I’ll start with the conclusion: Gronk’s fantasy value in 2017 was 138.48 points, scored in 13 games for an average per-game value of 10.65 points. Kelce was worth 143.05 points in 15 games, for an average of 9.54 per game. Ertz’s fantasy value was 104.18 points in 13 games, an average of 8.01. This puts the trio’s value far ahead of everyone else in the singleton positions (quarterback and tight end). The rest of the article is about what this means, my methodology, and highlights of the study.
The Fantasy Value Project disagrees.
The Fantasy Value of the Top Tight Ends
It is considered uncontroversial that the top 2 tight ends are Rob Gronkowski and Travis Kelce. How did their fantasy values compare in 2017, and how did Zach Ertz, who had a breakout season, compare to the top 2? This article aims to answer that question in terms of fantasy points. If my analysis is correct, this value represents an approximation of how many fantasy points the player contributes to a typical fantasy team, given the alternatives available to the team. Another way of putting it is that a player’s fantasy value is how many points a typical team should give up in order to acquire the player, or how many points a typical team should get in order to give up the player, if players could be exchanged for points. (leaving aside the issue that in most leagues, winning is based on wins, not season-long points).
For exclusive content from the Fantasy Value Project delivered to your inbox, sign up here.
As discussed in my last post, fantasy value is optimized differential contribution. That is, a player’s value to a fantasy team is how many points he scored above his potential replacement, assuming the team’s owner was making good decisions. As in the Deshaun Watson case study from the last post, I will assume that “good decisions” means startring the eligible player with the highest weekly projection from fantasydata.com. If a roster has Gronk and Benjamin Watson as its tight ends, and Gronk projects to outscore Benjamin Watson, then for the purposes of this study, a good decision means starting Gronk that week. If Ben Watson projects to outscore Gronk that week (i.e., if Gronk was injured or on bye that week), then Watson is the optimal starter and starting Gronk is not a good decision.
As the formula requires, if a player should be started, his value that week is the difference between his fantasy points and those of the next best starting option who is on the bench. A player gets 0 points towards his value if he is not the player who should be started that week, and negative points if he should be started but is outscored by the best replacement option on the roster who is benched.
As also discussed in the Watson analysis, this is a crude first pass that makes assumptions about who the available alternatives are on any given week for any given team. Because Gronk, Kelce and Ertz were projected to be top 12 tight ends before the 2017 season (Gronk TE1, Kelce TE2, Ertz TE10, according to Fantasy Football Calculator‘s 2017 ADP) and because typically fantasy players do not load up on TEs, I will assume that on any given week, the best alternative to Gronk, Kelce or Ertz was any TE projected to be between the TE13 and TE24 that week, with equal chance of being any of them. This seems like a reasonable enough working assumption at this stage, and it makes the task of computing fantasy values pretty easy: given that Kelce, Gronk and Ertz were each projected in the top 12 every week in which they were projected to play, they are assigned a value of 0 any week in which they were not projected to play (week 10 bye for Kelce, weeks 5, 9, and 14 for Gronk, weeks 9, 10 and 14 for Ertz), and the difference between Gronk, Kelce or Ertz’s fantasy point production and the average of the TE13 to TE24’s fantasy point production every other week. As always, I do not count the usually fantasy-irrelevant Week 17, and I am assuming a PPR league with one tight end starter, and no tight end premium or other special scoring.
I’ll spare the reader all the details of the calculation in favor of some highlights and summaries. The collective performance of the tight ends projected as TE13 to TE24 was weak. Only twice did they combine for over 100 fantasy points (weeks 2 and 10), which is the same as saying only twice did they average more than 8.33 points. Gronk’s and Kelce’s scoring was high variance, but each of them had only three games in which they scored less than 8.33 points (when they played – weeks 1, 7 and 11 for Gronk, weeks 3, 6 and 12 for Kelce). Ertz had less variance but was more of a steady, lower value producer. He finished below 8.33 points twice (weeks 11 and 13).
The TE13 to TE24 combined for an average of less than 5 fantasy points twice (weeks 6 and 14) and 5.25 or less two other times (weeks 11 and 16). Meanwhile, Gronk exceeded 20 points 6 times, and Kelce 5 times. Ertz only exceeded 20 point once, but exceeded 17 points seven times. Gronk and Kelce only had one week each where they did worse than the average projected TE2 (week 1 for Gronk, week 12 for Kelce). Ertz did it twice, in weeks 11 and 13.
How do these performances compare to other players? I have not calculated any running back or wide receiver values yet, because I am still figuring out how to determine their opportunity costs. But the top 3 tight ends tower above the top quarterbacks and the other top tight ends.
The projected QB1 in 2017 was Aaron Rodgers, whose best week, week 5 where he scored 24.04 fantasy points, gave him a value of 11.3 based on a comparison to the projected QB13 to QB24. Compare to an average weekly value of 10.65 for Gronk, 9.54 for Kelce, and 8.01 for Ertz. Rodgers’ second highest week (week 3, 7.15 points) was below Gronk, Kelce and Ertz’s averages, and he did not come close to the averages in any of his other 5 starts (generously discounting his week 6, in which he was injured in the first quarter). His season total was 27.96 (again, not counting his -12.32 performance in week 6), which gives him a per-game average of 4.66. Other top quarterbacks include Russell Wilson (90.42 fantasy points, averaging 6.028) and Tom Brady (52.47 fantasy points, averaging 3.5). As argued in the previous article, Deshaun Watson, because he kept being projected low, ended up with a low seasonal value of 20.55 despite his explosive output.
The other tight ends whose values I have calculated so far are Greg Olsen, Jordan Reed, Jimmy Graham, and Kyle Rudolph, whose respective season totals were -8.83, 13.6, 56.26, and 49.14. None of them cracked 3.5 fantasy points per game on average.
To restate what I said at the top: Kelce was the top tight end in 2017, scoring 143.05 points in 15 games. Gronk was just behind at 138.48 in 13 games. Ertz had 104.18 points in 13 games. This group was far ahead of the other top TEs, and well ahead of the top QB, Russell Wilson, who was worth 90.42 fantasy points in 15 games.
Sign up here to receive exclusive Fantasy Value content and special offers by e-mail.