As Week 1 of the 2021 regular season approaches, here’s what to look for when betting the NFL futures markets
A few days from now, Dak Prescott and Tom Brady will square off in the opener of the 2021 NFL season. Since February, bettors were able to bet futures. Last week, I explained how I prepare for another year of betting the NFL. But I also bet some futures during the offseason, primarily win totals.
Let’s talk about how sports betting enthusiasts can best approach those markets.
To find edges in the win totals markets, I use the season simulator nflseedR, an R package created by Lee Sharpe and Sebastian Carl. Hat tip to both for helping me tune some settings in the spring. The nflseedR package simulates an NFL season – for example, a thousand times – based on win percentages for each of the 285 games that you have to provide as an input.
But how does a football bettor arrive at win likelihoods for 285 future games?
Get on our list for weekly sports business, industry trends, interviews, and more.
I don’t want to get too mathematical, but a little bit of math is necessary. After the first two weeks of free agency, I compile my initial Power Rating, which evolves throughout the off-season and regular season. The rating describes the linear relationship between all 32 teams by spitting out a spread on a neutral field.
The next step is to turn those spreads into win percentages. I let Pinnacle and BetCRIS – the two sharpest NFL sportsbooks – do the work for me.
Moneylines (the implied win probability) and spreads aren’t always perfectly correlated, but over a large sample size, several seasons, you can run a regression on those two to turn your Power Rating spreads into an implied probability.
Now, you’ve arrived at win probabilities for every single game – on what’s effectively a neutral field.
We need some home-field advantage baked in.
How much is an NFL home-field advantage worth?
The value of home-field advantage (HFA) has been declining in the NFL for several years now. In 2019, it was practically non-existent; we observed and even bigger drop during the 2020 COVID season that featured far fewer fans. There are some (albeit subjective) observations we can make as to why that is:
- We deal with small sample sizes, so a season like 2019 includes a lot of noise.
- NFL offenses are evolving; play-callers use more motion and play-action.
- Teams have gotten more aggressive on fourth downs.
These factors make life easier for quarterbacks on the road than it perhaps used to be.
When applying home-field advantage, we should work with win percentages instead of points. Not only is the phrase “three points for home-field advantage” outdated, but it’s also not the best idea to strictly add or subtract a certain amount of points for home-field.
Every spread has a different value (and push frequency) because NFL games land more frequently on specific scorelines than other team sports.
All told, the most prominent score numbers are obviously 3 and 7. Imagine one handicapper estimates a home-field advantage to be worth two points; going from a spread of -2.5 to -4.5 is an entirely different story than going from -3.5 to -5.5 specifically because games will land a lot more on scores ending in 3 and 4 than on 4 and 5.
Therefore, it wouldn’t be smart to apply two points in both cases.
Back to the simulation — no matter how much HFA you want to apply for 2021, you can use different subsets or average it out. As a result, you get mean win projections for all teams based on your Power Rating.
Compare your results to the market numbers. If you find a decent discrepancy,consider taking a crack at betting NFL win totals.
How to deal with uncertainty
It’s important to note that you’ll arrive at projections that are conditional on your Power Rating — but your ratings will change, as we are dealing with tons of uncertainty. There will be future events that we cannot control. Hence, actual win totals end up differing by an average of about 2.3 games compared to sportsbooks’ preseason numbers.
Imagine Aaron Rodgers goes down with a torn ACL in Week 1 — knock on wood — or the team is befallen by a COVD-19 outbreak. That has ramifications for all of the NFC North regarding win-loss records, and additionally affects the other 11 opponents on the Packers’ schedule.
With superior coding skills, you could model out every last one of those scenarios. But otherwise, you have to think about specific ranges of outcomes for all teams and think about fat tails. Where’s the floor, and where’s the ceiling?
Combining the numerical output and subjective knowledge, I can arrive at buy points to bet win totals. Let’s say a team is priced at 7.5 at -110 in both directions, while my number for them is 8.3.
That’s a good opportunity to bet the over.
Tackling Coach of the Year futures
Another futures bet worth considering? The NFL’s annual Coach of the Year (COY) award, which was won last year by Cleveland’s Kevin Stefanski thanks to a surprise 11-5 season.
With that in mind, you want to find candidates at reasonable prices that can realistically win double-digit games, make the playoffs, and out-perform expectations. Sean McVay (2017) and Matt Nagy (2018) are additionally relevant examples, as their teams also won at least eleven games and out-performed expectations — their Vegas win totals, for instance — by a noticeable margin.
Improvement and expectations carry weight here. If the Chiefs go 14-3, Andy Reid won’t be considered for Coach of the Year, but rookie head coach Arthur Smith will be in the mix if a team like the Falcons goes 11-6 or 12-5 and earns a playoff berth and a revamped offense.
It’s certainly in the range of outcomes that Atlanta will have a top-10 offense after seeing what Arthur Smith did in Tennessee the past two years. Earlier in the off-season, Smith was available at 20-1 or even 25-1 in some places. Those numbers are long gone — the market has caught up and banged Smith down to as low as 10-1 at some sportsbooks.
And come Week 1, expect the COY trends to reshuffle yet again once we have actual results on the field.
