Russell Wilson has been the ultimate late-season fantasy QB

(“Today's Crazy Fantasy Stat” is an occasional offseason offering from PFF that highlights something that catches our eye and aids in our preparation for the 2017 fantasy season.)

Fantasy football is, of course, a season-long game, and generally speaking, the team that accumulates the most total points over the course of the season is going to be in good shape. (Obviously luck in the form of scheduling plays a big part, but the general truth is there.) The problem with that, though, is that it leads to us looking at season point totals as a barometer of success instead of finding the best performers in each given week.

(Subscribe to all our fantasy content, or get everything PFF offers with an All-Access subscription.)

My favorite example of this is new Packers tight end Martellus Bennett. He has developed a reputation as a fringy starting tight end in recent seasons based on his yearly totals. But in splits, his numbers look far worse:

In his first five games of the season, Bennett has been a superstar. After that, he’s been a relative dud. Is that predictive? No, not especially, but it’s worth acknowledging, when looking back on a given season and seeing that Bennett was the No. 7 tight end, that he wasn’t really.

To that end, today I want to discuss quarterbacks. Specifically, first-half QBs vs. second-half. I compiled every quarterback who has played at least half of his team’s games the last three years (plus Marcus Mariota, Jameis Winston and Tyrod Taylor, who all first became starters in 2015) and looked at their average splits by quarter- and half-season.

Three quarterbacks have averaged first-half and second-half totals at least 4.5 fantasy points apart over the last three seasons:

Over the last three years, Tom Brady and Philip Rivers have seen their second-half performances fall off in a big way. Perhaps not coincidentally, both are on the older side of NFL quarterbacks (though admittedly, Brady just had maybe his best season yet). Rivers has had to deal with serious attrition among his offensive weapons, with Keenan Allen, Antonio Gates, Danny Woodhead and others going down midseason in recent years, likely costing him some production.

Brady, meanwhile, has a different baseline than Rivers — his worse half was nearly as good as Rivers’ better half — and much of his higher first-half average is buoyed by his insane first half of 2015. In other words, yes, he has gotten worse over the second halves of seasons, but his “worse” was still plenty productive.

Russell Wilson, though, is the interesting one. The quarterbacks in this sample generally get worse as the season goes on (non-Wilson QBs in the sample averaged 17.2 ppg in the first half, 16.3 in the second) — which makes sense, as an NFL season is a wear-you-down grind — but Wilson has historically done the opposite. There have been noticeable reasons, particularly a first-half injury in 2016 and a crazy Doug Baldwin stretch in 2015, but regardless, Wilson’s second halves have jumped by 4.5 fantasy points per game the last three years; the next-best jump in the sample was Cam Newton, whose second halves have only been 1.7 ppg ahead of his first halves. Wilson and Aaron Rodgers (20.8 ppg) are the only quarterbacks in the sample to average 20-plus fantasy points a game over games 9-16.

Who knows if this is really who Wilson is, or if it’s just statistical noise. By the time a sample size is large enough to really know for sure if a player is a first-half/second-half player, he will have probably aged out of that information having any sort of usefulness. So I can’t confidently say Russell Wilson will have a better last eight games in 2017 than he will a first eight.

But if it’s Week 8 or Week 9, and Russell Wilson is struggling, and someone offers me a trade of him for Philip Rivers? I’m probably taking it.