That’s usually why I avoid QBs coming off career years. It’s generally counter-intuitive to believe a player coming off a career year will be undesirable, like Gary Barnidge this year. But career years are called such for a reason: They’re almost never duplicated. These are cases of something unusual happening, which usually means the opposite will occur in the near future.

-- 2016's Lessons Learned

So last week I analyzed how to identify if a QB had a career year, which built off of a previous piece on QB production after career years. Besides providing you with fascinating reading (if you're a stats geek like me), the point was to use this analysis to see which QBs had a career year if 2016.

While I won't try to cover everything from the previous articles again here, I will try to save you from going to the links on some key points. So to re-cap the definition of a "career year" as it applies to the 2016 QBs:

  • It was a QB's best season in BOTH total FP and FP/G, minimum 200 attempts, 25 attempts per game, and 7 games.
  • The career year had to be ranked in the Top 12 in FP/G for that season's QBs.

The scoring system I'm using is 25 pass yards = 1 FP, Pass TD = 4 FP, 10 rush yards = 1 FP, Rush TD =6, no turnover penalty.

Here's the 2016 Top 12 by FP/G:

Top 12 QBs in FP/G for 2016: Was it a Career Year?

Player

Age

FP/G

Max FP/G

Career High in FP/G?

FP

Max FP

Career High in FP?

Aaron Rodgers

33

24.9

27.3

No

398.0

409.4

No

Drew Brees

37

23.1

26.1

No

370.3

417.6

No

Andrew Luck

27

22.6

24.7

No

339.7

395.7

No

Matt Ryan

31

22.6

22.6

Yes

361.5

361.5

Yes

Tom Brady

39

21.7

25.9

No

260.6

414.0

No

Kirk Cousins

28

20.6

20.6

Yes

330.3

330.3

Yes

Ben Roethlisberger

34

19.7

20.5

No

276.2

328.2

No

Philip Rivers

35

19.4

19.6

No

310.9

313.6

No

Blake Bortles

24

19.4

22.5

No

310.1

360.1

No

Marcus Mariota

23

19.2

19.7

No

287.9

287.9

Yes

Dak Prescott

23

18.9

18.9

Yes

302.9

302.9

Yes

Tyrod Taylor

27

18.9

20.2

No

282.9

282.9

Yes

The chart starts with the names and ages (as of 31 Dec 16) of the Top 12 QBs last year, followed by the FP/G they posted last year. "Max FP/G" is each player's best year in FP/G through 2016; the next column answers the question if this was a new career high for him. The next three columns are the same thing but for Total FP rather than FP/G. The table is sorted by 2016 FP/G (3rd column). Attempts per game are not shown, but all 12 had over 25 attempts per game.

Seven players did not have career highs in either FP/G or Total FP: Rodgers, Brees, Luck, Brady, Roethlisberger, Rivers, and Bortles. So I won't address them.

Three fully met my definition of "Career Year:" Ryan, Cousins, and Prescott. Taylor and Mariota had career highs in FP but not FP/G. They fall into a more technical category of career year than the summarized definition above, so for now I'll keep them in the analysis.

In last week's article I outlined two ways to spot a career year. Age was a key determinant:

Career Year Age

Career Year

% of Career Years

Cumulative% of Career Years

22 to 24

8

9%

9%

25 to 28

46

49%

57%

29 to 31

21

22%

80%

32 to 34

12

13%

93%

35 to 37

7

7%

100%

A second was: QBs in the age 25-31 "sweet spot" for career performances who post Yd/G better than 117% of their previous high, or TDs more than 137% of their best season are more likely than not having a career year.

And there was one way to tell if it was NOT a career year: if a season did not top these statistical minimums (see full explanation of these stats):

Percent Change on Previous Maximums in Statistical Categories by Top 12 Fantasy QBs

Statistical Category

Minimum Change in a Career Year

Att/G

75%

Yd/G

83%

Total TDs

81%

TD %

27%

Y/A

48%

Comp %

73%

Here's my thinking then on whether the five QBs left from the Top 12 had a career year in 2016:

Dak Prescott: No. This is pretty obvious. Few QBs have a career year before age 25; only 5 have had one at Prescott's age of 23. Because he was a rookie under 25, I can't apply my other two criteria. Of note, if I'm wrong, is that QBs with over 20% of their value from running the ball (Prescott was at 21%) fall further the next season than more traditional pocket passers. But if I'm right, stick around until the end of the article. 

Matt Ryan: He was 31 in 2016. While it's possible for QBs to have a career year at 32 or older, 80% of all career seasons have occurred by the time a QB is Ryan's age. So that's a strong indicator that 2016 was as good as it gets for Ryan. On the other hand, Ryan's Yd/G were only 105% of his previous best and TD only 119%. So he didn't top the 117% Yd/G and 137% TD/G criteria. Ryan's percent changes in the six categories in the previous chart exceeded the career year minimums, so those don't indicate this was a NOT career year (sorry for the double negative). You can form your own judgment on the subjective criteria of losing his OC or having a hangover from the Super Bowl loss. But using my stat criteria, the message is mixed. Based primarily on his age, my final answer is yes, 2016 was his career year. 

Kirk Cousins: At age 28, he's right at the point where about half of all QBs have had their career years. So this is a 50-50 indicator. His Yd/G were 108% of his previous best and his 25 TDs were actually down from his career best 29 (86%), so neither of these point to 2016 as his career year. He topped all the CY minimums for percent change, so I can't rule out this was a career year on those grounds. Again, there are a number of subjective criteria to consider including his receiving corps, his new OC, his contract situation, and even where he'll be playing – a lot of that will become clearer as free agency plays out. Still, I think the numbers don't clearly point to a career year for Cousins (on the other hand, see this article about his future performance). 

Tyrod Taylor: Taylor's 2016 was probably not a career year because while he had his most total FP (282.9), his FP/G (18.9) were well below his 2015 high (20.2). His total was higher in 2016 because he played one more game, but even then, he only beat 2015/s total FP by less than one full point. The better question is probably if 2015 was a career year. He was 26 that season, which is an age at which most QBs still have their career years in front of them (only 38% of career years occur at age 26 or younger). But because he had substantially no QB experience before 2015, as with Prescott, I can't use my other criteria. As with also with Prescott, Taylor is a running QB – a third of his 2016 FP came on the ground, so if it was a career year, he also is likely to take a bigger drop than most career year QBs (but.. see below). At this point for Taylor, it seems like opportunity is the key in whether or not his career year occurred in 2015 or 2016. He may never get another chance to have 200+ attempts. But if he does, there are a number of QBs who had their career year after age 31 who had unusual starts to their career—like Taylor has—and didn't get a full shot at QBing in the NFL until later: Trent Green, Jeff Hostetler, Rich Gannon, Joe Theismann, and Doug Flutie to name several. And that's not mentioning Hall of Famers like Warren Moon and Steve Young. My prediction: if Taylor gets another year with over 25 attempt per game, he will top his 2015 and 2016 performance.

Marcus Mariota: Like Taylor, his Total FP (297.9) was a career high but his FP/G (19.2) was not. But in this case, Mariota beat his previous high total by about 50 FP and was only a half point below his rookie FP/G. He did post 137% of his previous best TD total, right at the cutoff for indicating a career year, but he only threw for 97% of his previous high Yd/G. (Mariota, although he got a lot of points running the ball, did not break the 20% threshold to be called a "running QB" in this study). Combined with his young age, 23, the same as Prescott, it seems like a good bet that 2016 was not his career year.

By this point, you may not remember why we even care if it was a career year. The next charts are meant to put that in context for you: 

Rank

Career Year

Non-Career Year

Change in FP/G in Year+1

Change in Rank in Year+1

Average Rank in Year+1

Median Rank in Year+1

Change in FP/G in Year+1

Change in Rank in Year+1

Average Rank in Year+1

Median Rank in Year+1

Top 4

-4.7

-8.8

11

8

-2.8

-9.0

11

6

#5-8

-3.6

-13.3

20

14

-1.7

-8.0

15

11

#9-12

-5.2

-19.6

30

25

-1.6

-8.3

19

15

All Top 12

-4.4

-12.9

19

15

-2.0

-8.4

15

11

Three weeks ago, I showed the career data. Now I've added the non-career side. In almost every measurement, Top 12 QBs in non-career years lose less of their fantasy value in FP/G and rank in the next year (Year + 1). For example, a Top 4 QB following a career year will see a drop of 4.7 FP/G (minus sign means decline or drop in scoring) but his non-career year counterpart only loses 2.8 FP/G. There are exceptions: those same Top 4 QBs have almost the same decline in rank, but career year guys lose a fraction less in fantasy rank (8.8 vs. 9.0). However, the median rank in the next season is better for the non-career year players.

The fact is…if you look at the bottom row in this table, almost everyone loses value (points and rank) after a Top 12 year. This is reality in the fantasy world: most players, regardless of position, decline after a good year, for a variety of reasons. Aging is a factor. So is health: to do well, a player has to stay healthy – and some (not all) of that is luck; chances are an injury will catch up to a "lucky" player in the next year. Also, coaches of successful players get hired away and chances are the new coach isn't as good. The reasons go on and on. 

Since career years by definition are better than non-career years, of course the decline is likely to be bigger after the former. The key is the magnitude – and of course, spotting the career year. 

One last point on this table: in terms of median rank in the next season, non-career Top 4 QBs do pretty well, coming in at #6. Yes, they will drop a few points per game, but they started at a higher level. The key is paying the right price at draft (or auction). Take them too high and you'll be disappointed. But in leagues (often high-stakes formats) that take QBs lower (or cheaper), there comes a point at which the QB coming off a Top 4 season, especially a non-career one, is a good value compared to a QB who was #5 thru #12. The flipside is trying to find an ascending player – because whatever the averages (means or medians) some players get better. They are just not as numerous. 

What about age? 

Age

Career Year

Non-Career Year

Change in FP/G in Year+1

Change in Rank in Year+1

Average Rank in Year+1

Median Rank in Year+1

Change in FP/G in Year+1

Change in Rank in Year+1

Average Rank in Year+1

Median Rank in Year+1

22 to 24

-7.8

-23.5

31

23

-1.0

-4.7

11

6

25 to 28

-3.9

-11.0

17

11

-1.8

-8.0

15

11

29 to 31

-3.9

-12.8

19

20

-2.5

-9.4

16

12

32 to 34

-4.1

-10.9

15

14

-1.8

-7.8

15

11

35 to 37

-6.6

-17.6

22

15

-2.4

-10.2

16

13

All Top 12

-4.4

-12.9

19

15

-2.0

-8.4

15

11

Young QBs (ages 22 to 24) are the most volatile. After a Top-12 season, those who did not have career years were the best group in this table at keeping their value: they saw the smallest decline in FP/G and had a median rank at #6 well above every other category. The problem is, the same age QBs coming off career years collapsed to the worst performance of any of these categories. And while it's rare for a QB to have a career year between ages 22 and 24, you can see the danger of making the assumption that Prescott and Mariota didn't have a career year – which I did above. Since we have more data on Mariota, it's a little easier to say this was not a career year for him (and by one measure, he did better at age 22). But he also will probably have a higher draft cost than Prescott (in part, because we know more about him).

The QBs aged 25 to 28 in some ways are the safest group. There is no difference in their median rank in the following season regardless of whether they had a career year or not. And in both categories their decline in FP/G is below average. For a player like Cousins, this gives you some comfort in drafting him (he's still likely to decline though).

Note that Tom Brady doesn't fall into these age groups. At some point age will catch up with him but there just aren't enough QBs who've played into their 40s to really draw any kind of stat-based estimate on what he'll do. I'd also be a little careful with the age 35 to 37 numbers (Drew Brees and Philip Rivers), although the sample of non-career Top 12 years appears to be a decent size (32 cases – only 7 career years, a warning sign), some of these cases are multiple years of the same guy: there are only 23 QBs in this group. 

Finally, there's the running QBs (those with over 20% of their FP on the ground):

Source of FP

Career Year

Non-Career Year

Change in FP/G in Year+1

Change in Rank in Year+1

Average Rank in Year+1

Median Rank in Year+1

Change in FP/G in Year+1

Change in Rank in Year+1

Average Rank in Year+1

Median Rank in Year+1

<20% Rushing

-4.0

-11.8

17

15

-2.1

-8.9

16

12

>20% Rushing

-5.8

-17.7

24

20

-1.6

-6.0

12

6

All Top 12

-4.4

-12.9

19

15

-2.0

-8.4

15

11

Running QBs coming off career years do badly in the next season, much worse than their pocket-based brethren. But those NOT following up to a career year actually are more reliable than average. This is a key caveat to what I said about Prescott and Taylor initially. In both cases, if they had a career year, they would be much worse choices than the other Top 12 guys. But if they did not – and my judgment is they did not, although in Prescott's case that's based solely on his age – they are likely to keep more of their value than the other QBs. This increase the risk-reward calculation on Prescott beyond what I said about his age group. In my opinion, it makes Taylor a great target if he gets a starting job at some point in 2017: his running value will remain and he's very likely going to be on a team that throws more than the 2016 Bills (they were last in the league in attempts, although Taylor took so many sacks that he held them down in that category; they were 30th in QB dropbacks. The QBs playing for Buffalo when Taylor wasn't got sacked at a higher rate than him, so it wasn't just his fault).

 

From all this, next time I'll make some estimates of how the 2016 Top 12 QBs will do next year.