
This is a guest post from Colin Davy, a talented data scientist that works at Facebook and has founded betscope.io.
Think of a bet you recently liked. Maybe it was a recommendation from someone you trust, maybe it was a system you think works, maybe it was based off The Power Rank’s analytics.
Think about why exactly you liked the bet. What about the game did you think would play out differently than the way the market implied it would? And most importantly, what particular market did you enter to act on that belief?
Chances are it was one of the standard bets (spread, moneyline, total), since those are the ones we’re most familiar with. But there are a lot of other markets available to act on your belief that the game will play out differently than the market believes it will.
Let’s take the Bills-Patriots game for tonight as an example, where the Bills are favored by 3. Let’s say you do your research, and you like the Bills to cover, so you bet on Bills -3. In a world where you are correct and the Bills do cover, all else being equal, all of the following are true:
- The Bills are more likely to cover their first half line of -2 and second half line of -1.
- The Bills are more likely to go over their team total of 21.5 points, and the Patriots are more likely to go under their team total of 19.5.
- Josh Allen is more likely to go over his passing yards prop line of 234.5 yards, and Mac Jones is more likely to go under his passing yards prop line of 200.5 yards.
- Stefon Diggs is more likely to go over his receptions prop line of 4.5, and Hunter Henry is more likely to go under his receptions prop line of 2.5.
You might have some intuition around this and might not even realize it. If you’ve played fantasy football, DFS, or any other contest where you’re constructing lineups of players, you have probably heard of strategies like stacking, where you might roster a QB/WR combo to potentially double up on big games from both players.
This taps into a powerful concept in sports betting: the idea of correlated outcomes. The same principles of fantasy apply to sports betting as well, where outcomes in one market are correlated to the outcomes in other markets.
I like to sum this concept for sports betting in a single sentence: if you believe one market is different, you have to believe every market is different.
Why is this so powerful in sports betting? Because knowing your full range of beliefs and correlated outcomes allows you to search for the best possible price across every market and book to act on your beliefs and get the best return possible.
There are price discrepancies across sportsbooks for every kind of bet imaginable, and being able to attack the market at its weakest possible points provides an actionable avenue for maximizing your long-term returns on your bets.
This sounds nice in theory, but it’s very difficult to do in practice.
- Scanning every price at every book is time consuming when done manually.
- Quantifying the correlated outcomes is difficult without some complex math.
- Matching correlated beliefs against all markets is hard to execute, especially when price discrepancies can be fleeting as books adjust their numbers.
Fortunately, we’re developing some tools that will help bettors address all of these issues. Sports betting is a growing industry with all sorts of new betting products, and with that growth comes opportunities to take advantage of these products and boost your bankroll.
The Software Tools
Hey, this is Ed again. I’ve known Colin for a long time, and I wrote about his predictive golf analytics over this past year.
Colin is a very talented data scientist, and I’ve had a chance to take a look at the software tools he is developing to put these ideas into action.
To make sure you don’t miss out on his content and software updates, sign up for his free email newsletter at Substack.