free NFL football picks
Home | Teams/Players | Free Picks | Office Pools | Research | Contact Us

NFL Playcalling Studies:
Part II -- League Wide Tendencies

If you missed part one, you may want to review NFL Team Playcalling Predictability.

We've assessed the league history from 1998 to 2002 (five seasons) to find out how often on average teams rush or pass in certain situations, considering the following factors:

- Down
- To-go yardage
- Quarter
- Score
- Field Position

From this we can state that NFL teams have rushed 53% of the time with 1st and 10 at their own twenty to thirty-nine yardline during the first quarter of a tied game...to 94% passes for a team facing 3rd and long in the fourth quarter down by a bunch.

A simple question is whether bucking the trends is a good idea -- by this we mean if a typical situation shows teams rush on say 70% of such cases, is calling for a pass (the "unpredictable" move) a wise risk to take? First off, we'll examine the results for the league as a whole in 2003, considering all plays during the course of the regular season:

NFL Playcalling Stats
Predictability
Range
Rush
Pass
Att
PSR
Big
Yds
EFR
Att
PSR
Big
Yds
EFR
.00 to .19
796
36%
19%
5.5
62
231
47%
8%
5.1
112
.20 to .39
2257
36%
14%
4.6
64
1624
49%
7%
5.2
110
.40 to .59
6844
44%
11%
4.1
77
6095
49%
9%
6.3
109
.60 to .79
3336
49%
8%
3.4
91
4898
42%
7%
5.6
85
.80 to 1.00
1190
47%
8%
3.3
87
5415
34%
7%
5.2
59

LEGEND: PSR = 'Play Success Rate' every play is recorded as either a success or failure depending on the situation, Big = Big Gain % of 10+ yards rushing, 20+ yards passing, Yds = adjusted yards per play, including penalties and ignoring non-plays like QB kneel-downs, EFR = 'Effectiveness Rating' a summary rating.

The 'predictability range' is simply how often NFL teams rushed/passed in that specific situation -- so the .00 to .19 range represents plays where the play call went against the league wide pattern (the league average called for the same rush/pass decision less than 20% of the time in similar circumstances) -- the most unexpected playcalls. On the other hand, the .80 to 1.00 range represents plays called in accordance with the NFL trends, and 80% or more of the time the league made the same rush/pass call as the team in the specific play.

What do we learn from the table? For starters, predictable rushes (rushing in situations when the NFL as a whole ran the ball 60%+) had the highest PSR and EFR, but lowest yards per carry and big gain rates...many of these plays presumably would have come on short yardage setups. However, it was unpredictable passes (those plays where the pass was called less than 40% of the time by the NFL) that had the highest EFR, but again with a reduced yards per play average.

Unexpected rushes (run called less than 40% of the time) had poor overall performance, but did see a high rate of big gains -- catching the defense by surprise once in a while.

Finally, woe be to the team in an obvious passing situation (80%+ pass call predictability), where the performance is dreadful across the columns -- a 34% play success rate, and a lousy 59 EFR rating.

Since it's such a Ying/Yang choice of two options the way we have it structured (we will break out play calls beyond simple rush/pass in future articles), it makes sense to present the table in a different view:

NFL Playcalling Stats, Rush/Pass Choice
Rush
Pass
Predictability
Att
PSR
Big
Yds
EFR
Predictability
Att
PSR
Big
Yds
EFR
.00 to .19
13%
36%
19%
5.5
62
.80 to 1.00
87%
34%
7%
5.2
59
.20 to .39
32%
36%
14%
4.6
64
.60 to .79
68%
42%
7%
5.6
85
.40 to .59
53%
44%
11%
4.1
77
.40 to .59
47%
49%
9%
6.3
109
.60 to .79
67%
49%
8%
3.4
91
.20 to .39
33%
49%
7%
5.2
110
.80 to 1.00
84%
47%
8%
3.3
87
.00 to .19
16%
47%
8%
5.1
112

Now we have it aligned so you can tell in a given 'prediction range' whether the run or pass call was the winner. Here's how it breaks out:

- Obvious passing situations (likelihood of a Pass 80%+)
Bad results either way (the 'Sylla and Charybdis' for offensive co-ordinators), but the unexpected rush has a fraction higher PSR and EFR, with big gains of 10+ (but still not a successful play if you get 12 yards on 3rd and 18!) occurring almost one out of every five tries.

- Likely passing situations (likelihood of a Pass 60%-79%)
Even with the defense aware that the pass is coming, it stil is the the clearly preferable playcall, with a higher PSR (42% to 36%), Yards per play (5.6 to 4.6) and EFR (85 to 64). Despite NFL coaches calling for the pass 68% of the time, they probably should pass even more.

- Balanced rush/pass situations (likelihood of a Run/Pass 40%-59%)
Here again the pass has been the much more rewarding playcall, meaning it's quite possible that for optimal performance, teams should PASS MORE under these circumstances. With an average yards per play of 6.3 passing to 4.1 rushing it's a big difference.

- Likely rushing situations (likelihood of a Rush 60%-79%)
Both options show some promising results, with a 49% play success rate whichever you choose. This might be an area where it's worth looking at the specifics of the offensive and defensive team to understand if it's better to attack the strength or the weakness of the opponent.

- Obvious rushing situations (likelihood of a Rush 80%+)
Despite some solid enough numbers running the ball, the pass calls have been reasonably effective under these conditions, which makes us think that NFL offensive co-ordinators could playing it too conservative here. 'Running out the clock' type events are certainly part of the context, but we'll take a look at a data set that excludes those moments next.

All right, but you may be skeptical -- there could well be some biases slipping from teams running out the clock or passing against prevent defenses. So rid ourselves of those cumbersome elements, let's rerun the numbers but only look at plays occurring in the first quarter...

NFL 1st Quarter Playcalling
Rush
Probability
Rush
Pass
Att
PSR
Big
Yds
EFR
Att
PSR
Big
Yds
EFR
.00 to .19
11%
42%
18%
5.7
72
89%
34%
7%
5.4
67
.20 to .39
32%
18%
13%
4.5
28
68%
31%
8%
6.1
71
.40 to .59
56%
43%
11%
4.2
73
44%
49%
9%
6.6
109
.60 to .79
66%
55%
9%
3.6
107
34%
49%
7%
5.2
109
.80 to 1.00
76%
60%
6%
3.1
127
24%
41%
4%
4.3
102

- Obvious passing situations (likelihood of a Run <20%)
Runs produce better average yards per play, and better play success rates, but were only tried 11% of the time. Mix it up a little folks!

- Likely passing situations (likelihood of a Run 20%-39%)
Wow! Our eyes blinked on this one. The rushes under these circumstances produced an 18% PSR which is shockingly low (and we're talking 143 attempts here, so not a trivial sample). Passing results weren't great, but clearly defenses are adamant about still keeping an eye on the run at this early stage of the game.

- Balanced rush/pass situations (likelihood of a Run 40%-59%)
One theory goes that teams like to 'establish the run' and maybe that explains why we see more rushes in this category even though passing has produced measurably better results overall.

- Likely rushing situations (likelihood of a Rush 60%-79%)
Both options are good, with rushes providing a very healthy 55% PSR, albeit with less upside than the pass.

- Obvious rushing situations (likelihood of a Rush 80%+)
Well, it looks like you are better off running in the first quarter when faced with these challenges, as rushes in 2003 had a whopping 60% PSR, and a dynamite 127 EFR. Short yardage passes are okay performers, but it's not worth pulling the surprise apparently at this stage of the game.

That's it for the league-wide patterns...the next article will go team by team through the above breakouts!


Rate this Feature
Check the circle that reflects your opinion
Poor   Fair   Good   Excellent

Comments

Email (optional)

We want your feedback! Tell us your thoughts

Copyright 2004 by TwoMinuteWarning.com, All Rights Reserved