The Titans vs. the Colts may seem like a boring 1-1 vs. 0-2 matchup, but for my interests, there may be no more exciting game on the schedule this weekend than Marcus Mariota vs. Andrew Luck in just the third game of Mariota’s career. After Mariota’s Week 1 performance — four touchdowns in his first 13 pass attempts (before getting yanked for the entire fourth quarter like he was a Pro Bowler) in a win over the hapless Tampa Bay Buccaneers — Mariota’s cleats are already in the Hall of Fame. So that leaves us to wonder: Will Mariota eventually join them?
It might seem kind of early to try to answer that question, what with small sample size and all. But one of my favorite exercises in broader stats is trying to find situations where extremely small samples are sufficient to make broader inferences. For example, I mentioned last year that games as early as Week 2 or 3 in the NFL really can be “must-wins” — even for teams that have an established record. (Indeed, it may be one of the rare situations in which talking heads underestimate the significance of individual games.) Great performances by young athletes in sports are another case where small samples may be signaling something.
Let’s take this as an opportunity to jump the gun.1
So how does Mariota’s break-in game stack up against other Week 1 debuts, and what have similar performances signaled in the past?
I’ve written previously about the most important things to watch for when trying to size up the career prospects of rookie QBs: playing time and raw production (rather than passing efficiency). Sure, situations vary and some QBs benefit from sitting on the bench to learn the game for a few years, but on average, the bench-riders are less successful than those who earn their way onto the field immediately and are able to move the ball when they get there.2 But this isn’t a usual case of rookie-watching.
There have been 75 QBs since 1960 who had at least 10 pass attempts in Week 1 of their rookie year. Seven of them are now in the Hall of Fame. (For comparison, there have been 870 quarterbacks drafted in this period, 12 of whom are in the hall.)3 Of those 75, only 17 put up 7.5 adjusted yards per attempt (AY/A)4 or better in that first game. Mariota was the 17th:
Six of these quarterbacks are still active, but of the 11 who are eligible for the Hall of Fame, five have made it.
In other words, 45 percent of eligible QBs who had debuts anywhere close to Mariota’s turned out to be all-timers. And make no mistake, that rate is fantastic — even for a No. 2 overall draft pick like Mariota. For comparison, here’s how many of the QBs drafted in the first round since 1960 and who retired before 2010 have actually made the Hall:
Just three of the 12 eligible QBs drafted No. 1 overall have made it, and just one of the 40 QBs drafted between Nos. 2 and 11 overall has made it.
So Week 1 performance is a better predictor of Hall of Fame chances than draft position. But there’s more: Standout Week 1 performances are about as good (or better) of a predictor of Hall of Fame status as entire standout first seasons. For example, as mentioned above, of the 11 eligible QBs with the top Week 1 performances (by AY/A), five have made the Hall. Compare that with the 11 eligible QBs with the top rookie seasons overall (by AY/A, minimum of 250 attempts) — just three have made the Hall.5
Voila, Mariota is set for immortality, right? Not yet. There are two issues with Mariota’s claim. One is that auspicious debuts have seemingly gotten more common in the recent pass-happy era, so perhaps their significance should be discounted. Two, Mariota had only 16 pass attempts, which is an even smaller sample than usual. Though, to be fair, he practically lapped the field: If he had had 32 attempts and gained 0 yards on the second 16, it still would have been one of the top debuts by AY/A, and he still would have made the chart above, somewhere between Warren Moon and Jim Kelly.
Which brings us to his “let-down” loss against Johnny Manziel’s Browns in Week 2. Mariota had just over 8 AY/A in that game — less than half what he produced the week before! So let’s see where that stacks up against the other hot-handed rookies:
Mariota still had the third-highest AY/A in Week 2 of any of these 17 rookies while recording the fifth-largest number of attempts.
While 15 of 17 QBs (unsurprisingly) regressed in Week 2, there are other interesting patterns: Looking at just the 11 Hall-eligible QBs, the six who didn’t make the Hall — the more fluky ones, if you like — were more likely to have larger performance declines and were more likely to have fewer pass attempts in Week 2 than the five who are in Canton. Those who made the Hall either maintained a relatively high (7+) yards per attempt (Kelly, Moon) or increased their attempts (Bob Griese, Fran Tarkenton, Moon).6 Moon — who was only sort of a rookie — is the only one who did both. Until Mariota, that is.7 In other words, Mariota’s Week 2 loss wasn’t really a let-down at all. If anything, it puts him in the upper tier of high-performing debuts, and that tier has an amazing track record.
Any QB can have a two-game hot streak. But the (Bayesian) calculus for a rookie is very different than for a veteran, which allows us to read more into such a small sample than usual. When looking at unusually good performances like this, broadly speaking, there are two possible explanations: This player just happens to be the lucky one of dozens or hundreds of competitors. Or the player is really something special. When you have a track record of being average, the odds tend to favor the former. But for a rookie, when anything is still possible, the odds favor the latter.
Charts by Reuben Fischer-Baum.
Reminder: If you tweet NFL questions to me @skepticalsports, there is a non-zero chance that I’ll answer them in Skeptical Football.
Check out win and loss projections and playoff odds for all 32 NFL teams.