To find value in a particular NHL player prop, a bettor must first predict what is likely to happen. Then, with a prediction in hand, it’s time to assign probabilities to both sides of the prop. Finally, a bettor must quantify his edge so he can decide whether to place a bet. For instance, “Will Auston Matthews Score A Goal?” is a prop many sportsbooks offered Monday evening before the Toronto Maple Leafs faced the Ottawa Senators. Here’s how FanDuel priced the prop:
It might have been a good bet, but a bettor must first analyze data from previous games and use his best judgment to estimate the expected number of goals Matthews will score. Otherwise, there’s no way to know how often the event will occur. Computationally, it’s a complex problem. Therefore, most bettors don’t have the slightest clue how oddsmakers price a market like this one, let alone how they should price it themselves.
%%offer%%
Fortunately, it’s a rather simple problem to solve with the help of an application like Microsoft Excel. However, we’ll do everything in Google Sheets since it’s free to use. The only thing a bettor must provide is a prediction. To come up with a prediction for whether Matthews would score Monday against the Senators, I used a method called bootstrap sampling, which helps us estimate the underlying distribution of a data set even when we have a limited sample size. Bettors can learn more about how I came up with my prediction in the book “Statistical Sports Models In Excel.”
Matthews had scored 11 goals in 14 games going into Monday’s game, and using the aforementioned method, a projection of 0.8 goals was generated. Now that we have an estimate, we just have to use the Poisson distribution function to find out, based on our projection, how likely it is that Matthews will score in a given game. Open either application and select the cell where you will place your prediction. In this example, it’s placed in C-3. The formula we are using is over in F-4. Drag it down a total of six rows and stop at F-10.
If everything has been input correctly, the spreadsheet should look something like this (See Point Spread Weekly chart)
Each probability corresponds with the number of goals in the cell adjacent to it. For example, the probability that Matthews will score exactly one goal is 35.9%. Cell C-6 contains the probability that Matthews does not score, which as you can see is 44.9%. Once at this point in the process, simply subtract that number from 100 and we’re left with a 55.1% chance that Matthews scores. In other words, we have estimated that “Auston Matthews Over 0.5 Goals” should be priced at -123 (55.1%). After removing the vigorish, we can see that the bookmaker is implying that Matthews’ chances of scoring a goal roughly resemble a coin flip, since the implied probability of the odds being offered (-110) is 52.4%. Therefore, betting on the Leafs forward to score at -110 should be considered a value bet since the odds of it happening are greater than the implied probability of the odds offered by FanDuel.
Matthews did indeed score — in fact, he scored twice — but it’s beside the point. He was chosen because he is the current league leader in goals and likely a popular bet to score on any given night, especially when the Maple Leafs face a lowly team like the Senators. This made Matthews a great test subject. This was simply a practice example to get bettors off and running in the right direction. Understand that using this process to predict goal scoring will produce positive results only if the bettor’s estimate is more accurate than that of the bookmaker.
