How much do we tend to know after five weeks? Bill Connelly compares five-week data to full-season data to find out if we should be worried about TCU and Baylor.
07 Jan 2005
Guest Column by Jason McKinley
We've all heard the conventional wisdom spouted out by commentators in the media: "In order to have success in the passing game, you must establish the run." "To be a winning team, you need to be able to run the ball on offense and stop the run on defense." "Points come from the passing game!" It turns out that if the numbers are explored, all of these statements are true. And all of them are false. It just depends on which team you examine.
I started looking at play selection because I live in St. Louis and grew tired of hearing about how the Rams needed to run the ball more. Maybe they do, but what is the basis of that argument? Does the offense operate more efficiently when more runs are called? Does it limit the opposition's possessions and time of possession? How can this be determined?
I began by trying to answer these questions for only the Rams, but about 10 seconds later I realized that the Rams' numbers in isolation meant nothing. So, my spreadsheet grew as I added data for all 32 teams. The basic principle is this: Play selection (run/pass) for touchdown drives for each team is compared to that team's play selection in non-touchdown drives. Each team's defense was analyzed by the same method.
As a modifier, each team's touchdown percentage is also factored into the equation. If that modification is omitted, we may look at a team like the Chiefs that ran the ball 52% of the time in touchdown drives and only 38% of the time in non-touchdown drives and conclude that their passing offense stinks. It doesn't; they scored a lot of touchdowns regardless of how they moved the ball.
The final numbers in pass offense and run offense indicates a likelihood of scoring a touchdown as play selection drifts towards more passing calls and more running calls, respectively. These are each compared to the league average and assigned a percentage based on how much better (or worse) a team is at scoring and preventing touchdowns than other teams as runs or passes increase in a drive. The set up is similar to that familiar Football Outsiders statistic, DVOA (Defense-adjusted Value Over Average) in that positive numbers are good for offense and negative numbers are good for defense.
For instance, the average touchdown drive this season featured 4.1 passes and 3.1 runs. A drive similar to that could then be considered "balanced." As a team deviates from that percentage of pass and run plays, it becomes more of a pass or run drive. Five passes and two runs? Passing drive. Two passes and four runs? Running drive.
I chose touchdown percentage because when I first began looking at this early in the season, net touchdown percentage seemed to correlate pretty well with winning percentage. (Offense minus defense only; special teams touchdowns are not included at all and defensive scores only count as a successful stop for the defense and an unsuccessful drive for the offense.) The correlation fluctuated a bit during the season and currently stands as .823. (Correlation coefficient explained here.) In retrospect, the correlation may have been better had I included field goal drives as well. My main problem with not including field goal drives comes up in situations where all a team is trying for is a field goal. However, the way I measured drives (kneel-down drives don't count) resulted in 5831 total drives on the season. How many of these featured a "field goal only" mindset? I don't know, but I doubt that it affects the results significantly.
The consistency shown by most teams over the course of the season has been noteworthy. Most teams that were good at putting together running drives for touchdowns and not doing as well passing (Cowboys and Bengals, for example) in week three were still trending that way at season's end. It is this consistency that led me to continue compiling the numbers each week and to ultimately come here to share my findings.
Some of the results end up similar to DVOA. Others look a bit different, since I am looking at one very specific topic -- play selection's affect on touchdown percentage -- and am not dealing with other variables (such as strength of schedule). Most interesting to me are the teams that do well offensively or defensively on one type of drive but not the other.
With that, here are the numbers for all teams, followed by a little more detailed analysis of the twelve playoff participants. Keep in mind what the numbers mean: As a drive trends towards passing or running (represented by "pass offense" and "run offense" in the table below), the number given is the percent relative to the league average that the drive will result in a touchdown.
Perhaps the Steelers' run offense number is a bit surprising. They are one of the best running teams in football by pretty much any measure, but as the balance of plays in a drive shift more towards running, their chance of scoring a touchdown falls below that of the league average. I speculate that a lot of that has to do with the Steelers' unwavering conviction to running the ball. Way ahead in the game? Obviously, run the football. Behind in the game? Don't panic. Stay the course. Run the football. Unrest in the Middle East? There is only one solution. Run the football. It's probably tough to get those late game touchdowns when your strategy is so clear to the opposing defense. At least that's how I would rationalize the skew. The upside to this is how effective their relatively few passes have been at getting the ball in the end zone. If for some reason the offense needs a touchdown late, their passing game seems to give them a very legitimate chance.
Will that actually need to happen, though? The defense has been better at stopping teams from converting both running and passing drives for touchdowns than any other team in the postseason. (The Ravens fared better at both in the regular season.) It's hard for me to imagine a team this good losing another game, except that I feel I could say the exact same thing about at least two other teams moving forward.
Many times a look at what a team does the worst gives insight to where they might be best exploited. Good luck finding that on the Patriots, at least with analysis as elementary as mine. Just where do the Patriots struggle regarding scoring or defending touchdowns? Well, I guess it must be in defending run heavy drives. As teams run the ball on them more their chances of converting for a touchdown zooms up towards 18.5% below the league average. Yikes.
Perhaps playing a defensive scheme that forces the Patriots to rely on passing or running a disproportionate amount can help. The numbers say no. They seem to be even better offensively than defensively and there is no let up whether passing or running. This team is good. Really good. We all knew that already? Then let us move on to an even more obvious team, the Colts.
Here's something interesting. As you watch the Colts play this Sunday keep this in mind: If you see the Colts run a pass play in a drive, they have been 162% more likely than the other 31 teams in the league to finish that drive with a touchdown. Oh, if you see them execute a running play, they have only been 132% more likely than the other 31 teams in football to finish the drive with a touchdown. If they have to move the ball on the ground during a windy outdoor game, these numbers say that they will be able to pull it off and even flourish.
Defensively the Colts seem to do all right, only slightly below average, as teams pass more. They have had significantly more trouble keeping teams out of the end zone when more runs are involved. In my initial, not at all reader-friendly version of the numbers, I would do things in terms of standard deviations to try to bring clarity. In those terms, the Colt defense is more than a standard deviation below the league average in terms of allowing touchdowns as teams run the ball more. Can this be exploited in a game where the Colts' offense also gets slowed? If so, we'll get another year of "Manning can't win the big one" rhetoric. Yippee.
This is a very well-rounded team. Like the Patriots, they are better than the league averages at everything I measure. What has really set them apart as far as my analysis goes is their offense. They are second in passing (Colts) and third in rushing (Chiefs, Colts). When Aaron predicted an upswing in the Chargers' offensive fortunes, I'm not sure he had that kind of touchdown productivity in mind.
This is as good a moment as any to mention a limitation of the numbers in this article. I do not in any way adjust for the strength of opponent. One quarter of the Chargers' games came against the Chiefs and Raiders, which grade out as the two worst defenses in the league in terms of touchdown percentage allowed. Did those games unduly inflate the Charger offensive numbers? Probably. However, my intentions are simply to show how play selection affects different teams' touchdown percentages. For a team like the Chargers, they seem to be successful regardless of play calling.
When the Jets run the ball, they seem remarkably average at getting the ball in the end zone. When they pass the ball a lot, they somehow get even more average. Of course, these numbers have been steadily declining since Pennington's injury, which is a concern as they play Saturday.
The defense has generally done a good job of keeping touchdowns off the board, but has been especially good at keeping teams from putting together a sustained ground drive. They've been nearly as good as the Steelers at keeping rushes from adding up to six points. A handful of teams have been able to work on the Jets through the air this season. Will the Chargers be able to replicate some of that success? If the Jets can secure defensive air supremacy, it should be a fun game.
This is a nice offense for scoring touchdowns regardless of play selection. Not great, but nice. Even if Plummer has a deadly, win-killing "Bad Jake" day, he'll still usually try to atone with a meaningless touchdown or two. But, if the Broncos fear "Bad Jake" they can always go to mostly a ground game without losing much.
The Broncos do have something seemingly going for them on defense when playing a good passing team like the Colts. They rate seventh in the league in keeping pass heavy drives out of the end zone (27% better than the league average). But, they are below average at stopping run heavy drives (7% worse than the league average). As noted above, the Colts are fantastic when running, so this may end up not making much of a difference.
If it's true that defense wins championships, then the Philadelphia Eagles are the only legitimate hope for the NFC. Even after mailing in the last two games, they still boast some impressive numbers across the board both offensively and defensively. They were actually showing better than the Patriots in the numbers I measure until the last two games were played. Now we'll get to see if they can overcome the long layoff and the loss of Terrell Owens. (He was a wide receiver for the Eagles. You may not have noticed him as he quietly went about his business, but he actually had a terrific year for them before getting injured in Week 15.)
Really, before clinching home field throughout the playoffs, the Eagles looked absolutely unstoppable according to the data I came up with. They still look great even after factoring in the last two games, but can they just automatically turn it on come January 16? Frankly, with the relative dearth of talent in the NFC, I wouldn't be surprised if they can.
I'll admit it. I don't understand how this team is 11-5. I've seen them play a few times, sometimes winning, sometimes losing, never looking particularly great in any way. Yet, here they are. Here's the Falcons summary: They don't score too many touchdowns as they call pass plays, they allow touchdowns at a fairly high rate when passed on and at an even higher rate when run on. They are, however, the eighth best in the league at scoring touchdowns as they call more runs (12.2% better than average).
Is rushing offense (as measured in this exercise) that good of an indicator of wins? Not this season. Running your way to touchdowns correlated with winning percentage to the tune of .477 this season. By the way, stopping running drives correlated with winning percentage at only .277. When measuring success in the way that I describe in this article, a team was much better off this season if they were successful at passing and stopping the pass rather than the traditional "run and stop the run" approach.
The Packer offense has been great, passing or running. They have outperformed the league average no matter how their drives break down. They have the second best offensive touchdown percentage of the remaining NFC teams. It will need to carry them.
The offense appears to be the necessary catalyst to a Super Bowl run because their defense gives up touchdowns at an alarming rate. They are not good against the run (12.6% below league average) and they are flat out awful against the pass, 63.4% below the league average. It certainly makes for fun games for people that don't have a rooting interest, but I wonder what it's like to be a fan of a team that has such wild, up-and-down, high-scoring, emotional games. As a Rams' fan during the Mike Martz era, I wouldn't have any idea about that.
The Seahawk offense has performed better than average in terms of touchdown percentage regardless of passing or running. It's their defense that has a discrepancy. The defense when teams put together running drives is 7% below the league average, ranking them 15th. Against pass-heavy drives, they are 18.1% below average and rank 24th.
So, it would seem that the defense that was such an asset in the first three games of the season has turned into a liability for much of the rest of the season. That doesn't sound like good news with the Rams coming to town after coming back to life a bit offensively in the last couple of games. However, the Seahawks' offense is certainly good enough to put up touchdowns on the Rams if it turns into a high-scoring affair.
The Rams and the Steelers are the only teams in the postseason that have really significant offensive touchdown success differential between running and passing drives. For the Rams, they are 11th in touchdown percentage as they trend toward passing drives and 21st in touchdown percentage as they run more. Most commentators I hear seem to want desperately for the Rams to run the ball more, usually because they feel it will open up the passing game and keep their defense off the field. Neither appears to be true. The passing game is usually pretty solid right off the bat. It leads to more points. It moves the chains, which in turn limits possessions. Also, the Rams' defense in the 13 full Marc Bulger games has had to face only 10.77 drives per game, which would rank as the fourth fewest in the entire league.
Of course, there were those other three games. We all know what happened. It was all over the news. On Sunday, December 5, journeyman quarterback Chris Chandler descended on the Edward Jones Dome, heavily armed and with a list of demands. Among them was the requirement that he take over as quarterback of the team following the second drive of the 49er game. Eight quarters, two offensive touchdowns and ten (Ten! In eight quarters!) Chandler turnovers later, federal marshals escorted him off the field. Much of the damage was already done. The Rams had managed to beat the 49ers, but had been beaten up badly by the Panthers and Cardinals. After the Cardinal game, a stunned Martz had this to say about the events: "It's tragic for this football team that that (quarterback) position could hold this whole football team hostage." That about sums it up, Mike. In the Marc Bulger games, the Rams' passing drives hummed along at a clip 33% better than the league average. Their running efforts moved to a level just above the league average in terms of touchdown percentage.
Defensively, the Rams rank a less than pedestrian 19th in touchdown percentage allowed. Either running or passing will get the job done against them. If an opposing team has a particular strength at either, they should have success sticking with that. If they don't, then the Rams have a chance. Unfortunately for the Rams, there aren't any "lousy at everything" offenses in the playoffs, including the Seahawks.
From the values I measure, the Vikings have the second widest disparity in the league between offense and defense (Chiefs). The Vikings run less than any other team, but when they do run, the chances of a touchdown drive resulting is 43% better than for the other 31 teams. When they pass, they get even better; third best in the league at 88%. I have a crazy hunch that the Vikings might try to throw the ball on the Packers at on Sunday.
On the other hand, they have the worst defense in the playoffs and the 30th ranked defense in the league in terms of touchdown percentage. And it doesn't matter what your play call preference is. They are 59% worse than average against the passing drives and 57% worse than average against running drives. I don't think it's a fluke that the two previous Packer-Viking games have featured both teams scoring over thirty points in each game.
Jason McKinley spent several years as a molecular microbiologist and is currently working in social services while studying to become an actuary. If you think that's confusing, you should ask Craig Krenzel what he's doing with his life. You can contact him by e-mail at firstname.lastname@example.org.