Fantasy Football: 8 Risky Picks Based on numberFire's Projections

In fantasy football, the term "risk" often comes with negative connotations involving the downside and potential for bust associated with it. Risky players often have the most upside and potential for value as well, though.

Fantasy football is all about taking risks, you just have to make sure they are the calculated kind. Our projections at numberFire include confidence intervals (or ranges of outcomes), and each one has a standard deviation baked in as well. For more on this check out Fantasy Football: 8 Safe Picks Based on numberFire's Projections, but for now, just think of standard deviation as a measure of risk (both upside and downside) in the projections.

Using this, we can pinpoint some players with the widest range of outcomes and make more well-informed decisions on whether that risk is worth taking at a player's cost. This isn't just the players with the highest standard deviation at each position. Everyone knows that rookies have wide ranges of outcomes and players coming back from injuries are hard to predict. Here we'll look beyond some of those archetypes and examine players you may not have realized carried the amount of risk they do, and decide whether that should draw you in or have you fading.

Quarterback

Dak Prescott, Dallas

Standard Deviation: 31.3

In two season of NFL play, Dak Prescott went from a rising star to a detriment to his team rather quickly. In his rookie season Prescott averaged 0.29 Passing Net Expected Points (NEP; numberFire’s metric that measures what a player contributes to their offense compared to expectation) per drop back. Last season that fell all the way to a -0.01 Passing NEP per drop back mark.

It’s might seem reasonable to expect Prescott’s rushing ability to balance this out, but in the past two seasons he’s ranked no higher than eighth in total rushing attempts among quarterbacks. His rushing value has come primarily in the form of touchdowns, ranking first among quarterbacks with six in each of the past two season. But touchdowns can be fluky, and a players floor can’t be built on a high variance foundation.

Prescott is barely going inside the top-20 quarterbacks per MFL10 ADP, so the opportunity cost is low. However, there are probably better bets to be made when looking for a late-round quarterback.

Blake Bortles, Jacksonville

Standard Deviation: 28.5

Blake Bortles is one of those other late quarterbacks I was alluding to. Yes, his Passing NEP per drop back averages out to zero over his career, but that has never held him back from outproducing his draft cost. And this year looks to be no different.

He’s being taken over 20 picks after Prescott, but numberFire's projections have just one quarterback between them. Over his career, Bortles has averaged a nearly identical amount of rushing attempts per season as Prescott, but he has also thrown 82.8 more passes per season.

Bortles' fantasy projections plays out similar to Prescott’s, with increased volume bringing down his standard deviation slightly. He makes a great arbitrage play over other late-round options like Prescott.

Running Back

Ronald Jones, Tampa Bay

Standard Deviation: 40.9

Rookies naturally have a high amount of variance in their projections. With no data from them at a pro level, it’s tough to be overly confident in our projection of any first-year player. This is especially true for Ronald Jones, who has our highest standard deviation among all rookies, as well as the second-highest among all running backs.

Jones was a prolific runner in college but failed to make much of an impact as a pass-catcher, peaking at 14 receptions in his final year at USC. He was taken with the sixth pick of the second round by the Tampa Bay Buccaneers, which would imply that he will see a large workload from week one. Peyton Barber may have something to say about that though.

Currently, Barber is considered the team's starter, which would place Jones on the wrong side of a committee. Jones has the talent to make an instant impact as a runner but Barber could turn him into just another guy ready to be cut in favor of a streaming defense. Let someone else take this risk.

Rex Burkhead, New England

Standard Deviation: 31.9

Rex Burkhead has such a high standard deviation because of his ambiguous usage. In 10 games of varying health in 2017, he saw 16 red zone touches (per Pro Football Reference), 7 of which were carries at the goal line. While that may not seem like a lot, this means that 25% of his carries were inside the 20, and 11% were on the goal line. If he can stay healthy and fill that role full-time, he'll smash his sixth-round ADP.

But the Patriots spent their 22nd overall pick on Sony Michel, who profiles as a back that can be used on all downs. Michel has the profile and the draft capital of runner who dominates a backfield. The Patriots are known to spread the love in their backfield, but this pick says otherwise.

Burkhead’s outcomes range from the top-scoring back on the NFL’s best offense to second-in-line behind a first round pick. He's the perfect selection in a "Zero RB" draft.

Wide Receiver

Amari Cooper, Oakland

Standard Deviation: 33.8

In his first two years in the league, Amari Cooper recorded 155 receptions for 2,223 receiving yards. The only players to ever exceed these marks in their first two seasons are Odell Beckham, A.J. Green, Michael Thomas, and Marques Colston (based on Pro Football Focus' Query Tool).

This isn’t saying that he’s guaranteed to be as good as these players, only that what he did in his first two years set him on a career path that very few players have been on. However, his 2017 season was a disappoint of epic proportions. Cooper posted career-lows in yards, receptions, targets, and touchdowns. In addition to the weak counting stats, he was also the least efficient we’ve ever seen. His 0.07 Target NEP per target was significantly lower than his previous two season marks of 0.34 and 0.19.

Based on Cooper’s early-career excellence and the departure of Michael Crabtree, a bounce-back should be in line. Even when baking in the risk of Cooper repeating his 2017 performance, he remains a buy in all formats.

Sammy Watkins, Kansas City

Standard Deviation: 28.9

The uncertainty surrounding Watkins comes on all fronts. His highs peak at a 2015 season where he averaged 80.5 yards per game, but at his worst he was a decoy, recording 38.9 yards per game with the Rams. He has shown "elite receiver" to be within his range of outcomes but "first-round disappointment" looks just as likely through his four years in the NFL.

Now he goes to a team transitioning from a very consistent passer in Alex Smith to an unknown entity in Patrick Mahomes. Competing with two established pass-catchers in Tyreek Hill and Travis Kelce, targets may also be hard to come by for Watkins. On the other hand, an off-season connection with Mahomes and regression from Hill could lead to a rebound season, which would allow him to easily return fantasy value.

The jury is still out on Watkins' talent as a player, targets are no guarantee, and the quality of those targets is unknown as well. He carries the highest risk profile of any receiver this season. If you’ve already employed a high-risk strategy and Watkins’ name comes across the queue, it may be best to let someone else roll the dice.

Tight End

Greg Olsen, Carolina

Standard Deviation: 30.3

Greg Olsen missed the majority of last season with a foot injury, but when we last saw him in 2016, he was dynamite. That year he commanded 21.6% of the Panthers’ targets on his way to 1,047 receiving yards. The previous two seasons were more of the same, with a 20% target share and over 1,000 receiving yards in both. The Carolina Panthers have made some big changes since 2016 though.

Last year Devin Funchess stepped up by converting 111 targets into 63 receptions for 840 yards. The Panthers also let Christian McCaffrey shine as a receiver, targeting him a team-high 113 times. This year, they’ll get second round receiver Curtis Samuel back from injury, and have added a phenomenal prospect in D.J. Moore through the draft.

Our last point of reference for Olsen pegged him as a perennial stud, but this team has gone through more turnover in his absence than they ever have during Olsen’s time in Carolina. As streaming czar JJ Zacahriason noted on a recent podcast, average draft position has been unreliable at nailing down successful middle-round tight end picks.

Olsen is an easy fade in favor of tight ends going later in the draft and of going with a streaming approach at the position.

Ricky Seals-Jones, Arizona

Standard Deviation: 18.2

While he barely ranks inside numberFire’s top-20 tight ends, Ricky Seals-Jones has a standard deviation inside the top-10 at the position.

As a former wide receiver at Texas A&M, RSJ’s best assets come in the receiving game. Conveniently, the Cardinals lost 213 targets from last year between free agency departures like Jaron Brown, John Brown, Troy Niklas, and more. However, RSJ put up fewer than 300 receiving yards last year. He has shown us little in terms of actual production, but he also presents huge upside as a wide receiver being used in the tight end spot.

In week one he will take on Washington, who gave up the seventh most fantasy points to tight ends last season. If he struggles in this soft opening week matchup, you can cut ties without losing much in terms of draft capital. He also has the upside of contributing in a big way as a final-round pick though, making his upside far out-weigh his risk.