FanPost

Yards versus points under the microscope: what really matters (for all the wonks)

There's a furious debate raging on BTB. Last year Dallas's offense was ranked #2 in yards gained and ranked #14 in points scored (for a delta of 12).One side says that’s Jason Garrett's fault. Their position is that there is an inherent problem with the Dallas’s red zone offense (e.g. Garrett playcalling, offensive line, etc.). The other side says that's nonsense. The same skills that you need to gain yards are what you need to score points. All you're witnessing is bad field position and random variation. So what do the statistics say?

The idea of red zone performance is exactly the type of claim I tend to be skeptical of. It feels to me like the "hot hand" in basketball and "clutch" hitting in baseball. When either phenomenon is subjected to rigorous analysis there's nothing there. The idea that an offense can be "clutch" or "choke" in red zone feels like the same type of claim. Currently people are ignoring Dallas's offensive productivity (#4 in yards gained) and are certain things are horribly wrong because Dallas has scored only 27 points. But doesn't that seem like the kind of claim that should be greeted with skepticism? Dallas did have a touchdown taken away by penalty and missed 2 field goals NFL kickers typically hit 80% of the time (Dallas would be #14 points scored with those points included). All this happened with 2 backup offensive lineman in one game and the starters back after a month off for the other game. Isn't it possible that the offense is okay? Are we really at the point where we completely ignore what is typically the primary measure (yards/game) of offensive performance?

The graphs below plot yards/game versus points/game across the league in 2009, 2008, 2007. If there's a correlation we'll see a pattern in the data. If the correlation is positive (i.e. more yards = more points) will see the data moving from bottom left the top right.

40563568_medium

via img707.imageshack.us

84174006_medium

via img828.imageshack.us

11441772_medium

via img843.imageshack.us

As you can see (and as you would expect) there's a strong correlation. Teams that gain more yards score more points. The R squared's were 0.76, 0.67, and 0.80. The way to interpret the R squared is that yards gained explains around 70%-80% of scoring.

Usually when looking at correlations we have to be careful about drawing conclusions regarding causality. The axiom is "correlation is not causation". However, in this case since a football team has to drive the field (i.e. gain yards) before they score, it's safe to say that yards gained calls the tune.The other 20%-30% is explained by defensive scoring (included in points totals), special teams scoring (included in points totals), field position (if a team consistently has bad field position it will hurt scoring), and variation in red zone efficiency.

I know what you’re saying ‘Really Fan, you wrote a fanpost to show that teams that gain more yards score more points. I recently discovered that grass is green and the sky is blue. Should I write a fanpost about that?’

So let’s dig a little deeper. Instead of yards/game vs. points/game, let’s look at yards/game versus points/100 yards so we can isolate how efficiently an offense scores (The rule of thumb is that teams score about 10 points for every 100 yards gained). Now it’s not intuitively obvious that teams will be different in scoring efficiency (i.e. points/100 yards). It could be that scoring efficiency is stable across the league. It would still be good to gain a lot of yards as a team would score more points (e.g. assuming constant efficiency of 10pts/100 yards a team that gains 400 yards/game would score 40 points, a team that gains 300 yards/game would score 30 points, etc). However, that’s not what we find. The charts below show the correlation between yards/game and points/100 yards. We see the same pattern as above (i.e. a positive correlation). Hence, there’s a double whammy: teams that gain the most yards are also the teams that tend to be the most efficient at scoring.

99358024_medium

via img529.imageshack.us

16844335_medium

via img512.imageshack.us

21689268_medium

via img27.imageshack.us

So far then we have found that the ability to gain yards is correlated with both total scoring and scoring efficiency (sounds pretty good, give me a team that can move the ball ... I trust points will follow).

Here’s another way to look at it. I calculated points per 100 yards gained across the league over the last 3 years. The chart below shows points/100 yards gained versus rank in yards/game (for example, in 2007 New England was #1 in yards/game and scored 8.9 points/100 yards gained, so New England is the 1st data point on the left of the 2007 series). Dallas is denoted by the black squares.

37083039_medium

via img210.imageshack.us

As you can see, in every year the teams that are best at gaining yards are also the most efficient at scoring points. The line is downward sloping from left to right. However, while there is a clear downward slope to the line there is also significant variability to the line (i.e. it's not a smooth line).

Therefore, I wanted to determine whether certain teams were consistently poor in terms of converting yards into points (although this is muddled by the inclusion of defensive and special teams scores in the points totals). I looked at the delta between team’s rank in yards/game and rank in points/game from year to year. If a team has an inherent problem in scoring points you might expect to see persistent deltas across years (e.g. last year Dallas was #2 in yards gained in #14 in points scored, a delta of 12. If Dallas has an inherent red zone problem we would expect to see that delta persist). The charts below show the deltas for 2007-2008 and 2008-2009 across the league.

32817528_medium

via img840.imageshack.us

85484516_medium

via img375.imageshack.us

This is admittedly the weakest part of the analysis. However, I’ve included what I found and leave it to the reader to draw your own conclusions. The problem is that the league changes so much it's hard to see stability across years for any statistic. One year an offense is #2 and the next year its #15. When I looked it yards/game across years to establish a baseline for comparison the R-squared for yards/game across 2007-2008 and 2008-2009 was only 0.07 and 0.51 respectively.

We see there is some weak persistence in the deltas. However, compare the pattern in these charts and the charts for yards/game versus points/game. The highest R squared in these charts is 0.25 which means that at most the intrinsic traits of the team explained 25% of the delta (2007-2008 suggests it's even lower than 25%). Due to the variation across years in all statistics it's difficult to draw any conclusion.

However , I also know that Romo was unusually poor inside the 10 in 2009 (44% completion percentage inside the 10 versus 63% overall). I also know that Brian Burke (see below) found that red zone performance is a myth. Further, I know that in 2008 the opposite was true about Romo, he was unusually good inside the 10 (73% completion percentage inside the 10 versus 62% overall). That suggests to me a strong possibility that we're seeing random variation.

To me, the evidence does not support the hypothesis that teams have red zone specific problems. I suspect that teams either have a good offense or bad offense. And isn't that just common sense? If a player or teams is good, won't they be good 90 yards from the end zone, 50 yards from the end zone, or 10 yards from the end zone? I would urge fans to be skeptical of claims otherwise.

Comments and criticism accepted gratefully.

Addendum I

While I believe that the overall quality of the offense dominates the other variables, the attractiveness of running relative to passing does increase in the red zone.

http://www.advancednflstats.com/2009/11/offenses-run-too-often-on-1st-down.html

For the vast majority of the field, the pass has a higher average payoff. Runs have a higher payoff only inside final 10 yards before the end zone, so it appears teams are passing too often there.

Therefore, the argument that an inability to run the ball hurts red zone performance is a valid. However, I suspect the overall quality of the offense trumps the effect of a weak running game.

If you're interested, here are the deltas between rank yards/game and rank pts/game over the last 3 years. I didn't observe a pattern of 'running teams' outperforming their yards/game rank (i.e. a positive delta). If you notice any pattern please leave a comment.

Deltas: Rank Y/G and rank Pts/G

2007

2008

2009

New England

0

-3

-3

Green Bay

-2

3

3

Dallas

1

-5

-12

New Orleans

-8

0

0

Indianapolis

2

2

2

Philadelphia

-11

3

6

Jacksonville

1

-4

-6

Cleveland

0

1

3

Seattle

0

3

-4

Cincinnati

-1

0

2

Denver

-10

-14

-4

Arizona

5

1

3

Minnesota

-2

5

3

Houston

2

-14

-6

Washington

-3

-9

-4

NY Giants

2

4

0

Pittsburgh

8

2

-5

Tampa Bay

0

-4

-1

Detroit

3

3

-1

San Diego

15

9

6

Tennessee

-1

7

-4

Baltimore

-2

7

4

Atlanta

-6

-4

3

St. Louis

-4

-3

-3

Oakland

2

0

0

NY Jets

1

7

3

Chicago

9

12

4

Miami

2

-9

2

Carolina

3

3

-2

Buffalo

0

2

2

Kansas City

0

-2

2

San Francisco

0

1

9

Bonus Table

% pass (pass plays/total plays)

Pts/100 Yds

Rank Pts/100 Yds

% pass inside 10 (pass plays/total plays)

Pts/100 Yds

NY Jets

40.7%

6.8

11

33.8%

6.8

Cleveland

48.8%

5.9

23

40.4%

5.9

Carolina

48.8%

5.9

22

34.3%

5.9

Tennessee

49.9%

6.3

17

38.3%

6.3

Cincinnati

51.1%

6.2

19

56.9%

6.2

Miami

53.4%

6.7

14

38.2%

6.7

Buffalo

53.6%

5.9

24

50.0%

5.9

Baltimore

54.2%

6.9

8

28.2%

6.9

New Orleans

54.8%

7.9

1

43.1%

7.9

Minnesota

55.9%

7.7

3

43.6%

7.7

Jacksonville

55.9%

5.4

28

39.1%

5.4

San Diego

56.0%

7.9

2

35.9%

7.9

Oakland

56.6%

4.6

31

40.5%

4.6

NY Giants

56.6%

6.9

9

42.5%

6.9

New England

56.8%

6.7

12

37.0%

6.7

Atlanta

57.0%

6.7

13

42.1%

6.7

Dallas

57.3%

5.7

25

51.7%

5.7

Denver

57.5%

6.0

21

48.1%

6.0

Kansas City

57.7%

6.1

20

60.4%

6.1

Pittsburgh

57.8%

6.2

18

45.8%

6.2

Green Bay

58.1%

7.6

4

58.3%

7.6

Tampa Bay

58.6%

5.3

29

62.5%

5.3

St. Louis

58.9%

3.9

32

51.2%

3.9

Houston

59.6%

6.3

16

42.3%

6.3

Washington

59.9%

5.3

30

48.8%

5.3

San Francisco

60.6%

7.1

7

48.9%

7.1

Detroit

60.7%

5.5

27

49.1%

5.5

Philadelphia

61.5%

7.5

5

56.3%

7.5

Chicago

61.7%

6.6

15

51.0%

6.6

Indianapolis

62.7%

7.2

6

58.8%

7.2

Seattle

62.8%

5.5

26

66.7%

5.5

Arizona

63.3%

6.8

10

52.1%

6.8

Total

56.5%

53.5%

Addendum II

While I was looking for the analysis of the attractiveness of running in the red zone I found that Brian Burke also looked at whether red zone performance "is real".

http://www.advancednflstats.com/2008/01/is-red-zone-performance-real.html

The table above lists 30 of the league's leading passers, and we would expect a handful of QBs to appear significant just by chance (this is known as a "Type I" statistical error). Further, the distribution of t-test values is evenly spread from 0.93 to 0.04. Rhere is no bunching of values toward the significance level of 0.05.

From this, it appears that despite all analysis to the contrary, there is nothing special about any particular quarterback's ability inside the red zone as compared to outside the 20. It's simply a random subsample of overall performance. If a QB is a good passer, he'll probably be good inside the 20.

Another user-created commentary provided by a BTB reader.