Making Point Spreads by Oddsmakers and Using Algorithms

Betting on sports is becoming quite the hobby, as more and more states legalize across America. Besides, watching games is way more fun when you’ve got a little skin in the game. As you'd imagine, the NFL is king, drawing the vast majority of the bets placed in the United States every year.

The point spread. The “great equalizer.” Just how do sportsbooks  and oddsmakers come up with this all-important number? Most of the time, oddsmakers try to come up with a point spread betting infrastructure that creates the least risk for the sportsbook and effectively guarantees a profit even when your bet wins.

Many people assume that the point spread is designed so that both outcomes in a game—(1) the favorite covering and (2) the underdog covering—are equally likely given the relative strength of the teams. That’s close, but not quite right. And at times, impossible. The majority of the time, sportsbooks try to create a point spread with a margin of victory that attracts equal betting on both teams. Sometimes this is equivalent to a results-oriented spread—one which creates a truly level playing field—but not always because the public isn’t necessarily a rational actor.

The public may counter with their own betting lines by developing their own numbers. At their core, a sports betting algorithm functions by analyzing past and real-time data to predict the likelihood of different outcomes in a game. This prediction is based on a variety of factors, including team performance, individual player statistics, weather conditions, and even odds offered by bookmakers. These algorithms use statistical models to analyze these factors and create probability estimates for the outcome of a game. These are the public’s own weapons. For example, a sports betting algorithm might assess the probability of a specific team winning based on previous performance and other variables. Bettors can then use this information to compare with odds from different sportsbooks and decide whether to place a bet. This process is often referred to as identifying value bets—where the odds provided by a bookmaker are higher than the algorithm's probability of a certain outcome. This is where sharp action comes into play. 

Machine learning algorithms are now commonly used in sports betting. These algorithms improve over time by learning from past data and refining their predictions based on new information. By processing large datasets, machine learning algorithms and models can identify patterns that humans might miss, which can lead to more accurate predictions in complex scenarios like live betting.

Why do sportsbooks want equal betting on both teams most of the time? Simple. It’s a risk-free way for them to make money.If there is equal money on both sides, there is no risk for the sportsbook. They don’t turn a huge profit on any single game but they do it from a much bigger tray and they do it a couple of million times. Over time, it adds up a lot.

As a collective, the general betting public is not always a rational actor. People will bet on certain teams more heavily than others, even when that team doesn’t deserve it based on its quality of play. Teams that attract an irrational amount of betting money are known as “public teams.” You will find them in every sport. They tend to be the high-profile teams with massive fan bases and (sometimes but definitely not always) history of success.

There is no universally accepted grouping of public teams, but you won’t find much argument that the teams listed in the table below qualify.

Here’s a list of the most recognized public teams in America. 

Dallas Cowboys

Green Bay Packers

New England Patriots

Alabama Crimson Tide

Notre Dame Fighting Irish

Golden State Warriors

Los Angeles Lakers

Duke Blue Devils

Kentucky Wildcats

UNC Tar Heels

Chicago Cubs

Los Angeles Dodgers

New York Mets

New York Yankees

Toronto Maple Leafs

When oddsmakers list public teams at so-called accurate betting odds, the public is likely to bet on them more heavily and potentially put the sportsbook at risk of a loss if the public team covers the spread. For that reason, teams like the Dallas Cowboys and Duke Blue Devils and others mentioned above will often be bigger favorites than the analytics indicate they should be.

Many sportsbooks are willing to “take a position” in certain circumstances, which basically means that they will set/leave the point spread so that the sportsbook needs a certain result in order to come out in the black. When you see one sportsbook open a point spread at a different number than its competitors, those oddsmakers are likely taking a position on that game. They know that the lower/higher spread will attract bets on the team that is laying fewer/getting more points, but they are okay with that because their analysis says that the other outcome is more likely. Bookmakers don’t always need to be correct when they take a position; they just need to be right often enough to be profitable on the whole. Chances are that the sportsbook you’re using has done just that. It’s still in business after all.

Monte Carlo simulations are used to model the probabilities of different outcomes by running simulations thousands of times. This technique is especially useful in sports betting as it accounts for the inherent uncertainty and variability in sporting events. Many companies and individual bettors have become interested and have started using artificial intelligence-based algorithms to predict game outcomes. For example, several popular sportsbooks use algorithms to adjust odds in real-time based on ongoing match data. Bettors can also use platforms that provide algorithm-driven predictions to help them make betting decisions.

It’s also important to recognize that while algorithms can identify value bets, they are not a guarantee of profit. Sports betting algorithms offer valuable insights for bettors looking to make more informed decisions. However, they are not foolproof and should be used responsibly. By understanding the inner workings of these algorithms, how they are developed, and their limitations, bettors can enhance their experience without falling into the trap of over-reliance on technology. No matter how sophisticated the algorithm, gambling always involves risk. Use algorithms as tools for responsible gambling, and never bet more than you can afford to lose.