Print

QB Stats and Future Performance: Sack Rate

You are viewing free content provided by FantasyGuru.com. Why not consider subscribing today?

by Mike Horn, Staff Writer

Published, 4/16/14 

This is another in my series looking at what stats can tell us about QB's future fantasy performance.
 
Today I'm looking at what sack rate says. Sack rate is the percentage of time a QB gets sacked out of all the times he drops back to pass. The formula is sack rate = sacks divided by (pass attempts plus sacks).
 
I looked at all QB seasons with at least 100 pass attempts since 1988. I wanted to have a decent sample size of drop-backs against multiple opponents to screen out bad or fluky games.
 
This chart shows how sack rate has varied over time:
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
The years from 1988 to 2013 are along the horizontal axis, with sack rate on the vertical axis. The thick gray line indicates the average sack rate across the entire period – 6.5% for the QB seasons in the sample. In the '90s it fluctuated up and down until 1997. For the next 11 seasons, through 2008, sacks rates generally trended downwards. While the sack rates in the last several years were all below average until 2013, they have been steadily rising.
 
One other thing I learned looking at this data: Sack rates generally drop as QBs get older and more experienced. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
This chart is organized like the previous one, except that QB age is along the x-axis. The downward trend appears to end after QBs pass their 38th birthday, but the numbers after that are from only a very few QB seasons so are a bit unreliable. There is some bias in these numbers because good QBs are the ones who keep getting 100+ attempts as they age, and good QBs are generally less likely to be sacked.
 
I divided the QB seasons in my sample into four categories:
 
High sack rate: over 9.1% (the average sack rate of 6.5% plus one standard deviation of 2.6%).
Above average sack rate: between 6.5% and 9.1%
Below average sack rate: between 3.9% and 6.5%
Low sack rate: under 3.9% (the average minus one standard deviation.
 
I looked at how QBs that fell into these categories did in the year that they posted those rates, which I call the base year "Y." Then I looked at how they did in the next season, "Y+1."
 
Here are their stats in the base year:

Performance in Base Year by Sack Rate Category
Category
# of Players
AVG GP
AVG Pass Attempts
INT Rate
Pass FPG
HI
191
10.5
24.2
3.5%
11.5
ABOVE
355
12.2
28.3
3.2%
14.1
BELOW
435
12.6
30.1
3.0%
15.7
LOW
130
12.5
31.5
3.0%
16.8
Total
1111
12.1
28.8
3.1%
14.7

The table lists the sack rate categories, how many QB seasons (# of players) fit into that category, the average games played by each player (AVG GP), the average pass attempts in the games the QB played, the interception rate of the QBs in the seasons sampled (INT Rate = total number of interceptions by all QBs in each category divided by their total number of pass attempts), and then their PASSING fantasy points per game played (Pass FPG). I want to stress this is not their total FPG but just those that resulted from passing, with 20 passing yards = 1 FP and each passing TD = 4 FP.
 
QBs in the HI category played the fewest games, had the fewest pass attempts per game, the highest INT rate and fewest passing FPG. Now the fewer GP could be because these QBs are more likely to get hurt or just because they weren't as good as the QBs who were sacked less often. But notice that they had fewer pass attempts per game as well. Sacks are drive killers and therefore probably contributed to fewer opportunities to throw. These QBs threw interceptions at a slightly higher rate – as a group, they weren't taking sacks just to be conservative and avoid throwing into coverage.
 
Looking through the other categories, generally as the QBs' sack rates improve, their performance goes up. There isn't much difference between the BELOW average and LOW sack rate categories in GP or INT Rate, but the LOW QBs score about 1 FPG more from passing.
 
It's not clear if there is causation here; that is, does taking fewer sacks as a percentage of drop-backs make these QBs better, or if it's just correlation. But if the latter, then sack rate is a marker or indicator of better quality players.
 
All that though just says that sack rate describes a player's past performance. Does it tell us anything about future performance?
 
For example, do QBs in the HI category continue to get sacked at higher rates in the next season? They do, but there is some regression to the mean as you can see from the raw numbers:

Category in Base Year
Category in Y+1
HI
ABOVE
BELOW
LOW
Total
HI
33
48
19
3
103
ABOVE
47
107
73
9
236
BELOW
21
84
158
43
306
LOW
1
9
50
37
97
Total
102
248
300
92
742

Here's how to read this table: The first column is the categorization of QBs by the base year performance or year Y. The next four columns are the categories they fall into in the next year (Y+1). Note I've used only QBs who had at least 100 pass attempts in both seasons. The last column is the total number of players in each category in the base year. The totals in the bottom row are the totals in the categories for Y+1. The yellow highlights indicate the most populated category in each row – the base year's HI sack rate QBs are most often found in the ABOVE average category in Y+1.
 
If you look at the HIGH and LOW base year QBs, they most often move into the ABOVE and BELOW, respectively, categories in the next season. They are moving from an extreme category to one closer to the average – i.e., regression to the mean. However, in general, they stay on either the positive or the negative side of the average. Here's the percentages of where the QBs go to next year (the number at each intersection of categories divided by the total at the end of the row, on the far left):

Category in Base Year
Category in Y+1
HI
ABOVE
BELOW
LOW
Total
HI
32%
47%
18%
3%
100%
ABOVE
20%
45%
31%
4%
100%
BELOW
7%
27%
52%
14%
100%
LOW
1%
9%
52%
38%
100%

What these figures show is that 32% of the base year's HI sack rate QBs finished at the HI level again, while 47% were sacked at an ABOVE average rate. Only 3% of them made it all the way down to the LOW group. Overall, 69% of QBs who were in the HI or ABOVE average sack rate categories in "Y" finished in one of those two groups in Y+1. And 71% of those who took sacks at a LOW or BELOW average rate in the base year stayed in those categories in Y+1.
 
Here's the percentages calculated in a different way:

Category in Base Year
Category in Y+1
HI
ABOVE
BELOW
LOW
HI
32%
19%
6%
3%
ABOVE
46%
43%
24%
10%
BELOW
21%
34%
53%
47%
LOW
1%
4%
17%
40%
Total
100%
100%
100%
100%

In this table I've calculated where the QBs in next year's categories came from: the number in each intersection of categories divided by the totals at the bottom of the columns in the raw numbers table. The vast majority of QBs who did poorly in sack rate in Y+1 came from the poor categories in the base year (67%). And those who did well in sack rate in the following season did well in the base year (73%).
 
Note that the persistence in poor or good performance happens in spite of the general tendency of QBs to improve their sack rates as they age; every QB in the previous three tables was one year older in Y+1 than in the base year.
 
So, the bottom line: What can we learn about fantasy performance from last year's sack rate?

Performance Next Year by Base Year Sack Rate Category
Category in Base Year
# of Players
AVG GP Base Year
AVG GP Y+1
Pass FPG Base Year
Pass FPG Y+1
HI
103
11.8
11.9
12.5
13.2
ABOVE
236
12.9
12.4
14.8
14.9
BELOW
306
13.6
13.2
16.4
16.5
LOW
97
13.8
13.5
17.1
17.4
Total
742
13.1
12.8
15.5
15.7

This table again includes only QBs who had back-to-back seasons of 100+ pass attempts since 1988. Besides listing their base year categories and the number of players in each of those, I've shown their average (AVG) games played (GP) in both the base year and Y+1, as well as their pass FPG in both seasons.
 
Generally, players who got sacked less in the base year played more games and had more fantasy success passing the ball in the following season as well as in the base year.
 
Of course, passing the ball is only part of what we care about for fantasy. Here's the list of the top-20 QBs in FPG in 2013:

Player
Category
Rush FPG
Russell Wilson
HI
3.7
Aaron Rodgers
ABOVE
1.3
Alex Smith
ABOVE
3.3
Ben Roethlisberger
ABOVE
1.0
Cam Newton
ABOVE
5.9
C. Kaepernick
ABOVE
4.8
Nick Foles
ABOVE
3.1
Robert Griffin
ABOVE
3.8
Ryan Tannehill
ABOVE
1.9
Andrew Luck
BELOW
3.9
Andy Dalton
BELOW
1.9
Drew Brees
BELOW
1.5
Jay Cutler
BELOW
1.1
Matt Ryan
BELOW
0.3
Philip Rivers
BELOW
0.5
Sam Bradford
BELOW
0.4
Tom Brady
BELOW
0.1
Tony Romo
BELOW
0.3
M. Stafford
LOW
1.2
Peyton Manning
LOW
0.2

You may notice that the QBs in the HI and ABOVE categories tend to be the QBs who scored more FPG rushing the ball than those who were sacked at a lower rate. In fact, the higher sack rate QBs averaged 3.3 FPG vs. 1.1 FPG for the others. And even that number for the lower sack rate players is inflated by Andrew Luck – without him, they averaged only 0.7 FPG running the ball.
 
So while sack rate tells us something, it doesn't tell us everything. In fact, it may only be meaningful because it correlates with running the ball. So with that in mind, take these comments with a grain of salt:
 
·         Sack rate is one of the only stats that I've looked at so far that is negative for the 2014 outlook for Aaron Rodgers and Nick Foles.
·         And it is one of the only stats so far that is positive for Matthew Stafford and Andrew Luck.
·         Jay Cutler and Sam Bradford have a history of injury risk, but their 2013 sack rate category is an optimistic sign for their 2014 health – at least, relative to other QBs.
·         I'd be most worried about QBs like Ben Roethlisberger and Ryan Tannehill, who have poor sack rates in 2013 without providing compensating FPG production running the ball. Of course, upgrades to their offensive lines could change their outlook.

 

Bad draft picks are down 1,399%. Must be the Guru!

Back to the top