Almost two months ago I started playing around with models on Pine Sports. It's a pretty cool site, and I had one model that backtested particularly well. It's an NBA moneyline model. As I started tracking results from the model, I thought it would be interesting to see how the spread performed in games the model identified as good spots for moneyline bets.
The sample is up to 142 games. Still too small to draw significant conclusions, but large enough to be interesting.
Basically, the question this analysis can give insight on is, are the points worth it to take the spread over the moneyline in an NBA game?
Here is how the results from this model break down so far.
Moneyline Bets: 64-95, +13.5 units
Spread Bets: 74-70, -0.41 units (I didn't start tracking spread bets for about a week, that's why there are more ML bets than spread bets.)
This alone indicates there is more value betting the moneyline, but there's another angle I want to explore. The logical reason for playing the spread would be that there are enough games where a team loses but keeps it close that a spread bet wins even when the ML bet loses. How often has this happened?
There have been 95 games where the ML bet lost.
In those games, if you bet the spread on that same team, you would be 26-69.
Of course, winning 26 bets instead of losing them seems nice, but the price makes it not worth it. 72.63% of the time, when a team loses a game they don't cover the spread. Again, this is only for the specific model I'm using with 142 game sample size. I'm not saying this will hold forever, but it is worth watching and I'll continue tracking. It does make me wonder if the spread is worth it in the NBA.
Comments