rundown of Michigan's riser
Third Down Numbers and Other Stats
There's been plenty said on just how bad the third down defense has been, and I thought I'd chronicle that for you. For starters, our NCAA rank is currently #66 out of 120 FBS teams when it comes to overall defensive 3rd down conversion percentage (how often the opposing offense succeeds). We are listed at 38.78% with 38 conversions in 98 attempts.
From going over box scores, I found only 97, so note that discrepancy now. I'm not worried about one missing right now. Also worth noting, I used ESPN's box scores, not Brian's UFRs. So that may cause discrepancy if you go back and check plays there.
I'm not going to offer much more than interesting stats in this. I'll let you guys draw your own conclusions and leave them in the comments. Any thoughts or explanations are welcome.
So let's take a look at the different third down plays the defense has gone up against by yardage:
|Yards To Go||Conversions||Attempts||Percentage|
There's obviously a couple outliers out there. The 3rd and 18/24 plays against MSU and Iowa respectively definitely throw a wrench in the numbers. The number that is the most disturbing, though, has to the 3rd and 6 metric. Let's take a slightly closer look at that:
|WMU||3||6||pass||23||Fly play where a blanketing Warren dives and WR comes up with it|
|EMU||3||6||rush||13||Brown misreads zone read with running qb|
|EMU||3||6||pass||12||Umbrella coverage, missed tackle|
|EMU||3||6||rush||-4||2nd team scrubs were in|
|Indiana||3||6||rush||0||Rollout pass turned scramble for no gain.|
|Indiana||3||6||pass||18||3-man rush, as hit, throws skinny post against Mouton for 15 yards|
|MSU||3||6||pass||0||Stevie Brown Interception |
|MSU||3||6||pass||9||Crossing under routes confuses our LBs|
|MSU||3||6||pass||15||Woolfolk stares down QB in man coverage instead of WR. Misses route. Misses tackle to allow 1st|
|MSU||3||6||pass||0||Blitz house, man open but thrown wide|
|IOWA||3||6||pass||10||Curl short of the two guys we have deep on that side. Warren backed off presnap.|
|IOWA||3||6||pass||33||Pumpfake by Stanzi to a laid out Stross on a fly-ish route.|
Other than that pick and the four yard TFL against EMU by the scrubs, that's horrid. It doesn't seem to be laid squarely on blitzing too many, umbrella coverage, or anything in particular.
When you throw in those really long conversions, it looks pretty ugly. So what do you have to compare these numbers to? I've got two things. Brian did some extensive DIY Third Down Efficiency studies during the first few years of his blog, something he hopes to return to in the future, IIRC. There you can see that the normal conversion rate on a 3rd and 1 is ~68% (2007 statistics I believe). Michigan is outdoing that by about 7% on defense.
As you move down that trend line, however, you can see Michigan starts to approximate that line really quickly, then the extremely long conversions start to skew the results.
Also, we can look at how Michigan has done against opposing defenses.
|Yards To Go||Conversions||Attempts||Percentages|
As you can see, Michigan is doing much more poorly on offense when it comes to converting on third down. That said, we're also much better on converting on short yardage. When we get within 4 yards, we've got a very high percentage chance of converting.
Going back to the D for a minute, one of the other problems I'm noticing is how much worse we are on 1st and 2nd down. I'm not sure of too many metrics to gauge this, so I thought about a way to get a decent metric on this. While the standard 3 yards per play average will be fairly successful, it's probably not the best way to describe how successful you are. I decided to go with an arbitrary metric of half the distance needed instead. So, for example, if it's 1st and 10, 5 yards would be considered a successful pick up. So on a 2nd and 5, a 2.5 yard pick up would leave you with 3rd and 2 or 3. I would argue if you're able to do this, you'd probably be slightly more successful than just averaging three yards per snap.
I'll admit this metric is just my opinion, and I welcome ideas for a better way to measure success on 1st and 2nd down.
So with my metric in mind, here's the type of stats I'm seeing.
While Michigan does a decent job of stopping a team on 1st down, about 40.9%, second downs, Michigan is quite a bit worse on second down, around 53.8%. This is understandable as you generally need less yardage on 2nd down while still getting about the same number of yards. To explain, Michigan averages a 1st and 10.38 and gives up an average of 5.807 yards. Meanwhile, one second down, they average 2nd and 8.41 and give up an average of 5.629. The opposing team gains between 5-6 yards per play [ed. -cringe] on both first and second downs, while in my metric, they should need less.
I guess, if anything is good news, on third down, we face an average of 3rd and 6.56 and hold an average of 5.18 yards per play, over half a yard less per play than 1st or 2nd down.
I'll probably be playing with these stats a bit more in the next few days. Unfortunately, most of my stats don't involve personnel, so that complicates things.