Who’s better: Brady or Manning? Who deserved to be MVP in 2017? More and more, these types of questions are being answered with advanced stats like Wins Above Replacement (WAR).
Like many advanced stats, WAR began in baseball and is slowly leaking into the other major sports. Before we dive into what WAR can do for you as a sports bettor, let’s take a moment to understand how it’s calculated and some of its limitations.
For those of you interested, the WAR model I use is heavily based on a paper written by three Carnegie Mellon Data Science graduate students which can be found here (I’ll wait while you read it…).
For those of you who didn’t make it through the 40-plus page paper, WAR consists of play-by-play data and a couple computer models. One model calculates a value (generally either Expected Points Added – EPA – or Win Probability Added – WPA) for each play and the other splits that value among the players involved (QB/WR/RB/TE). Add it up over the course of a season and compare each player’s value to a set of “replacement” players.
The best way to think about replacement players is those who can be cheaply/freely obtained on the market – guys who might be on the roster bubble or could be obtained for a low draft pick.
FIND THE LATEST NFL WEEK 1 BETTING LINES ON OUR LIVE ODDS PAGE
Let’s get a feel for typical values of WAR for QBs. Looking at 2017, Russel Wilson comes in as the No. 1 QB with 6.25 WAR. That is, if the other Seahawks play the same but Chad Henne/Mike Glennon/Derek Anderson is the QB, you’d expect the 2017 Seahawks to finish around 3-13.
Towards the lower end starters, we have Trevor Siemian, Deshon Kizer and Brett Hundley all as worth about 1 WAR, and 2017 Joe Flacco worth slightly less than replacement – the only starter with that distinction.
Now on to some specific situations and a look at how WAR can help you as a sports bettor.
Nick Foles vs. Carson Wentz
Eagles fans get more Nick Foles and we’ve seen the Vegas line move four points in response to the news – going from Eagles -5.5 to -1.5. That’s a bigger move than many expected, but is it too much? Not enough?
Before his injury, Wentz was on pace to be the No. 1 QB by WA, and still came in third, worth 5.9 wins to the Eagles. Nick Foles came in and was decent, logging 0.3 WAR in his short stint as the starter. Over his career, per 500 plays Nick Foles has been worth 2.45 WAR, a respectable number that would put him as a top backup or low-end starter.
Carson Wentz over his career averages 3.6 WAR/500 and was worth 5.7 WAR/500 last season. That leaves Wentz as worth somewhere between 1-3 wins more than Nick Foles across a full season’s work. Translating that to expected points per game, we get Wentz being worth somewhere between 2.2-6.7 points per game. If you believe Wentz is who he was last year and are willing to discount his rookie year (put me in this camp) then the line move isn’t enough and the Falcons become an attractive bet Week 1.
Ryan Fitzpatrick vs. Jameis Winston
The Bucs will be playing without their starting QB versus the New Orleans Saints in Week 1 as he sits out a suspension, and the line has moved 2.5 points since the announcement, with Tampa Bay opening as a touchdown underdog on the road and sliding to -9.5 with Winston out.
Winston has been worth 3.37 WAR/500 in his career, versus Ryan Fitzpatrick’s 2.56 WAR/500 career rate, with a similar number for Fitzpatrick over the last three years. That equates to a difference of about 1.75 points, suggesting the market is slightly overreacting to Winston’s suspension.
As for Brady vs. Manning, unfortunately play-by-play data only goes back to 2009, so I can’t give an answer on that. But what about most valuable since 2009: Brady? Rodgers? Nope. Drew Brees tops the list. Certainly helps to have missed just two games over the last nine seasons.
