Chip Kelly’s Run/Pass Ratio

Posted on August 14, 2013

0


One of the most exciting parts about Chip Kelly’s arrival has been the mystique. Fans, and the 31 other NFL teams, have little idea of what exactly the NFL version of his offense will look like. How much will he run the ball?

Looking for hints, many have noted Kelly’s 2012 Oregon R/P Ratio of 686/373 (64.8/35.2). Of course, often this statistic has been presented with a disclaimer noting that Oregon was often playing with the lead, resulting in more run calls to keep the clock moving. Indeed, CFBstats.com offers some situational stats that can help us get a slightly better sense of Kelly’s R/P tendencies, and Kelly’s R/P ratio falls to 58.6/41.4 when we look at only the first half.

But this can be adjusted even further. College stats treat QB sacks as rushing attempts. Pass plays that ended in sacks or scrambles counted as runs at Oregon. (Scrambles are counted as rushes in the both the NCAA and NFL.) So how did Kelly do relative to NFL coaches in terms of calling passing plays? To find that we need to estimate how many of Oregon’s runs were actually scrambles or sacks. Then we can adjust the results, and compare them to NFL ratios.

Process:

In the 2012 NFL season, in the first halves of games, throws occurred on 88.2% of drop backs league wide (1.097 drop backs per throw). The lowest ratio of drop backs per throw was New Orleans at 1.03, while the highest was Seattle at 1.20. In general, teams with more mobile QBs tended to have higher rates of non-passes on passing plays, which makes sense. When looking at all regular season drop backs last year, Drew Brees and Peyton Manning’s ratios were 1.04 drop backs per throw, while Michael Vick, Robert Griffin III, and Colin Kaepernick all had ratios of 1.19, and Russell WIlson‘s was 1.21. Since Kelly just about always had strong scramblers at Oregon, it seems reasonable to estimate his QBs were on the higher end. But since we know Kelly deplores sacks (based on his NIKE Coaches Clinic Presentation — relevant excerpts here) and emphasizes quick throws, and also that Oregon had plenty of games against inferior pass rushers, I’ll just use the overall average of 1.1 drop backs per throw. So, yes, these are inherently pretty rough estimates. The point is that they get us close to what we’re looking for.

Thus, Kelly’s true number of pass plays (throws + sacks + scrambles) can be estimated as Throws*1.1. And whatever amount that adds to his pass play count, that amount must be taken out of his run play count.

Results:

With that approach here are Kelly’s Estimated First Half Run Play / Pass Play Ratios over his four seasons as head coach at Oregon:

Estimated First Half Plays

Run

Pass

2012

ORE

54.5%

45.5%

2011

ORE

47.4%

52.6%

2010

ORE

46.6%

53.4%

2009

ORE

51.4%

48.6%

Using play-by-play data (no estimates), here is what NFL teams did in the first half, for comparison (sorted by highest Run Play Percentage):

2012 First Half Plays

Run

Pass

KC

49.8%

50.2%

WAS

48.5%

51.5%

SEA

47.3%

52.7%

NYJ

45.3%

54.7%

MIA

44.6%

55.4%

MIN

44.4%

55.6%

CLE

42.5%

57.5%

BUF

41.8%

58.2%

CHI

41.6%

58.4%

SF

41.0%

59.0%

BAL

40.8%

59.2%

NE

40.0%

60.0%

CAR

39.7%

60.3%

CIN

39.5%

60.5%

HOU

39.2%

60.8%

JAC

39.0%

61.0%

TB

38.6%

61.4%

DEN

38.0%

62.0%

DET

37.4%

62.6%

STL

37.0%

63.0%

NYG

37.0%

63.0%

TEN

36.9%

63.1%

ARI

36.7%

63.3%

OAK

36.7%

63.3%

PHI

36.0%

64.0%

PIT

35.9%

64.1%

SD

35.8%

64.2%

DAL

35.2%

64.8%

NO

34.5%

65.5%

IND

34.1%

65.9%

ATL

33.5%

66.5%

GB

33.0%

67.0%

NFL

39.3%

60.7%

Thus, Kelly’s Oregon R/P Ratio is indeed high, but not too far out of line with what some teams have done in the NFL.

EXTRA POINTS

I gathered data all the way back through Kelly’s University of New Hampshire OC days, though I elected not to post it for a variety of reasons. I couldn’t find any situational splits for it, so it is difficult to present it in context, and they still did a lot of winning and running the ball, so there’s nothing really different/interesting to glean from it.

I did notice that Ricky Santos, who was not fast and ran a 4.82 40-Low, still had 119+ rushes in each of his three seasons. NFL teams average about 50 sacks plus scrambles per 13 games, so Santos must’ve been doing a decent amount of keeping the ball on the read plays. Matt Barkley ran a 4.85 40-Low. Nick Foles ran a 4.95 40-Low. So Kelly has shown a willingness to let slower QBs run the read-option and keep the ball. I’m not at all suggesting it would happen at the same frequency as Santos. Santos was running it against slower completion than Foles/Barkley would face, and was executing it before defenses saw a lot of it. But it’s interesting.

Advertisements
Posted in: Eagles, Uncategorized