Photo: Imagn Images

In last week's column, I introduced True Fantasy Points (TFP) as a method for quantifying regression to the mean (RTM), and then applied it in the context of identifying buy-low/sell-high candidates at QB. We can do the same thing for any fantasy position, and not much has changed on the QB front, so today I'll look at WRs.

Calculating TFP for WRs requires us to answer five questions (Note: I'm only considering points from receiving because WR rushing performance is a crap shoot.):

1. What unit of observation should we use? Targets or something else, perhaps?
2. How many observations before reception rate stabilizes as a measure of "true" reception ability?
3. How many observations before yardage rate stabilizes as a measure of "true" yardage ability?
4. How many observations before touchdown rate stabilizes as a measure of "true" touchdown ability?
5. What are the league averages for reception rate, yardage rate, and touchdown rate in 2014 only?

With fellow FootballGuy Chase Stuart and I independently finding similar results in our research, I've become increasingly convinced that WRs should be judged on a per-route basis rather than a per-target basis. That is, if you want to know about a WR's "true" fantasy scoring ability, you should look at his receptions, yardage, and touchdowns per route run rather than his receptions, yardage, and touchdowns per target. The only potential snag in following this advise is that, while target data is widely available (e.g., here's the link to FBG's target stats through Week 5), you can only get routes-run data via a premium subscription to Pro Football Focus. The good news is that, although copyright prevents me from giving you the routes-run data, nothing prevents me from using it to calculate TFP and report that to you instead.

So now that we've decided to use routes run as our unit of observation, here are the answers to the other questions:

• Receptions per route run (RPRR) stabilizes after a WR has run 188 routes.
• Yards per route run (YPRR) stabilizes after 351 routes run.
• Touchdowns per route run (TDPRR) stabilizes after 882 routes run.
• Considering only the routes WRs have run in 2014, the current league averages .121 RPRR, 1.54 YPRR, and .009 TDPRR.
• Considering all the routes WRs have run for his current team going back to 2007, the current league averages are .129, 1.79, and .011.

From there, it's a matter of plugging those numbers into a formula you can find elsewhere, and then subtracting our TFP results from actual fantasy point totals through Week 5. This produces a +/- stat that indicates the extent to which a WR is scoring above (or below) what we would expect after accounting for RTM.

TFP through week 6: wide receivers

Below is a table showing TFPs and +/- stats for players among FBG's Top 72 WRs Forward either by standard FBG scoring or by PPR scoring. For your convenience, I've sorted it by +/- in PPR leagues:

As an example of how to read the table, take a look at the row associated with Jordy Nelson. In standard scoring leagues, his routes run should have produced 44.7 points if we apply RTM based only on 2014, while those same routes run should have produced 48.2 points if we apply RTM based on his entire Packers career. Compared to his actual standard point total through Week 5 (76.5 points, not shown), Nelson's outscored his 2014 expectation by 31.8 points and outscored his Packers career expectation by 28.3 points. Nelson's TFPs and +/- stats for a PPR scoring system are interpreted in much the same way.

Before I use the numbers to detail a handful of buy-low/sell-high candidates in both types of scoring systems, I'd like to briefly call attention to a couple of WRs that illustrate the limitations of TFP that I mentioned last week.

DeAndre Hopkins looks like a sell-high candidate according to the table, but he's a perfect example of two TFP blind spots. First, he's a second-year player at a position where rookies tend to fare mediocre at best, and his rookie year was no exception. Therefore, it's plausible that his TFPs (and +/- stats) are weighting Hopkins' 2013 performance too heavily; an idea that's further supported by how much lower these values are for 2014 only. In short, this season may very well be his "new normal," and TFP hasn't picked up on that yet because about 80 percent of the routes he's run for the Texans came last season. Second, TFP also doesn't know that Houston made wholesale strategic changes this year, going from a Gary Kubiak offense that perennially produces only one viable fantasy WR -- and apparently continues to do so in Baltimore -- to a Bill O'Brien offense that spreads the ball around. Again, it's plausible that 2014 is a better indicator of Hopkins' "true" fantasy scoring ability than 2013 and 2014 combined.

On the other end of the ledger, Robert Woods is a perfect example of the other three TFP blind spots. One one hand, it doesn't know that (a) the Bills have three games remaining against top 8 defenses in WR points allowed, as opposed to only one against the bottom 8; and (b) Buffalo's offense is one of the least pass-happy in the NFL. On the other hand, however, TFP also doesn't know that Woods has been dealing with an ankle injury for the past few weeks, nor does it know that the Bills just benched an awful novice at QB in favor of a mediocre veteran.

standard buy-low candidate #1: brandon marshall

Seriously, guys and gals, go trade for Marshall right now. The remaining schedule for Chicago WRs may not be all that favorable, but Marshall is wildly underperforming his "true" fantasy-scoring ability. To boot, there's both a statistical and non-statistical explanation for it. With respect to the latter, there's obviously the injury issue. With respect to the former, get this: Marshall's True YPRR with the Bears is 2.13, but he's at 0.95 so far this season! Furthermore, other than a highly uncharacteristic strategy against the Packers in Week 4, Mark Trestman and company have called for a pass at least 65 percent of the time, regardless of whether or not the Bears were ahead or behind on the scoreboard.

STANDARD BUY-LOW CANDIDATE #2: CORDARELLE PATTERSON

Patterson's +/- rankings are the product of underperforming his True YPRR in Minnesota by 0.41 and underperforming his True TDPRR by .011. Based on the number of routes he's run this season, that translates to 65 fewer receiving yards and two fewer TDs than he's actually amassed so far. Among the per-route metrics I use to calculate TFP, TDPRR is by far the most susceptible to randomness, and Patterson's somehow yet to score a TD despite running over 150 routes in 263 offensive snaps. It also helps that Vikings WRs have a favorable schedule going forward, and Minnesota's passing game as a whole figures to benefit from Teddy Bridgewater's return to action.

standard sell-high candidate #1: brian quick

Admittedly, it's entirely possible that Quick is squarely in the "young breakout player" blind spot of TFP. That said, almost everything else points to him trending downward over the course of the rest of 2014. Through Week 5, he's averaging 2.48 YPRR and .023 TDPRR, but his "true" ability based on over 300 routes run with the Rams is 1.72 YPRR and .013 TDPRR. On a percentage basis, that means he's overperforming his True YPRR by 44 percent and overperforming his TDPRR by 77 percent. To boot, TFP14 says just about the same thing as TFP: His yardage total is 38 percent higher than it should be and his TD total is 109 percent higher than it should be.

Meanwhile, the other TFP blind spots don't work in Quick's favor. The remaining schedule for Rams WRs is about as "meh" as it gets, and he plays in Brian Schottenheimer's run-happy offense.

(As was the case last week with Andrew Luck, full disclosure: I have Quick in the FBG staff league, and I'm contemplating selling high on him. Again, I wouldn't advise you to do something I wouldn't consider doing myself.)

standard sell-high candidate #2: a.j. green

Green is Cincinnati's undisputed No. 1 WR, and he's got a very favorable schedule going forward. I'm only including him here because I have to publish a crazy stat. Among "established" WRs (for lack of a better term), Green is No. 1 by far in terms of how much he's overperforming his True YPRR. Over the course of nearly 1,800 routes run since being drafted by the Bengals, Green's averaged 2.31 YPRR; he's currently at 3.53 YPRR.

To put that in perspective, consider the following. Due to an early-game toe injury, Green only played five offensive snaps in Week 2, so let's ignore that game. Using his average number of routes run in Cincinnati's other three games, Green's 3.53 YPRR means he's currently on pace for 1,571 receiving yards in 15 games. That's worth over 50 fantasy points more than the 1,028 yards you would expect from his "true" yardage-producing ability across the same number of routes.

It's also worth noting that Marvin Jones and Tyler Eifert will be returning from injury in the next month or so. Oh, and late-breaking news is that Green had to be carted off the practice field on Wednesday due to an aggravation of the aforementioned Week 2 toe injury. 

ppr buy-low candidate #1: roddy white

White ranks 35th in receiving points despite ranking 19th among WRs in number of routes run. That's because his RPRR and YPRR so far in 2014 are both well below their "true" values based on his time in Atlanta. White's True RPRR of .152 and True YPRR of 2.06 suggest he should have 11 more receptions and 160 more yards than he does right now. Non-TFP factors also suggest an up arrow: The Falcons have five games remaining against bottom 8 defenses, and their offense continues to have one of the strongest pass identities in the league.

ppr buy-low candidate #2: marques colston

Similar to White, Colston is massively underperforming his True RPRR and True YPRR. Given the number of routes he's run so far this season, he should have 27 receptions for 362 yards; instead he has 15 receptions for 248 yards. He also has a couple of other things working in his favor. First, I identified Drew Brees as the #2 buy-low candidate last week, so Colston figures to benefit if TFP is right about his QB. Second, with New Orleans on their bye week, the Colston owner in your league is likely to be willing -- if not downright eager! -- to trade him.

ppr sell-high candidate #1: jordy nelson

Compared to his performance over the past six years in Green Bay, Nelson is going bonkers in 2014. As I mentioned earlier, the league averages this season for RPRR, YPRR, and TDPRR are .121, 1.54, and .009, respectively. Through Week 5, Nelson currently sits at .210, 3.24, and .025!!! (Even bold, underlined italics and exclamation points don't do his outlying performance justice.) What's more, as he's been mostly a "yards and TDs guy" in Green Bay, Nelson's career True RPRR is only .138, or 13.8 catches for every 100 routes run. Translating that to the number of routes he's run so far this season, Nelson's 34 receptions are 12 more than what you would expect from his "true" reception-producing ability. Regardless of what you think of him or the Packers offense, it's highly unlikely that Nelson can sustain this pace for the rest of the season. And considering his astronomical perceived value at the moment, you probably won't be able to get more for him in the future than you can in the present.

ppr sell-high candidate #2: DEZ BRYANT

Nearly everything I just wrote about Nelson can also be written about Bryant. He's overachieving no matter which per-route stat we look at -- albeit to a lesser extent than Nelson -- and Bryant's another case of a "yards and TD guy" becoming a reception machine out of nowhere. In Dallas, his "true" ability is 14.3 catches for every 100 routes run; right now he's averaging 20.5 per 100. In other words, he has 32 receptions, but he should have 22 based on the number of routes he's run this season. And of course, Bryant's also a player whose value is through the roof at the moment, so maybe it's time to get rid of him while the getting-rid-of-him's good.

Follow @IR_DannyT