It's kind of a macro level UFR... althought I have to say I would have rather you called the Rouchambeau verdict, the "Respect my Author-i-tay Metric"
Tales from the Dorkside: Rochambeaued Yourself
[EDIT: Updated to include sacks as failed pass attempts. Also corrected calculation error in success tallies. Text edited accordingly]
A basic premise of offensive football is that the single most important objective is to get to the next first down. A good offensive coach will Rochambeau his way to meeting this objective over and over again until either the scoreboard explodes or time expires. Coach Rodriguez has devised an offensive scheme that allows his teams to do just that and we have witnessed its power despite what we all witnessed last Saturday. But, was last Saturday’s performance due to poor execution, poor play selection, or was Penn State’s defense impenetrable?
First Things First
We need some definition for what a successful play is. Like others before me, I think it makes sense to characterize a play’s success based on the requirements at hand. A reasonable proposal is to use a Rolling Stones approach—get what you need. In essence, select plays that have a high probability to help you reach your next first down. Holding another basic premise that fourth down should be avoided the goal schedule by down is 33% on 1st down, 50% on 2nd down, and 100% on 3rd and 4th down; Regardless of the distance to go before the play starts. So, 1st and 10 requires 3 yards, 3rd and 1 requires 1 yard, 2 and 15 requires 8 yards, and so on. There are other ways to do this but, this is simple and fair. Sacks have been counted as failed pass attempts.
Turnovers are counted as a double fault, -1 for the unsuccessful play itself and -1 for the opportunity cost of losing the next play. I thought about an escalator here since presumably a turnover on 1st down is more costly than a turnover on third down, but there’s nothing to say that you won’t turn the ball on the next play, so a successful play only guarantees you the opportunity for one more play. A fumble is faulted regardless of who recovers since recovery is random and the mere opportunity to lose possession is not sexy at all.
Here’s what the breakdown going back to the beginning of the season looks like. Go Go Gadget Chart!
The Cartman column is the Rochambeau Verdict. Thumbs up means I think the play mix made sense as reconciled against the given success rates and thumbs down means I disagree with the play mix. I know, I know…I’m not a coach, but the data speaks for itself. I won’t spend a ton of space explaining everything else as hopefully the chart is fairly easy to understand. So for the reading impatient, I’ll roll straight into bullets.
- The running game is fine; even without David Molk.
- Between a talented freshman QB, a raw and athletic QB, and receivers who have lapses in concentration, the passing game is not ripe.
- A 50 % success ratio as described above seems to be adequate for offensive efficacy. When one phase of the offense dips too far below that Michigan needs to have the other step up or we look like the offense that we’re all trying to forget (2008).
- Going into the season, everyone “knew”* that we should expect Tate to play like a freshman a lot if not most of the time. He’s basically done that for four straight games. I’ll give him a pass on EMU because the running game was so damn dominant that it didn’t matter and he only passed 13 times (the fewest of the season).
- A 2-to-1 run to pass ratio is where it’s at. Again we “knew”* that we should depend on the running game going into the season, you know, because we were going to start a freshman QB and all.
- WMU/ND. Team came out on FIRE. All cylinders firing and no film for opponents to study. The low R/P ratio during ND was ok since both phases were getting it done. ND was a glimpse into the future.
- EMU. We were just better than them. Beat ‘em with one arm tied behind our back.
- INDIANA. First signs that Tate is indeed a true freshman. Passing success rate of 33%. Obviously, he pulled a rabbit out at the end but, uhh…a little close for comfort. Hence green meh-ey face.
- MICHIGAN STATE. They came in with a good game plan…at the expense of a bad game plan;*cough* WISCONSIN *cough*…Whew! ’Scuse me! The slowed down the run significantly and forced the pass offense to execute which they didn't do for most of the game.
- IOWA.Despite Tate’s struggles, the offense was effective and Michigan earned an opportunity to win. The running game was on point and the play mix was run heavy, which made sense. The game came down to a coaching catch 22 where only a victory bails you out but a loss leaves you with some ‘splaining to do. Hard to feel too pissed about losing that game…so long as you’ve finally gotten over the turnovers. I have. Hence, red meh-ey face.
- PENN STATE. Wuh-woah, Cartman wants to kick RichRod in the nuts. The passing game was not sharp, at all, but the R/P ratio was still pass heavy. At least for an instance where the pass game was not sharp and the running game was highly effective.
*scare quotes used because many of us seem to have forgotten that Forcier is a true freshman.
The problem last Saturday was that the scissors weren’t sharp, the paper was soggy, and the rock was…uh…not used enough. Zooming in a layer to the quarter by quarter break down of the Penn State game gives us this chart.
The play calling in the 3rd quarter was, unfortunately, too pass heavy. I suppose that was somewhat understandable being down 15 after the opening possession of the second half. But, there was a lot of time left to make up two scores and the running game was doing well and running is the scheme’s bread and butter and the passing game hadn’t been doing well and you have a true freshman QB. Does that mean Rich Rod sucks? Of course not. It just means that Rich Rod needs to hire a platoon of geeks to crunch data during TV time outs and half time.
In all seriousness though, it felt like we weren’t running the ball enough on Saturday and that’s what prompted me to dig into all this noise. I do have the benefit of hindsight, but even the data available at half time was enough to say that we should have come out running. Whatever, RichRod doesn’t need to explain himself to me and I don’t need or want him to. All coaches will have a lapse in judgment at one time or another and the players also need to be able to execute the plays that are called, regardless of play mix.
I must agree with you - it should be called that.
Interesting analysis overall. Interested to see how this plays out for the entire year.
Really interesting take -- thanks for the perspective.
As I understand it, your only beef is with the 3rd quarter, right? Yes, we started that quarter only down 9 pts, but by the end trailed by 22. With the way the offense kept imploding that 22 pts felt pretty daunting. Plus we averaged 2.8 yards per carry for the game. Sorry, but it's hard to see being too critical about the run/pass ratio on this one.
Brandon Minor and Carlos Brown averaged 4.8 ypc in the first half vs. PSU. For the whole game, all rushers averaged 4.1 ypc when sack yardage is stripped out.
On our first possession of the second half, down 15, we went 3 and out: incomplete, incomplete, sack. The defense forced PSU to punt on the next drive and then Denard fumbled. Then PSU scored to go up 22.
Not killing Rich over it, but Michigan should have called used Brandon and Carlos more coming out of half time.
since neither you nor I are invested in a strong defense or condemnation of the playcalling. However, I wonder what the YPC numbers look like after the first TD drive. After that, PSU seemed to have adjusted well and even in the 1st half our offense looked anemic both on the ground and in the air.
Nitpick: how do you get 4.1 ypc after sack yards are stripped out? In the box score I see 40 rushes for 151 yds gain, 41 yds loss. Sacks are 5 for 26 yds. Taking those out there are 35 carries for 151 - (41-26) = 136 yds. That's 3.9 YPC, not dramatically different than your figure of 4.1, but still worse. If you're excluding Tate as a rusher, you can't do that given the QB role in this offense.
I'm not in your face man, I just like discussing things.
Answer to your questions:
1. Minor/Brown first half ypc after TD drive: 5 ypc.
2. One of the "runs" was a TEAM rush when the ball was hiked out of the endzone. That was a two yard loss. Everyone who actually rushed the ball was counted.
The 5 YPC after TD drive does support your point.
What's the "all" column?
Success percentage for RUN and PASS combined.
Good work. However, I do disagree slightly with the "goal schedule by down." I think 50% of yards needed on first down would be an improvement. If you polled coaches and asked them if they would take 2nd and 7 or 1st and 10 most would take 1st and 10, IMO.
Thanks for the good work.
The goal schedule is a minimum, not a target. Of course 1st and 10 is better than 2nd and 7 but 2nd and 7 is better than 2nd and 8, 15, 20, and so on.I'm not arguing that a team should be happy with 2nd and 7 or only go after 3 yards on 1st down. Rather if you don't get at least 3 yards on 1st down, you're behind schedule.
If you get a pre-snap penalty or holding call on first down, you're starring 1st and 15 or 1st and 20. Gunning for 8 or 10 yards on first down is not necessary; going for 5 or 7 in that case is more reasonable. The generic goal of 33% is reasonable for this analysis, IME.
Good analysis. Looks like we have to mix it up a little more.
I am off to Champaign tomorrow night.
If you omit sacks because they are lapses in execution rather than play calling, why are you punishing turnovers double? Those are certainly execution problems not play-call problems.
You could take the view that if a defensive playcall is perfectly suited to what the offense will do, their guys will be in perfect position to meet the ball carrier immediately and strip it or have a coverage guy waiting to make a pick. But it would probably make this more accurate if each turnover was examined to make a rough guess of whether the D lucked into it or not.
although it might be complex to grade each sack, I've heard NFL QB's rate a coverage sack as > 4 seconds after the snap. Essentially if the defense can sack the QB before the routes are run, then it is failed execution on the OLine and you can't "grade" the success of the play.
You could argue that a coverage sack was a lack of execution on the receivers, BUT I think there should be a play call punishment if the defense is expecting pass, and covers it.
For example, if the QB scrambles for positive yardage, but does not meet the goal (a 5 yard scramble on 3rd and 10) then your system is counting it as a bad play call. Likewise if the QB throws it away, 0 yards, but no sack is counted as wrong playcall.
As to the turnover penalty, it is simply an opportunity cost thing. A sack doesn’t cost you the opportunity to gain a first down in and of its self whereas a turnover does.
As for sacks, after thinking about it some more I agree that they should be counted as a failed pass. When I wrote that it made sense to exclude them from being a failed pass but I can’t remember why and can’t construct a convincing and reconcilable argument to maintain the position; +1 for Jeff. I will update the charts and text later.
Best MgoChart ever. I know the analysis must be top rate--but the pictures and colors and stuff are all good enough for me.
And so what if I buy Playboy for the pictures?
there was a lot of time left to make up two scores and the running game was doing well.
In all seriousness though, it felt like we weren’t running the ball enough on Saturday.
Thank you, someone else said it finally.
Wasn't our opponents' gameplan earlier this season to limit the opportunities for Michigan's "quick-strike" offense?
So then what do we do on Saturday when faced with a team that shreds our secondary with long plays on short fields? Clang, clang, clang incompletions and go 3 and out.
Brown fumbling makes an "mostly-run" argument a bit less foolproof, but the run was more successful (as shown) and after halftime, you have plenty of time for long run drives and defensive stops.
Qualifier: I could only listen to WTKA and didn't watch.