The Importance of TANY/A* for QB Prospects
Last week I summarized my recent research on QB efficiency metrics, introducing a new metric — Total Adjusted Net Yards per Attempt, or TANY/A. It adds rushing into the equation, which I believe can improve the evaluation process for quarterbacks, especially for prospects coming out of college. TANY/A* takes it a step further and adjusts for schedule and era, and that is what I’ll be using here to evaluate past and present quarterback prospects.
What does a Top Prospect at QB look like?
With each attribute, we see a trend as we move down the draft. First-round quarterbacks are efficient, tall, and young, while seventh-rounders are none of the above. It makes sense — who wouldn’t want all three of those attributes in a QB? However, there are some interesting trends between the endpoints:
- In the first round, top-eight picks are the cream of the crop for all our attributes. There are some exceptions at the micro-level — Matt Ryan and Jake Locker each had a low TANY/A*, Ryan Tannehill/Eli Manning/Carson Wentz weren’t young when drafted, and several QBs measured under 6’2″ — but almost every top-eight pick at the position is a strong combination of efficient, tall, and young.
- In the rest of the first round, all three attributes are still prioritized, but there’s a drop-off in efficiency.
- There’s another drop in efficiency in the second round, but even more so for height. We don’t see short quarterbacks (6’1″ or shorter) drafted, but age and efficiency are prioritized with guys of average height (6’2″ or 6’3″).
- The third round flips, as we see a heavy emphasis on height, despite there still being efficient QBs of a similar age available later.
- The fourth round looks like the second round in terms of efficiency over height, but these passers come with a bit of an age tax.
- The fifth round shows teams prioritizing youth somewhat. Otherwise, the drop-off in efficiency continues through the end of the draft, where darts are being thrown with no attribute standing out.
- The undrafted quarterbacks (who play at least one game in the NFL) are even less efficient and shorter, but are younger on average than seventh round QB picks.
How does Draft Position affect the Hit Rate of QB Prospects?
We see above significant differences in the attributes of top prospects and late-round quarterbacks. When looking at hit rates, there’s a similar trend:
Here, “Lasted” means the prospect accrued at least two seasons worth of attempts (1,000 total — passing and rushing) in their NFL career. A “Hit” is a quarterback who met that threshold with average or better efficiency, a TANY/A* above 5.00. I filtered the data to those drafted through 2013, to give the players four years to meet the attempts mark (the length of a typical rookie contract).
Again, quarterbacks drafted with top-eight picks are clearly ahead of the rest. Other first round picks stick around for two seasons nearly as often, but only hit at half the rate. Day-two prospects see big drop-offs in both Last% and Hit%, 42.3% and 19.2%, respectively. Day-three quarterbacks are significantly worse, with a combined 5.9% hit rate.
Do Highly-Efficient QBs Outperform Draft Position?
I mentioned in my primer last week that college TANY/A* had a relatively decent correlation to both draft capital and NFL success. The question that naturally follows is whether NFL teams are accounting for college efficiency enough. So I broke the chart above into two groups, those with a college TANY/A* above and below the average for that range of draft position. In the chart below, those are signified by “High Eff.” and “Low Eff.” The averages for each range can be found in the first chart above.
- Looking at Last%, there is some variance between high and low efficiency quarterbacks in specific ranges, however…
- In the first round, High Eff. and Low Eff. quarterbacks last at an identical rate of 78.6%.
- For Day 2, 42.9% of High Eff. QBs last vs. 41.7% of Low Eff. prospects.
- For Day 3, 6.1% of High Eff. QBs last while it’s 8.6% for Low Eff.
- Looking at Hit%…
- For top-eight picks, High Eff. and Low Eff. quarterbacks hit at an identical rate.
- Day 3 quarterbacks hit 6.1% of the time if High Eff. vs. 5.7% for Low Eff. prospects.
- From picks nine through the third round, 33.3% of High Eff. quarterbacks hit, while that number is 15.8% for Low Eff. guys.
Overall, quarterbacks with below-average efficiency for their draft range last 31.1% of the time, but have a 16.4% hit rate. Quarterbacks with above-average efficiency for their draft range have an identical last rate of 31.1%, but hit more often with a 23.0% rate. In summary, after accounting for draft position, college efficiency does not affect a quarterback’s chances of accruing two seasons worth of attempts in their NFL career. However, QBs do see a modest bump in hit rate, specifically in the early rounds.
What type of prospect is primed for immediate success?
The first question to answer is which quarterback prospects get immediate opportunity, and the answer is simple. The ones drafted highly most often see immediate playing time. Here are the percentages of rookie quarterbacks to average at least 250 total attempts as rookies, filtered by their draft positions:
- Picks 1-8: 70.0%
- Picks 9-32: 52.6%
- 2nd-3rd Round: 28.6%
- 4th-7th Round: 4.3%
There are a couple obvious reasons why this is true. Quarterbacks drafted higher are less likely to have a good/established veterans ahead of them. Even if there is competition, highly-drafted signal-callers are more likely to be good in general, thus more ready to start early in their careers. It’s also possible that teams feel more pressure to give their premium investments snaps, whether that’s coming from a development point of view or to justify the draft capital spent.
The other question to ask is which quarterback prospects, if given immediate opportunity, are primed for success as rookies. This answer is a bit more complicated, but certainly more interesting. The attributes I have found significant for projecting quarterback prospects include draft position, college efficiency (TANY/A*), age, height, and the percent of college production they accrued through rushing (call it Rush% for short). For projecting career success, efficiency and draft capital are the most important attributes, with the others being minor but still relevant. However, when I checked which attributes correlate with rookie success, another attribute shined above the rest. Here are the correlations (r^2):
- Rush%: 0.157
- TANY/A*: 0.094
- Draft Position: 0.025
- Height: 0.010
- Age: 0.000
The two pieces of college production/efficiency data are the most important by far. When I created a model to predict rookie year TANY/A*, Rush% made up 60% of it, college TANY/A* 30%, and funny enough, age took the last 10%, with draft position and height being insignificant. Typically, being younger is good for a prospect’s long-term outlook, but being older helps as a rookie. The model improved on Rush% to get a correlation of 0.194. There is still a ton of unexplained variance, but that’s life as a football prognosticator.
So if you’re looking for a quarterback to succeed as a rookie, you want a mobile one who was efficient in college, and being older helps a bit. The player probably needs to be an early-round pick to get opportunity, so draft position is still important here even if it didn’t show up in the model. I’ll touch more on the 2018 class later this week in another article, but here’s how the top guys rank in rookie readiness, among 45 first-round picks since 2004:
- Lamar Jackson: 4th
- Baker Mayfield: 12th
- Josh Allen: 19th
- Mason Rudolph: 28th
- Sam Darnold: 30th
- Josh Rosen: 45th
At least five of the above quarterbacks should be selected in the first round of the draft next month. This is a historic quarterback class in that regard. In my next article, I will go into detail on what makes this crop of rookies so historic, and I will provide a preliminary set of rookie quarterback rankings.