the bet to win. Any other result is a losing bet. The team that wins the bet is considered to ‘cover’ the spread. So, if the score ended Eagles 27, Chiefs 26, the Chiefs would have covered the spread and a bet on Chiefs +2.5 would win. The final term is ‘odds’. The odds signify how much a sports bettor needs to bet for every $100 dollars they want to win. Negative odds imply that a bettor needs to bet more than $100 dollars to win $100. For example, the typical -110 odds means that a bettor has to bet $110 to win $100. This is why sports bettors need to win 52.4% of their bets and not 50.1% to be profitable. Some bets will have positive odds. This means the expectancy of a winning bet is less likely. However, it will be more profitable if it wins. For example, if a bet has +100 odds a bettor has to bet $100 to win $100 (www.investopedia.com).

Below is an example of the interface for a basketball game between the New York Knicks and Washington Wizards on the DraftKings sportsbook. The ‘Game Lines’ section includes the spread, total, and moneyline bets (moneyline is choosing the winner). The odds for the total and spread are all at the typical -110. The New York Knicks have a spread of -2, which means they have to win by more than 2 for the bet to win or by 2 for the bet to push. Any other result is a loss. The Washington Wizards have a spread of +2, which means they have to lose by less than 2 or win for the bet to win or lose by 2 for the bet to push. Any other result is a loss. The total line is set at 225. This means that for a bet placed on the over to win, the teams have to combine to score more than 225 points. For a bet placed on the under to win, the teams have to combine for less than 225 points. If exactly 225 points are scored a push occurs. The moneyline odds are set at -130 for the New York Knicks and +110 for the Washington Wizards. This means that the New York Knicks are favored to win.

Calculating Decimal Odds

Decimal Odds for Negative Sports Odds: 1 – (100 / Negative Sports Odds)

Decimal Odds for Positive Sports Odds: 1 + (Positive Sports Odds / 100)

Example for Sports Odds of -110: 1 – (100 / -110) = ~1.91

Example for Sports Odds of +150: 1 + (150 / 100) = 2.5

Implied Probability vs Real Probability

In sports betting the to distinguish implied probability and real probability. Implied probability is the likelihood of an outcome based on the odds set by the sportsbooks. This value is calculated by just doing 1 divided by the decimal odds.

Implied Probability for Odds of -110: (1 /1.91) * 100 = 52.36%

Implied Probability for Odds of +150: (1 /2.5) * 100 = 40.00%

So, if a team has odds of +150 to win the game then their implied probability is 0.4. This indicates that the team has an implied probability of winning 40% of the time. Sportsbooks will often set their odds, so that the implied probability is actually slightly higher than the real probability of a team winning. Since, Sportsbooks do this, bettors need a metric to determine the quality of the bet they’re making. This is done by using expected value.

Here is a useful calculator for finding odds and win amounts.

Expected Value

Expected value is on average the amount of money a sports bettor can expect to win on a wager. Therefore positive expected values indicate a higher quality bet as their is more value. Sports bettors should refrain from placing bets with negative expected values most of the time. Expected value is best used when the sports bettor has a different predicted probability (real probability) than the implied probability from the odds given by the Sportsbooks. Expected value can be calculated using the following equation.

Expected Value = (Probability of Winning * Amount Won Per Bet) – (Probability of Losing * Amount Bet)

For example, the Seattle Seahawks play the Arizona Cardinals. The Seattle Seahawks have -130 odds to win and the Arizona Cardinals have +110 odds to win. This indicates that the Seahawks have an implied probability of winning of 56.5% and the Cardinals have an implied probability of winning of 50%. Now, if a sports bettor believes that the Seahawks will win, but the probability is less, like 52%, expected value can be calculated to see if betting on the Seahawks to win at -130 is a quality bet.

Amount Bet: $25

Amount Won if Bet Wins: $19.23

Amount Lost if Bet Loses: $25

Expected Value = (0.52 * 19.23) – (0.48 * 25) = -2

So, the expected value will be negative for a $25 bet when the sports bettor believes the odds of winning are at 52%. This indicates the bet should not be made because it doesn’t provide value.

In general, expected value is saying that when a sports bettor believes the probability of a favored side winning is less than the implied probability given by the Sportsbook, the bet should not be made and vice versa. Also, if the sports bettor believes the probability of the underdog side winning is greater than the implied probability given by the Sportsbook, the bet should be made and vice versa.

So, using expected value, a sports bettor can now make a bet that takes into account the likelihood of their bet winning against the probability given by the Sportsbook

How Sportsbooks Set Their Odds

Why do sports books set the odds on the typical spread or total bet to -110? It all comes down to money and risk. Sports books really aren’t trying to profit too much off spread or total bet. These are the most common types of bets, so in turn the most amount of money is going to be wagered. This creates risk for the sports books if a large percent of the money ends up on one side of the total or spread. To minimize that risk, sports books aim to set lines that will result in the money being evenly split on both sides. This is also why lines are adjusted during the week. If one side of the total or spread has significantly more money bet on it, the total or spread will move, making the other side of the bet more enticing, which the sports books hope will even the money out. The odds are set at -110 because this further reduces the risk of the sportsbook as they only have to win 47.6% of the bets placed. As an industry, sports books agree on this way of setting odds, which means the consumer has to accept the fact that they are betting at a slight disadvantage. While, setting the lines this way does reduce risk it also reduces the possible profit. Sports books are able to do this because of how much they profit off of parlays and other non-typical bets. A parlay is when multiple straight bets are combined into one large bet. For the bettor to win, all of the straight bets must win. Parlays are popular because they give an odds boost. So, to the bettor it appears that they are gaining an advantage by combining their bets into a parlay. However, the odds of all the straight bets winning significantly go down as each straight bet is added. So, while parlays appear like a great idea to bettors, the sports book actually has a significant advantage and sees most of their profit come from those type of bets.

Project Challenges

This project will not be easy or guaranteed to be successful. Some quick background research and general knowledge of the sports betting field, reveals that creating a model that will correctly predict the total result more than 52.4% of the time will be challenging. There are a few reasons for this. First, finding useful NFL data that also contains sports betting data or can be combined with sports betting data is not an easy task. Second, there are just not that many strong predictors of the total result, and the ones that are strong tend to uncommon. Third, sports books have been perfecting their models for setting total lines for a long time. As a result, they are very effective and difficult to beat. While there are many challenges to this project, they create plenty of opportunities to practice machine learning and there is the exciting possibility of creating a profitable model.