Moving the (Stati)Sticks: Oregon State

Submitted by Adam Schnepp on September 17th, 2015 at 12:13 PM



Looking at last week’s data was an exercise in finding where the minute differences were; this week’s data is an exercise in finding where there aren’t massive ones. Michigan gave up a long, low play-count drive to begin the game and then shut down the Oregon State offense. It shows in the number of plays run (M: 76, OSU: 55) and in yards per play and scoring opportunities and so on, all the way up to the final score. Michigan turned into 2007 Ohio State in the second half and experienced a strikingly similar result.

Michigan’s domination of Oregon State did have a few bumps along the way. The first drive was as bad as it looked. Oregon State averaged 7.9 yards per play and had three explosive plays (two rushes and one pass) in a seven-play drive. The drive was anything but methodical; their success rate was just 42.9%, but the two long runs and one long pass counteracted their inability to stay in favorable down-and-distance situations. Michigan wasn’t bled to death, but was definitely bludgeoned. A low success rate proved to be foreshadowing Oregon State’s undoing as their explosiveness fizzled after the first drive, and that alone was enough to take them out of the game.

Oregon State followed up their big first drive with a two-play drive that ended in a turnover and a three and out that gave Michigan the ball at their 32 yard line, which is a pretty favorable place to start. OSU pulled things together momentarily, however, and had three eight-play drives through the rest of the first half. The problem with those three drives is that only one of them ended in Michigan territory, and that resulted in a turnover on downs after moving the ball 50 yards (6.25 YPP). The other two drives went 31 yards (3.87 YPP) and –17 yards (-2.125 YPP), with the negative yardage drive the result of the snap so high above the punter’s head that it almost hit the hanging camera.

[After THE JUMP: There’s a field position section so of course I used the crazy sky snap gif]

The second half was even worse for OSU. They had sixteen total plays for 9 yards in the second half, which means they were moving at a brisk 0.56 YPP. Michigan had a 14 play, 73-yard drive in the middle of the fourth quarter. I can throw a dart at the drive charts and I’m almost positive I’ll hit one that was longer and went for more yards than Oregon State’s entire second half output. A half yard per play. A half yard per play!

Michigan’s second half provided about as stark a contrast as you’ll find this season outside of the uniform maize and the on-field maize paint. Michigan had five explosive plays (four runs and one pass), which is especially noteworthy when framed by the zero explosive runs Michigan had at Utah. Their drives lasted 12, eight, five, three, 14, and two plays and resulted in 18 points. Oregon State’s second half drives lasted four, three, three, three, and three plays. They were methodical, but not in the way you’d like to be if you aren’t ripping off huge chunks of yards on pretty much every play.

Even more impressive than Michigan’s ability to sustain drives was their run/pass mix. At Wednesday’s press conference, Passing Coordinator Jedd Fisch said that an ideal offensive balance would be 50% run and 50% pass through three quarters before switching to 80% run and 20% pass in the fourth quarter. Michigan got a little ahead of themselves, running on 64% (16/25) of plays in the third quarter and 83.3% (20/24) of plays in the fourth quarter. The piece de resistance was the 14-play drive in the middle of the fourth quarter on which Michigan called nine straight runs, a third-down pass that went for twelve yards and converted, then four more runs that culminated in a touchdown.

If you were at the game you probably saw Oregon State’s mascot Benny the Beaver dancing near midfield before the game started. Nothing he did was inflammatory, but it wasn’t really entertaining, either. He was just sort of there, going through the motions such a person in his position is required to go through. You could say the same about Oregon State’s defense.


Last week I detailed Bill Connelly’s five factors. They are: explosiveness (35%), efficiency (25%), field position (15%), finishing drives (15%), and turnovers (10%). In parentheses are the weights he’s assigned to each factor’s importance in winning a game. We’ll use this as a sort of road map through the advanced box score, which also includes some basic information like number of plays and total yards.

Using yards per play as a proxy for explosiveness, Oregon State averaged 2.96 YPP and Michigan averaged 4.57 YPP. Getting outgained by more than 1.5 yards on a per-play basis goes a good way to making sure you lose, especially considering the weight this category carries in determining game outcomes.

Efficiency can be thought of as how often a team stayed in favorable down-and-distance situations, and Michigan was just alright at this against Oregon State. Their success rate was only 51%, but a moderate efficiency rate is tenable when your opponent’s is just 24%. Yeesh.

Field position is another category where Michigan lapped the field, with an average starting field position of their 37.8 yard line, compared with Oregon State’s average start at their 27.2 yard line. There’s a little bit of frippery here thanks to Oregon State’s sky snap.

Pro: That’s the closest anyone’s come to winning the stuffed animal you get for hitting the hanging camera. Con(s): This isn’t warm ups, bottom line is he didn’t get the stuffed animal, field position, general gif-iness, field position again.

The advanced box score uses number of drives, scoring opportunities (counted as a first down inside the opponent’s 40), and points per scoring opportunity as its measures of finishing drives. Michigan and Oregon State both had 12 drives, but Michigan doubled up OSU’s scoring opportunities (six to three) and averaged a solid 5.67 points per opportunity while Oregon State posted just 2.33 points per opportunity.

Michigan’s turnover margin was –1, but their expected turnover margin was –0.22. The actual difference was –0.78, but the effect was negligible in a game where they took four of the other five factors and in which those four weren’t really that close.


Moving from five factors to four, we turn to resident stat master The Mathlete to look at some cumulative statistics. The Mathlete uses four factors to evaluate teams. A refresher on what they mean:

Conversion rate = [1st Downs gained]/[1st Down plays (including first play of drive)]. A three and out is 0/1. A one play touchdown is 1/1. Two first downs and then a stop is 2/3, etc.

Bonus Yards = [Yards gained beyond the first down line]/[Total plays from scrimmage]

This is an adjustment to how I have previously calculated, to account for the plays a team runs.

Field Position = The expected point difference per game for where a team’s offense starts and where a team’s defense starts. Each drive is given an expected value based on the start of scrimmage, all of the drives for the offense and defense are totaled and compared. This accounts for all elements of field position: turnovers, special teams, drive penetration etc.

Red Zone: Points per red zone trip (TD’s counted as 7 regardless of PAT)

I think the best way to look at this is to use a table that tracks Michigan’s progress as the weeks go on. Keep in mind that the numbers in the table below are cumulative, not from that week’s game.


  Field Position Conv Rate Bonus Yards Red Zone
Week 1 21.0 73.0 1.52 5.7
Natl Rk 60 30 59 27
Week 2 25.0 68.0 1.63 5.8
Natl Rk 77 70 95 30

When I talked to The Mathlete this week he explained Field Position as the expected points an average team would score given the starting field position of each offense. I think Michigan’s Field Position going up is pretty intuitive given the field position stat we saw in Connelly’s advanced box score.

The increase in Bonus Yards dovetails nicely with the increase in explosive plays, though the national ranking takes a big hit. That’s not a huge surprise considering Michigan’s lack of explosiveness against Utah and that many of Michigan’s explosive plays against Oregon State got them three or four yards past the first down marker, particularly in the second half. The dip in Conversion Rate and subsequent slide down the national rankings, however, is worth keeping an eye on.


  Field Position Conv Rate Bonus Yards Red Zone
Week 1 27.9 73.0 1.64 5.7
Natl Rk 47 44 20 30
Week 2 25.1 67.0 1.50 6.1
Natl Rk 51 58 23 88

Combine the Oregon State stats with the Utah stats and the defense looks better in every area save Red Zone points, which makes sense considering Oregon State got there just once and scored a touchdown. Still, Michigan’s national rankings fell in every category; the small sample size and early season caveats both apply here. If UNLV is as advertised, these two tables could be very interesting next week.



September 17th, 2015 at 12:41 PM ^

Also, I find it odd that Connelly uses PPP as a proxy for explosiveness. Would seem that % of plays that exceed, say, 10 yards, would be the way to go (since a team that runs 2 plays of 0 yards and 1 play of 10 yards is different than a team that runs 3 plays of 3.3  yards, or some such variation). But I might not understand the tool very well. Nice write up.

Adam Schnepp

September 17th, 2015 at 12:59 PM ^

In his original post he has PPP, PPP+, and yards per play as ways to measure explosiveness. The advanced box scores are trimmed down a bit because he posts one for every game, so they just have YPP. You've got a good point, though.

And by frippery I just meant the starting field position number is a little gaudy because of one giant miscue. Also, I try to shoehorn that word in wherever because it's awesome.


September 17th, 2015 at 1:04 PM ^

In spite of the preseason Harbaugh hype, the team the team the team is flying under the national radar.  Zero votes in the rankings; 45 teams received at least a vote.  We are getting very little mention in the BTN and ESPN college football shows.

I think that is great, but with another probable rout this weekend and our only loss being by a TD at a (now) nationally ranked opponent, probably short-lived.



September 17th, 2015 at 1:45 PM ^

Utah game was close and decided by a pick-six.  Oregon State was one-sided.

It's not the job of statistics to surprise, so it's not an issue that they pretty much retell the story we all saw.  It's a productive exercise, but I can understand if some find it a bit dry.


September 18th, 2015 at 2:25 PM ^

The only thing that wasn't close about the Utah game was turnover margin:  3 important ones vs one that was as useful as knocking down the pass would have been.  I understand that had the scoreboard been different they might have opened up on us, but the fact remains that we outgained them.

I am not saying we should be ranked.  I am saying that the Utah loss is likely to look less bad as the season goes on and we push higher into national rankings in various offensive and defensive categories.  And I agree that assuming we don't stumble this week and BYU doesn't stink against UCLA, the game against BYU looks to be the key to how the program will be judged nationally in the early to mid season.

And I am also saying that Michigan would be favored against some of those 45 teams that received at least one vote.