To Gronk, or not to Gronk?

We've faced this fantasy football dilemma for years. Do you spend up at the tight end position, snagging a player like Rob Gronkowski, or do you wait during your draft and hope to find this year's breakout guy?

Do you Gronk or Kelce, or do you, I don't know, Kittle or Burton?

Let's use some math and a little bit of logic to break down one of the more complex questions in fantasy football.

The Flaws of Value Based Drafting

The argument for drafting a Gronk or Kelce type tight end -- an elite one -- is that they're so much better than the rest of their position that, in turn, you gain a significant edge over your leaguemates by owning one of them. If you've got an elite tight end, because there are so few of them, your fantasy squad is winning at that position in a big way. Or, at least, that's the logic.

This is a Value Based Drafting, or VBD, way of thinking about things. As the legendary Joe Bryant penned years before the first Obama administration (OK, sure, sure, I'll stick to sports), Value Based Drafting says that "the value of a player is determined not by the number of points he scores, but by how much he outscores his peers at his particular position."

In today's fantasy football world, this may seem obvious to many. But when you think about it, Bryant's work on the subject has been the reference for so much analysis. Without the concept, fantasy football strategy would be lost.

Fantasy owners will use Value Based Drafting during their drafts -- that's why it's called what it's called -- but the same logic can be used in hindsight to see how valuable a player was during a particular fantasy season.

Take the 2017 season. In PPR formats, last year's top players at running back and wide receiver -- Todd Gurley and DeAndre Hopkins -- outscored the rest of the guys at their position by a good number of fantasy points.

Gurley was about 180 points better last season than the 12th-best running back, while Hopkins scored 87 more points than the 12th-best wideout. From a VBD perspective, Hopkins' point differential versus the 12th-best wideout was about the same that we saw between last year's RB4, Kareem Hunt, and the RB12.

Here's the same method but for quarterbacks and tight ends (Russell Wilson and Travis Kelce). Because most leagues start roughly half the number of quarterbacks and tight ends than they do running backs and wide receivers, we'll look at slightly different baseline measures.

Using VBD methods, you start to see that Gurley was the best player in fantasy football last year. (Surprise!) Almost any way you look at it, Gurley was the player giving you the largest advantage in fantasy football.

Now, this is just a simplistic way of doing things. Realistically, with Value Based Drafting, you'd want a more intelligent way to find baselines at each position and go from there.

But how do you do that precisely?

Exactly.

That's one of the first problems you may run into with VBD. That's not a huge deal, though. We can move past that.

A bigger issue with VBD is that you need to nail your projections. The exercise we just walked through is based on historical performance. It already happened. We knew that Todd Gurley was the top running back and that DeAndre Hopkins was the best wide receiver. Entering a draft, you don't know this. And because projections are going to have a lot of variance, not accounting for that variance can get you into trouble.

Moreover, there are a lot of assumptions that are made with VBD analysis. Russell Wilson, for example, was roughly 93 points better than the QB12 in fantasy football. Hypothetically, that means he'd be giving you a 93-point edge across the season over the squad in your 12-team league with the worst quarterback situation.

But fantasy football isn't a yearly game. It's a weekly one. You're making weekly lineup decisions while utilizing the waiver wire. superflex league study last month, in a competitive market, when the demand increases, the supply decreases. And the opposite holds true as well.

You're starting just one tight end in a traditional fantasy football league, so the demand is lower than what you see at running back and wide receiver. You need the position less. That's why we've seen the 12th tight end -- the last theoretical starter in a 12-team league -- drop off draft boards far later than the 24th running back and wide receiver through the years.

This creates a similar situation as we see with quarterbacks. Because reasonable tight ends are available so late, there's often a higher opportunity cost involved when you decide to pass on a wide receiver or running back early in your draft.

Opportunity cost, if you're unaware, is the loss of potential gain from other alternatives when one alternative is chosen. So when you draft a tight end early, you're forgoing the opportunity to have a top wideout or running back. And because those positions dry up much faster, it becomes more vital to snag them early and often.

Let's think of this another way, using the Value Based Drafting principles that we talked about previously. Pretend the top tight end -- who's still on the board at Pick 15 -- is projected to score 120 more points than the 12th-best tight end this year. Meanwhile, you're also considering drafting the WR7 and the RB9 at that spot. And both of those players are projected to outscore the 24th-best player at their positions by 100 points.

Should you take the tight end? Probably not, because that 12th tight end will be drafted around Pick 120, while the 24th-ranked running back and wide receiver will go five rounds earlier. So even though the difference in points between the top tight end to the 12th-best one is largest, it's not the optimal selection because of the supply and demand of the position.

This means that a tight end like Rob Gronkowski needs to be that much better than we might initially expect versus all other tight ends. Comparing his projection to a baseline projection isn't enough.

Predictability

There's more to player value than comparing the fantasy point totals from one player to the next. That's especially true at the beginning of the season when we're forecasting -- because our projections aren't going to be flawless, we should probably find other angles to value players properly.

One of those angles is predictability, or how good we -- as a fantasy football community -- are at predicting the success of players at each position.

An easy way of showing this is to chart average draft position (courtesy of MyFantasyLeague.com) by fantasy points scored. Because average draft position (ADP) tells us how we're feeling about a player at the beginning of the season, comparing it to how a player actually finished gives us some sense of how well we were able to forecast.

The chart you're staring at displays ADP versus PPR points scored over the last seven years at tight end, running back, and wide receiver. (We shouldn't care about quarterbacks here, R Squared of 0.26, while it's 0.34 at wide receiver and running back.

Things look slightly different when you break things down into distinct groupings.

For example, take a look at the R Squared values among only top-48 running backs (by ADP), top-48 wide receivers, and top-24 tight ends.

OK, so the relationship between ADP and fantasy points scored is still weakest at tight end when we shrink our overall sample and focus on more relevant players. Noted.

But take a look at what happens when we continue to block off players.

Among top-12 tight ends, there's been zero correlation between where a tight end was drafted and the number of fantasy points he's scored over the last seven seasons. When you consider the 12th tight end by ADP over the last seven seasons has been selected, on average, at Pick 120, that's pretty remarkable.

It's clear, though, that this has to do with tail-end TE1s than the mid-range ones. Because, as you can see, there's a decent correlation -- at least compared to running back and wide receiver -- among the top-six tight end group. This tells us that we may not be so bad at projecting who top tight ends will be. It's just that the later-round tight ends have really thrived over the last handful of seasons.

Weekly Predictability

Of course, this is looking at things from a yearly perspective, when fantasy football is played weekly. And that's where elite tight ends may have a unique edge.

If you read the aforementioned superflex study, you might recall the talk that surrounded the Coefficient of Variation, or CV for short. Instead of re-explaining the concept, here's what was said in that piece:

A lot of you know what standard deviation is -- it's a measure used to quantify the amount of variation in a set of data. A dataset with a high standard deviation means the data points within that set are located further away from the set's average. When the standard deviation is lower, each point will tend to be closer to the mean.

Comparing standard deviations from one dataset to the next can be difficult because each set is unique. Some may involve super small numbers, while others may have incredibly large ones.

So, in order to compare one set of data to the next, smart people created the Coefficient of Variation (CV), which takes a dataset's standard deviation and divides it by its average. This allows us to see the extent of variability in relation to the mean.

Basically, the higher the CV, the more variability in your data set. The lower, the more predictable.

Got it? Maybe not, but let's roll on.

Using 2017 data, here's a look at the CV averages among various groupings -- similar groupings to the R Squared analysis above.

For more detail, the TE1 to TE6 group includes the top-six tight ends in season-long scoring last year. So we're looking at Travis Kelce, Rob Gronkowski, Zach Ertz, Delanie Walker, Evan Engram, and Jimmy Graham. On average, those tight ends had a CV of 48.53%, which was similar to what we saw out of last year's RB1s. To put this another way, they were roughly as predictable on a week to week basis as a random RB1.

Looking at the remaining groups, one thing sticks out: tight end predictability drops similarly to running backs, even when we account for the fact that you're starting one tight end versus two or more backs. Unlike the other "onesie" position, quarterback, tight ends do see a significant dip in CV. And since we seem to have a better grasp as to who those top tight ends may be -- at least in the early rounds -- a competitive advantage arises: owning an elite tight end gives you consistency at an inherently inconsistent position.

The Gronk Edge

Earlier, we saw that, using Value Based Drafting methodology, last year's top tight end, Travis Kelce, wasn't more valuable than the top running back and wide receiver, Todd Gurley and DeAndre Hopkins.

But which running backs and receivers was he more valuable than from a VBD standpoint?

Yeah, yeah, I know -- I just wrote about how Value Based Drafting is flawed. And it is. But just humor me for a second.

One of the VBD flaws noted above was the difficulty in finding correct baseline players at each position. It's not really the main beef I have with VBD, but it's a small part of it.

Fortunately, years ago, Frank Dupont penned an article on Rotoworld that helped solve this issue. In it, he showed that in a normal 12-team league that starts two running backs, two wide receivers, and one tight end, the baseline running back should be RB34, the baseline wide receiver should be WR31, and the baseline tight end should be TE18.

Let's run with that.

With these new, intelligently-thought-out baselines, Gurley was still the best player in fantasy last year, scoring 242 more points than the baseline back. Kelce scored about 117 more points than his baseline tight end.

Kelce's value, per these new baselines, compared closest to last year's RB7 or RB8 (in between LeSean McCoy and Carlos Hyde) and Keenan Allen, who was the WR3.

According to Value Based Drafting, Travis Kelce was easily worth a second-round selection.

And that's what we're fighting against this season with Gronk. He's currently being drafted in the late-second round on FantasyFootballCalculator.com, near the RB13 and the WR9. And Gronk was actually pacing towards a higher fantasy total last season than Travis Kelce was.

Utilizing the same baselines as well as numberFire's season-long projections for 2018, Gronkowski's VBD outlook compares to our RB7 (Kareem Hunt) and our WR2 (Julio Jones).

Gronk seems like a screaming bargain. At least for VBD advocates.

Bringing It Together

To reiterate what was said at the top, there's more to player value than simply, "He's this many points better than the rest of his position."

As we've learned, the cost associated in acquiring a particular player should be factored into any value equation. That helps us understand opportunity cost better. And in addition, the predictability aspect of the game can't be overlooked.

So, should you Gronk, or should you not?

Remember when I said that Rob Gronkowski has to be that much better than the rest of his position in order to live up to an early-round price tag?

Well, that may be the case.

It's not as though Gronk is currently being selected where a Value Based Drafting methodology would tell you to draft him. He'd be a late first-round pick under VBD. Instead, he's a late second-rounder, making the opportunity cost in selecting him much lower than you'd see at a VBD-related price.

And, remember, the value of Gronk isn't just that he'll outperform his position, but he'll predictably outperform his position.

In truth, though, your 2018 tight end approach should center around a boom or bust strategy. You get Gronk given his cost isn't egregious, or you wait to get a low-end TE1.

The fact that there's been no correlation between fantasy points scored and average draft position among top-12 tight ends over the last seven seasons is telling: late-round tight ends have balled out. They've jumped into the elite tier of tight ends consistently. I mean, since 2011, the 12th tight end by ADP has outscored, on average, the 2nd, 4th, 5th, 6th, 7th, 8th, 10th, and 11th players at the position. Yes, small sample size alert, but it's not just that 12th tight end -- the TE9 by ADP has outperformed the TE2. The TE15 has done better than than TE4.

And thanks to supply and demand, when you can get those players so late, it makes less sense to grab that position early.

There's also the giant elephant in the room, and that's Rob Gronkowski's health. While anyone can get knocked out, we shouldn't ignore that Gronkowski has missed time in each of the last six seasons. That makes him slightly less appealing early in drafts. And it may push you off of him entirely.

So I'll ask one more time: should you Gronk, or should you not?

You can Gronk. Or, at least, data and game theory suggests that you can.

But should you?

All I ask is that you avoid those middle-round tight ends.