Packing the Stats: Rushing to Conclusions

ALLGBP.com All Green Bay Packers All the Time

Packing the StatsAfter Sunday afternoon’s 30-20 victory by the Green Bay Packers over the St. Louis Rams, I listened to Jason Wilde’s weekly appearance on ESPN Wisconsin’s radio show “Pack Attack.” The conversation immediately dove into a debate between Wilde, Bill Johnson, and Homer about the effectiveness of Alex Green’s rushing attempts. While he made 20 rushing attempts the entire game, Green only netted 35 yards for a 1.8 yards per carry average. His longest run was for 15 yards.

On one side of the debate was Jason Wilde, who maintained that making the attempts to run the ball was more important than their overall yards per carry. He posited that the defense’s linemen would have to account for a run, even if it wasn’t for significant yardage. That means they couldn’t just “pin their ears back” and go after the quarterback each down.

Opposing this idea was Bill and Homer, who both insisted that Green’s yards per carry was unacceptable and would need to get better in the future to ensure offensive success. They claimed that if the running game isn’t making traction, then the defense doesn’t really have to worry about it, period. (Jason Wilde eventually called them “stubborn” in their opinions.)

So which matters more – yards per carry or total rushing attempts? This really piqued my interest from a statistical standpoint, and I decided to head over to Pro-Football-Reference.com to being my research. My sample data was all games (regular season and postseason) within the past ten years (2002-2011) that matched the rushing criteria below.

(You can download the complete Excel file here: rushing_stats.xlsx)

TOTAL RUSHING YARDS PER GAME (2002-2011)
TOTAL YARDS W L T W-L% COUNT GBP W-L% GBP COUNT
60-79 212 513 1 29.20% 726 30.40% 23
80-99 342 527 0 39.36% 869 68.60% 35
100-119 435 427 0 50.46% 862 61.30% 31
120-139 393 291 0 57.46% 684 75.00% 20
140-159 352 189 0 65.06% 541 61.10% 18
AVERAGE YARDS PER ATTEMPT (2002-2011)
YDS/ATT W L T W-L% COUNT GBP W-L% GBP COUNT
1.0-1.9 42 77 1 35.00% 120 33.30% 3
2.0-2.9 322 363 0 47.01% 685 65.20% 23
3.0-3.9 786 768 1 50.55% 1555 64.00% 50
4.0-4.9 746 745 1 50.00% 1492 52.50% 40
5.0-5.9 353 324 1 52.06% 678 64.00% 25
6.0-6.9 129 122 0 51.39% 251 20.00% 5
TOTAL ATTEMPTS PER GAME (2002-2011)
ATTEMPTS W L T W-L% COUNT GBP W-L% GBP COUNT
10-14 10 193 0 4.93% 203 33.30% 6
15-19 72 566 1 11.27% 639 13.00% 23
20-24 264 836 0 24.00% 1100 42.50% 40
25-29 614 636 0 49.12% 1250 74.40% 39
30-34 739 303 2 70.79% 1044 83.80% 37
35-39 568 99 0 85.16% 667 83.30% 24

These results, I think, are extremely interesting, and they help support the case made by Jason Wilde. One of the most surprising things is how little of a correlation there is between yards per attempt and winning games. The data suggests that anything after 2.0 YPC is not much different than flipping a coin to see who wins. While correlation does not equal causation, the correlation is pretty low, so having a big YPC average doesn’t automatically mean a better chance of winning.

Now, contrast that with total attempts per game, and the difference is quite striking. The chance of winning a game increases significantly with every additional 5 attempts made per game. Even in total rushing yardage, the success rate goes up as the yardage goes up.

If we look at this with some hypothetical numbers, things become a little more obvious. Let’s say Team A runs for 100 yards in a game on 25 carries. The average yards per attempt would then be 4.0. Good production all around, I would say. (While each statistic individually gives about a 50% chance of winning per the charts, please note that we haven’t looked at the success rate given statistical combinations or cross-references.)

Now, let’s say Team B also runs for 100 yards in a game, but they only do it on 10 carries. That’s 10.0 YPA, but can an offensive be sustained for an entire game off of 10 carries? Also, the numbers could suggest more of a boom-and-bust running performance. Big plays are good, but if they can’t be sustained with longer drives, it gives the other team more time to score.

Similarly, if Team C matches the YPA of Team A with 4.0, but they only do it on 15 carries, that’s just 60 yards for the entire game. If you don’t have a big passing attack to balance out the offense, it could make it harder to win.

Which brings me to my next point. You’ll notice that the Green Bay Packers’ winning percentages are almost all higher than the league totals. This can be attributed to the fact that, in the past ten years, the Packers have had some of the best passing offenses in the league with Brett Favre and Aaron Rodgers. They don’t need as strong of a running game when they can make things happen through the air.

For example, in the 80-99 total yards category, the Packers (68.60%) are well above the league winning percentage (39.36%). Other teams in the past ten years to be significantly higher in this category are the New England Patriots (75.80%) and Indianapolis Colts (57.60%).

If we look at the attempts per game in the 20-24 range, the Packers (42.50%) are almost double the average (24.00%). Again, other teams significantly higher in this category are the Indianapolis Colts (58.10%), New England Patriots (57.10%), Philadelphia Eagles (39.00%), and New Orleans Saints (37.00%).

Notice anything yet? All of those other teams have boasted some pretty significant passing attacks.

Finally, we do have one more things to take into account. Football Outsiders devoted themselves to debunking the myth that a team has to “establish the run.” In their 2012 Almanac, they write:

. . . There is no correlation whatsoever between giving your running backs a lot of carries early in the game and winning the game. Just running the ball is not going to help a team score; it has to run successfully.

There are two reasons why nearly every beat writer and television analyst still repeats the tired old school mantra that “establishing the run” is the secret to winning football games. The first problems is confusing cause and effect. There are exceptions, but for the most part, winning teams have a lot of carries because their running backs are running out the clock at the end of wins, not because they are running wild early in games.

The second problem is history. . . Optimal strategies from 1974 are not optimal strategies for 2012.

Their conclusions are drawn in a 2003 article (their first ever) that looks at first quarter running production vs. fourth quarter running production, then comparing that to success during the season. My only concern with this approach is that it’s a little short-sighted and doesn’t have a big sample size. Is the sample size significant enough? Sure, but we could be looking at a bigger picture. Plus, they’re only looking at two quarters of a game, which doesn’t really speak to an overall running performance.

Obviously, there is more to winning a football game than just running the ball. Offenses have to have some semblance of balance in order to be effective. As Football Outsiders notes, simply giving carries to the running back is not a solution. Still, if you consider the effect it has on a pass rush against an elite quarterback like Aaron Rodgers, there is some substance to the strategical advantages.

And as I mentioned before, correlation does not equal causation. There are so many aspects to the football game – e.g., defense – that can make a much bigger difference in regard to the outcome. Yes, being able to run more yardage on more attempts correlates to a higher rate of success, but in order to get those attempts, you have to perform well in other areas.

Finally, the point from Football Outsiders about being able to run the clock out on those extra attempts to seal a game is about as important as the rest of it. Teams that can add those extra carries and still be productive to maintain a lead will generally be picking up more wins.

All that said . . . who do you think wins this debate? Jason Wilde or Bill and Homer?

——————

Chad Toporski, a Wisconsin native and current Pittsburgh resident, is a writer for AllGreenBayPackers.com. You can follow Chad on twitter at @ChadToporski

——————

  • Kevin

    Wilde wins. I think something that wasn’t mentioned that plays a big role for the Packers is play action, which I would think depends more on number of carries than yards per carry. Overall though, thes stats don’t lie. Hopefully McCarthy has the commitment to stick to the running game even when its not working (like against the Rams), and hopefully Green and the O-line can get something going

    • Mike

      Exactly! We don’t even need to look at the stats tables. In the Indy and Seattle games, I don’t believe we averaged a great ypc in either game, but in the halves where McCarthy did stick with the run, Green Bay outscored the opposition 33-10 (33-3 if we’re being real), but were outscored 34-6 in the halves he didn’t call runs – total of less than 10 in 30 minutes worth of football. Stick to the run! Like Kevin said, if teams are even stepping to stop the run, that split second hesitation helps the play action game immensely!

  • G

    I think this is a secondary outcome. Teams in the lead tend to run more late in the game to burn clock. In a sense, winning causes increased rushing attempts, rather than the other way around.

    Brian Burke does some great analysis of this at Advanced NFL Stats.

    • Chad Toporski

      I’ll have to check his stuff out. Thanks for suggesting it!

  • Elle

    This is really interesting, thanks for the article.

    But G is correct.

    From a statistical standpoint, I would like to see the correlation between a high median YPC vs a high average YPC. Because a lot of the higher YPC averages are bolstered by 40 yard runs and such. Those are big plays, but one 40 yard run and a bunch of 2 yard runs that add up to 80-100 yards in 20 attempts is not going to be as effective or as useful, in my opinion, as a running back that literally rushes for 4-5 yards on every attempt.

    • Chad Toporski

      Not sure I could gather those types of stats, but I think it would be great to know.

      • Elle

        Yes I figure that would be incredibly difficult to garner, which is why I don’t fault you for it. Still a great article.

        • G

          Check out “success rate”

  • Tarynfor12

    I agree with Wilde however a poultry 1.5 per carry average isn’t enough unless your passing is a consistant unstoppable dynamic.
    MM is interested in keeping defenses honest but the level of honesty is suspect with averages of less than 2 yds per carry.
    I’m not one that believes the Packers need to run the ball 45% especially since we don’t have the OL or the RBs capable of success with that formula but a solid 20 and if the game allows a 25 touch game with at least a 3 yd average..extra icing for this offense.

    • steve cheez

      “a poultry 1.5 per carry”- are you saying our RBs are turkeys?

      • Tarynfor12

        LOL…didn’t even realize it.

  • Lucas

    I would love to see this work done with passing. While we’re (you’re) at it, what about offensive attempts combined?
    MM speaks of increasing the number of chances/plays from scrimmage by playing fast so as to increase the number of possessions in a game. Is he on to something? Great work as always, Chad.

    • Chad Toporski

      I’ll add the passing game to my list!

      Thanks, Lucas.

      • steve cheez

        No good deed goes unpunished!

  • http://googlechrome madtowne

    Good article…Wilde wins…Homer likes
    to stir things up for better ratings!

    The running game is an important part
    of the offense, but it’s more than the
    halfback. It’s also a quick hitter to
    the fullback, a quarterback scramble,
    a sneak, a bootleg, an end around, a
    reverse, and a lateral. While all are
    classified under the running game, these
    players must be accounted for by the
    defense. Green, Kuhn, Rogers, and Cobb.
    Tough guys to defend in MM’s system!

    • Chad Toporski

      I think this is an excellent point. The use of Cobb, Kuhn, and even Rodgers are part and parcel of the run game.

      (To note: The statistics used above are not limited to running back production.)

  • Dan

    I think all of this is a moot point if a team has a good to great O-line that can both run block and pass block at a consistently high level. If its a pass, block the guys in front of you and give your QB time, if its a run, open a hole for your back. The problem is the Packers do not have a consistently good O-line. Our QB is hit way to much and our RBs don’t have much room.

    I’m not saying that with a good O-line it should be ok to pass 90% of the time, but, with a good,productive O-line, balance percentages can be tilted either way, run or pass.

    • Chad Toporski

      And I think that shows with the ability of teams like the Packers, Patriots, and Colts in the past 10 years to win with less of a run game.

      Offensive “balance” doesn’t mean a 50/50 split, but it does necessitate the use of both the run and the pass at a high enough level based on the talent at each position group.

  • Greg

    I tend to agree with Wilde. Bill and Homer are kind of using backwards logic. You can’t say a defense can just forget about the run because they are holding the opponent to low ypc. What if they achieved that low ypc soley BECAUSE they were focusing on the run? You can’t treat these things like they’re in a vacuum. That’s like saying I don’t have to take out the garbage anymore because my house isn’t currently overflowing with garbage (because I’ve been taking out the garbage). Bill & homer are getting tunnelvision in the stats and forgetting about what causes the stats.

    Now, I’m sure defenses aren’t selling out to stop Green and rushing production could definitely improve, but I do think rush attempts can serve a purpose, even if they don’t go for a ton of yards. Predictability leads to Rodgers getting hit repeatedly and sacked 8 times. Ypc is also going to suffer if a leading team is trying to run out the clock, since the defense will often catch on and start playing the run. Variety is the spice of life.

  • David

    The differences is this..

    While Linemen and Linebackers have to rush the passer and stop a runner because they have to deal with both.. that isn’t true for DB’s. Play Action works primarily against DB’s… it’s when they are caught looking to see if Adrian Peterson has the ball that their coverage suffers or worse they blow the coverage completely and give up a huge pass play. So DB’s against the packers don’t have to play the run… they can sit back and worry about just the WR and TE. The reason this doesn’t kill the Packers like other teams is Aaron Rodgers.

    • Chad Toporski

      True, though I would also add that play action can buy some time for the routes to develop. It can help “freeze” the pass rush for a moment and also prevent linebackers from dropping immediately into coverage. This allows receivers to open up not only deep, but in the mid-ranges, as well.

    • Greg

      to add to chad’s point, DB’s aren’t the only ones that cover. LB’s are involved in coverage as well. as a former DE, I can attest that our defensive scheme & film study each week heavily emphasized opponent tendencies, no matter who the opposing skill players were. matchups play a part, but tendencies do as well

      • Mojo

        Greg makes a good point here. If we use an extreme example of a good passing team (say the Packers) who decide to abandon the run and go with the pass nearly all the time (lets say just 5 run attempts per game) any d-coordinator is going to recognize this tendency. I would expect the coordinator to go all out to stop the pass, telling his LB’s and CB’s to worry only about coverage . The pass rush would be constant and the Packers passing attack would suffer mightly.

        In a sense the Packers would be playing right into the defenses hands by continuing to pass only. If the Packers stuck to it, over time the effeciency of the passing game would go way down with picks and incompletions.It would be a disaster.As Greg said tendencies are extremely important when evaluating an opponent.

  • buckeyefanindaphils

    I think it also matters if you can run when you HAVE to. Like 2nd or 3rd/1-2 yds. or goal line. If you have success running on those downs, you have more success with play action and teams are more likely to sell out on the run. If you get 6-7 yrds on 1st down, it makes things much easier, or as Urban Meyer says “stay ahead of schedule”.

  • Jake B

    I agree with some previous commenters that the correlation between attempts and win % is slightly artificial…to even have the option to increase your rushing attempts you have to be getting 1st downs and so are more likely to be scoring.

    I’d look at the correlation between total rushing attempts in a game (I guess you have to do whole years but that’s quite dodgy imho…) and PASSING yards per attempt. Then, see if that correlation changes in strength if you look only withing teams/games with a high rushing YPC vs those with a low YPC

  • http://allgbp.com/2012/10/23/ Dman

    Do believe Holmgren use to say that they passed to set up the run

  • Scott

    That is correlation not causal. I can think of three reasons attempts is likely to be a result of winning, not a cause:

    1) A lot of rushing attempts occur after a team has already built a big lead and is running out the clock. They probably won’t get a lot of yards per carry then because the other team knows it is coming.

    That doesn’t mean rushing attempts does help keep a defense honest, I just don’t think those statistics carry a lot of weight.
    2) In converse, low rushing attempts happen if a team is really far behind late in the 2nd half. They’re going to abandon the run, thus resulting in less carries.

    3) Offenses that are moving the ball wind up in more situations where it makes sense to run (first down, 2nd or 3rd and short) whereas offenses that are struggling wind up in a lot of 3rd and long situations where teams tend not to rush.