Welcome to this week’s Stacking the Box Score, where I’ll be quantifying what went down between the Kansas City Chiefs in their most recent game — in this case, Sunday’s 23-16 win over the Denver Broncos.
As always, here’s a quick primer on some of the stats and data I’ll be using. If you already feel comfortable with these metrics, skip the link and keep reading!
Chiefs vs. Denver Broncos, December 15th
Not much to note in this win probability chart — other than only one play had greater than a 10% swing: the Chiefs’ first touchdown pass to Tyreek Hill. After that, it was smooth sailing with no huge plays for either team.
Something cool happened on Sunday — other than the absolute dominance the Chiefs had in all three phases of the game.
With his first catch, tight end Travis Kelce recorded his fourth-consecutive 1,000-yard season. He’s the only tight end in NFL history who has ever done it.
It’s an incredibly impressive feat — and Kelce is undoubtedly one of the best tight ends in the game. But as I’ve covered in this series before, yards might not be the best metric for evaluating player production.
So I looked at tight end performance using some other metrics. Our play-by-play dataset only goes back to 2009, but we know that today’s tight ends play a bigger role in the passing game than ever before; it’s fair to assume that without adjusting for era, tight ends of the past decade would outperform Hall of Fame tight ends like Mike Ditka in these metrics. Let’s take a look.
My first instinct was to plot Success Rate (how often a target to that player improves the teams’ chances of scoring) and Expected Points Added (EPA) per target. Using these metrics, how do the tight ends of this decade look?
Here we see players like Tyler Eifert, Ladarius Green and O.J. Howard — who certainly aren’t bad players, but might not be the names you’d guess to be the most dominant tight ends of the decade.
The problem is that we’re using metrics (EPA and success rate) based only on when a player is targeted. But winning targets (getting open on routes often enough for your quarterback to look your way) is a big part of being a successful NFL receiver.
So ideally, we’d want EPA per route run — but in lieu of that, we can look at two stable (that is, relatively consistent year-to-year) metrics that correlate well with production: air yards (the total amount of intended yards gained through the air on targets to that player), and yards after the catch. To make it fair, we’ll look at these on a per-game basis.
This is starting to look better. The annotated tight ends are certainly the ones who play the biggest roles in their teams’ passing games. Since 2009, Kelce ranks second in the league in yards after the catch but is well behind George Kittle — the YAC king. But as you see, on average, Kelce is targeted on deeper routes.
What about which tight ends have had the most impact on the games they’ve played — both by how much they helped their team score and how much they helped their team win? We can get a sense of that by plotting EPA against Win Probability Added (WPA)
Now we’re getting somewhere! On a per-game basis. This is likely the best we can do at determining which tight ends have been the most dominant. Kelce doesn’t top this list, but he’s head and shoulders above everyone except Rob Gronkowski — who has arguably had to compete for targets with fewer teammates than Kelce has.
In the second tier, former Chiefs tight end Tony Gonzalez stands out — along with Kittle and AFC West tight ends Darren Waller and Hunter Henry. In the third tier, we see notable players like Jimmy Graham, Antonio Gates, Greg Olsen and Jason Witten.
Our metric looks pretty good — but it is on a per-game basis. What about longevity? Let’s look at which tight ends have added the most to their team’s chance of scoring since 2009 (or since they entered the league).
This isn’t very fair to players like Gonzalez, Gates and Jason Witten, who entered the league well before 2009; we’re only capturing the latter portion of their careers. However, we can see a trend similar to the last graph: Rob Gronkowski has been the most dominant tight end in the NFL since 2009 — with Kelce not far behind.
So when we’re talking about the best tight ends of the decade, it’s clearly a horse race between Gronkowski and Kelce — and Gronkowski has the edge. Can we find any other differences between them?
When we look at the two players’ EPA distributions (how often particular EPA values occur in their plays) we see that both players have similar frequencies of negative plays. Kelce has a higher frequency of positive plays in the lower range, but Gronkowski has a slightly higher frequency of big plays — that is, greater than 2 EPA.
Is there anywhere Kelce has an edge?
Well... yes! When we break down the EPA per target by downs, we see that on fourth down, Kelce has actually been a more efficient receiver than Gronkowski. In fact, around 75% of his targets on fourth down have gone for positive EPA — while the best 75% of Gronkowski’s targets have only gone for at least -2 EPA.
So there you have it. Advanced metrics say that among tight ends, Kelce and Gronkowski own the past decade — and it isn’t particularly close. But between the two, Gronkowski has the edge.
But Kelce has one advantage Gronkowsi can’t touch: he’s still playing. His metrics can still improve.
Hope you enjoyed this deep dive into the data. If you have questions, please feel free to leave a comment.