Stop Managing Down & Distance, Part 2: A Wisconsin Case Study

That's sort of a given, isn't it?  I don't know a lot of teams in the past 30 years that had an RB who produced more offense than their QB.  

A more interesting comparison would be to see if, at elite schools, top WRs are now producing more yards/TDs/PAN than feature RBs.  

Mathlete: please don't take this the wrong way. I respect your methods of analysis and particularly your pure curiosity about the game quite a bit. I would like to know how much organized football you played as a kid. 

I have to think they use it as a constraint play. Someone (Seth?) was talking about constraint plays the other day, and you need them. I agree the game has changed, passing is huge, etc. But if the defense knows you're going to pass every down, they "pin their ears back" get after the QB, play nickel or dime, etc. What I'm seeing as your central argument here is that teams run the ball too much. What would be interesting is what percentage of plays should be running plays in order to maximize your offense. Pass plays are typically > than run plays, but you need the run to open up the pass and vice versa. So if teams are running too much, how much is the "right" amount according to the mathlete? That's what I'd like to see this data turn into.

06 WVU = Pat White (running threat), 07 Florida = Tebow (running threat)

I'm wondering if the "QB vs. RB" comparison is somewhat clouded by the emergence of the QB that serves as a primary running threat?  Both teams had RBs, but woe be the defense that failed to account for the QB running the ball. 

Also .. if I'm not mistaken, Borges is quoted as saying "You run for yards, you throw for miles" or something to that effect.  In other words, he gets the passing game angle to this.  But as others point out, there's a question of balance.

I think it's a glaring omission to discuss the relative value of Ball and Willson without discussing the basic structure of the offense - run-heavy, with lots of play-action. In other words, a large chunk of the passing plays, of Willson's production, were predicated on the threat of Ball running the ball.

Could an argument be made that a defense is less willing to get chewed to pieces with 5.5 yards-per-play metronomic consistency than shut down the passing game, which will, even with an extremely efficient QB, produce 0 yard gains about 35% of the time?

... if your constraint plays are making you efficient or you should just do the efficient thing more.  In other words, it's possible that if Wisconsin had shifted carries from Ball to Wilson, they would have been worse off.  This is a difficult hypothesis to test because it relies on so many interactions and we don't know how to weight them.

One game that seems instructive from the NFL is the Patriots loss to the Steelers (yes, I'm a Steelers fan).  The Patriots pass a lot and exceptionally well; while they have an efficient running game, it's mostly a change of pace, something to do to keep defenders thinking.

The Steelers essentially ignored the Patriots running attack, rushed five and played five defenders in short zones with one deep safety.  They designed their defense to take away the thing the Patriots did most often and succeeded.  (Obviously it helps having great players and having an offense that held the ball most of the day, but the last point is a chicken-and-egg issue.)

I should note that I don't doubt that many coaches, particularly in the Big Ten, are fairly conservative - coaches everywhere are, because the consequences for an anti-conventional tactic that fails are large (namely, getting fired).  It's easy to say that adopting this measure will mean 1% more wins in the long run but not so easy to do when most coaches won't see 100 games to get that extra win.  (It's worse when it's the 3 extra wins in year 6 with 2 extra losses first pattern ... you just won't see those wins often enough.)

I issue I see is you have more variables than you can control for.  If you're playing against a Greg Robinson coached defense (complete a Tony Gibson coached secondary), is the traditional runningback obselete?  

Heck yes it is.  Look at that cushion on the corners and how horribly trained the secondary is.  Commence bombing the defense and forget three years and a cloud of dust.  

Are you playing a top defense (Alabama, LSU, Ohio State, insert your choice here for 2011)?  

Suddenly a stable run game is key.  Couple of points:

• You mention offenses are more complex.  So are defenses.  There are times the D-Coord outsmarts the offensive coordinator.  At that point having the safety net of "3 yards and a cloud of dust" ball is very useful.  You can call your offense around that while you figure out how to adjust your passing game, etc.
• Repeated pounding of the rock will get the offense pumped up and remove the defense's will to live.  As a high school safety I dreaded going up against the teams with manchild offensive lines and knowing I was going to spend 4 quarters being run over by large RBs who might later test positive for steroids (two of them did).
• Regarding the point about, that kind of RB also increases the odds of DBs cheating into the box for run support and opening the door for the QB to do better.
• Look at Michigan vs Wisconsin 2010.  Standard run game vs bad defense = massive yardage gains and control over the clock.

I don't see any reason to declare managing down and distance as dead.  Teams that can average 4 yards a carry and exert their will in the red zone (dominate offensive line) force the defense to exert 100% of their effort to stopping the steady advancement.  This opens the door for defense mistakes, too many people in the box, and explosive players to rack up gaudy numbers.  What I see your work showing is more that defensive coordinators are more concerned over "death by a thousand papercuts" than "stopping the really explosive guy who might produce in spurts".  

My one question is regarding turnovers.  You state "Even at their best great running backs at similar value to decent quarterbacks" but I find that hard to believe.  A decent QB will turn the ball over much more than a great running back, hell, a decent QB will turn the ball over more than a decent running back.  I may have missed it in parts 1abc but I think its obvious that an int is much more detrimental to an offenses production than a no gain running play.  Sure RBs fumble on occasion but if we're talking "great running backs" there's no way they fumble more than a QB turns it over.  In 07 Henne had 9 ints compared to Hart's 2 fumbles, and that's a great QB compared to a great RB.  Wilson had an uncommonly low number of ints (4) last year which I think helps his stats significantly.  Obviously I haven't gone through the numbers like you have and it may be something that isn't as big an issue in the end but it was the one thing that jumped out at me so I thought I'd ask.

I like the metric, but you are basically rehashing Walsh's West Coast Offense.  Everyone shoudl run Andy Reid's offense, because that makes the best use of running backs in a era when QBs are more valueable?  Hmmmm.  My metric wil be called Freakish Athletes with Super Talent (FAST).  With this metric I will show that if you have really good players on both sides of the ball, you will win a lot.  Whereas, if your team is full or cones, you will likely lose.

Thank you again for the analysis. I love reading your posts.

This is a nice example that a QB (even paired against an elite running back) is almost always going to provide the biggest difference between a team winning and losing. 

The thing, in particular that CRex hit on, is the point is to play variance and the odds in a way that maximizes a teams chance of winning. Sometimes, it's a high variance strategy and sometimes grinding it out makes more sense. The game that immediately comes to mind is Florida vs Tennessee during Lane Kiffin's year as coach. Most people expected the Volunteers to be run out of the stadium, but the game was pretty close all the way through. Tennessee had a pretty good defense and a terrible offense, so they basically did nothing high variance except throw a couple of deep balls. If those passes hit, Tennessee had a great shot at the upset. They didn't, so Tennessee lost, but I think the strategy of slowing that game down made a lot of sense.

The offense, defense, special teams, and the opposition all have to be factored in to determine the best strategy for a given game. Wisconsin knew it could reliably hammer out yards on the ground, still score a bunch of points, and minimize the number of opportunities for their defense to get torched. In essense, they played to shield their weakest link (defense) while reliably scoring a bunch of points.

I feel like most coaches go into games understanding they won't score every time and probably won't stop the other team every time, so they try to play the odds in a way that gives them the best shot to win at the end.

Having more than one rushing threat is certainly in vogue right now. It is neat that your numbers reflect and support this trend.

Two quibbles I would have with your data though. The numbers that show 199 QBs and 1 RB (or 97 and 3) don't so much suggest to me anything about the QBs or RBs themselves but instead suggests to me that the system is greatly biased towards QBs. The second point offers some explanation for the first. Your system demonstrates Wilson as being much more important than Ball. As you mention, although not strongly enough in my mind, the offense is structured for Ball to be the workhorse and Wilson to be the glory boy--your measure of 'plays over 7 yards' is a perfect example. Wilson gets those 7 yard  plays because Ball does what he did. You can't easily seperate the two. Wisconsin is a poor example of this because Wilson is very good but I'm sure somewhere in those 199 QBs that are more important than every RB but one, you will find a mediocre QB completing easy passes that his RB set up for him.

The passing game is king in football right now--partially because of rule changes--but one glance at the two teams that played for the crystal football will tell you that the running game is not dead.