FiveThirtyEight Should Be Thankful It’s Not An Actual Sportsbook

Grant Lucas September 7, 2018 615 Reads
FiveThirtyEight

Legal sportsbooks should not just be your destination to place a wager. (Quick reiteration of legal sportsbooks. Talking to you, offshore book, among others, that doesn’t deserve to be named because you for some reason still exist and are for some reason setting lines on if Colin Kaepernick gets a signature shoe or if Donald Trump will slam Phil Knight for Nike bringing on the controversial quarterback.)

Whoa. Shake it off.

Anyway, legal sportsbooks are not just for sports betting action. They’re educational tools composed of educated information crafted by highly educated folks. Plays, players, games, teams, leagues, sports, all handicapped by savants who rarely receive their proper due.

Meanwhile, holed up in a dark room lit only by the cold light of a computer screen and serenaded by the depressing soundtrack of the clacking of keyboard keys and the scratching of pencils-on-paper, “experts” provide their take on the 2018 NFL season.

One resource combines analytics, gut feelings and knowledge; the other composed of overanalysis and algorithms. One (hint: the legal sportsbook) mans the keg tap while running the lit playlist at a house party; the other (hint: FiveThirtyEight) lives at said house but was banished to the basement after spending an hour trying to decide what genre of music paired best with the craft beer using thousands of computer simulations.

Simulations that prove nothing

FiveThirtyEight recently unveiled its 2018 NFL predictions, utilizing a model that uses what it calls “Elo ratings,” which measures head-to-head results and quality of opponents. Per the site: “This forecast is based on 100,000 simulations of the season and updates after every game.”

“Good work, everyone,” sighed the simulation supervisor, drawing in a deep breath of stale, chick-pea-puffs-infused air.

Using much less time, and enjoying the light of day as a result, bookmakers churned out win totals and odds to make the playoffs and win the Super Bowl. After the wiseguys got out their lines, FiveThirtyEight shared its computers’ simulation-derived work. For argument’s sake, we compared FiveThirtyEight’s predictions with the odds offered by FanDuel Sportsbook, which powered by longtime European bookmaker Paddy Power Betfair.

Of the 32 teams in the league, FiveThirtyEight had 23 with win totals that stood within one of those produced by FanDuel Sportsbook. Of that group, 15 were virtually exact or EXACTLY exact with FanDuel’s numbers. Fun fact, if nothing else. But what stands out with FiveThirtyEight should relieve the company that it’s not actually a sportsbook.

Value where value should not be

Last year’s FiveThirtyEight predictions are interesting to study. (In fairness, even sportsbooks get bad beats; and unforeseen injuries — like for Green Bay, Indianapolis and Houston — certainly hurt projections.)

Heading into Week 1 of the 2017 season, FiveThirtyEight gave the Philadelphia Eagles a 2 percent chance of winning the Super Bowl, one of the lowest in the league. They were forecast to finish third in the NFC East with an 8-8 record. Of course, Philly went on to go 13-3, win the division and the franchise’s first Super Bowl.

Other surprises cropped up: Jacksonville winning the AFC South after being projected to take last, and reaching the AFC Championship game; the Los Angeles Rams winning the NFC West after having the fourth-lowest percentage to reach the playoffs in the league. Sportsbooks also succumbed to such surprises. But the 2018 predictions from FiveThirtyEight are … surprising, to be nice about it.

Obviously, the company (sorry, its simulations) have the usual suspects atop the Super Bowl favorites: New England, Philadelphia, Minnesota and Pittsburgh. Same with FanDuel Sportsbook.

Let’s say FiveThirtyEight was a sportsbook. Where’s the value?

Oh, I’ll feed you, baby birds.

Start with the teams that are overrated (compared with FanDuel Sportsbook’s odds). Kansas City, per FiveThirtyEight will win the AFC West with 9.5 wins (8 wins by FanDuel). The Chiefs are the sixth-biggest favorite (12th by FanDuel) to reach the playoffs and the seventh-biggest favorite (16th by FanDuel) to win the Lombardi Trophy.

Never thought I’d say this, but Alex Smith is no longer there for the Chiefs. Their playmakers (Kareem Hunt, Tyreek Hill) are now on the radars of opponents. And KC’s defense is fragile. With the Chargers on the rise and the streakiness of the Raiders, the Chiefs have a difficult road ahead. Under on wins; yeah, OK, on playoffs; end of season.

FiveThirtyEight’s algorithm (dang robots) also overvalues Buffalo, which inexplicably reached the playoffs last year but not by their own doing. The company/fictional sportsbook has the Bills winning 8 games (FanDuel more accurately with 6.5) and with reasonable playoff chances at 36 percent. FanDuel gives Buffalo the longest odds at the postseason, which, without Tyrod Taylor and no other weapons seems more reasonable.

The greatest discrepancies between FiveThirtyEight and FanDuel Sportsbook come with Green Bay, Houston, and San Francisco.

Consider that in only two seasons the Packers have not made the playoffs with Aaron Rodgers as a starter: 2008 (his first year as the play-caller) and last year (when he was injured early in the season). Yet FiveThirtyEight projects Green Bay to take third in the NFC North with 7.1 wins (10 by FanDuel), a paltry 23 percent chance at the playoffs (FanDuel with sixth-best odds), and a mere 1 percent crack at the title (FD against with sixth-best odds).

A similar story is told with Houston, which was 3-4 and in the division title hunt before starter Deshaun Watson went down. FiveThirtyEight gives the Texans 6.3 wins (8.5 by FanDuel) to finish last in the AFC South and the fourth-lowest chance at reaching the postseason (ninth-best by FanDuel). San Francisco is also predicted to take last in its division with 7.3 wins (8.5 by FanDuel) with a slightly better playoff percentage than Green Bay but still in the lower half of the league (FanDuel with upper-half odds). This FiveThirtyEight forecast is for a team that features a QB, Jimmy Garoppolo, that went 5-0 as a 49ers starter last year and added offensive and defensive weapons in the offseason.

Certainly pocketing the potential to tell everyone, “I TOLD YOU SO! I CALLED IT!” while parading around the room like Rocky Balboa is the American dream.

But being able to take on the “experts” (FiveThirtyEight allows readers to do week-by-week picks to beat its projections) and scream, “I TOLD YOU SO!” is even sweeter.