Soliciting your takes on Net Success Rate vs. PSU
Brian has been posting this chart from Parker Fleming @ cfb-graphs.com here and there throughout the year. After several games, Michigan was at the top of the chart.
Against PSU, the chart suggests we underperformed. From what I can tell, offensive net success rate is whether you get 50% of the yards needed on 1st or 2nd down, or 100% of the yards needed on 3rd or 4th down. and Defensive Success Rate is whether you prevent the other team from getting that.
I did not watch the game. I listened on the satellite radio and had the PSU home announcers. the vibe was they got really excited about a lot of their defensive stops on our run plays, but then eventually we did score and wear them out. From the PSU home announcer broadcast, I did not have the sense that we DOMINATED them as has been discussed in the media. I did get the sense we did what was required to secure a W.
So I guess this chart for PSU would indicate that we failed to get 5 yards on 1st down a lot of times and JJ, Blake, and Donovan bailed us out on a few drives, but overall we punted a lot.
It doesn't really matter for PSU, but makes me wonder how this will play out in THE GAME in 2 weeks. Does OSU try to shut down the pass like PSU did and make us rely on the run? But then also do they bring a capable offense that can score 20-30 points vs. our run game that might score 20-25?
Anyway, just soliciting your thoughts on this.
November 14th, 2023 at 11:52 AM ^
When you say within 1 score, you mean that they were down by 8 and needed a TD and a 2-point conversion just to tie.
Michigan's biggest danger was getting JJ mugged on a pass drop and giving up a TO that either gave PSU great field position or was returned for a score.
Seriously. That was PSU's best chance to win. Michigan refused to give it to them.
After Michigan scored to go up 14-3, Penn State had exactly one possession where they could, at any point, take a lead; it was the opening possession of the second half. Michigan recovered a fumble, wisely drove into FG range and bled clock, re-asserted an 8-point lead, and then kept the game away from Penn State like the tall dad holding a ball over his head to keep it away from his 8-year-old: might not be very entertaining, and the kid may do a lot more, but the only way that child gets the ball is if dad lets them.
November 14th, 2023 at 12:19 PM ^
This is the correct take. OP: I suggest you listen to the podcast, if you haven't. That might help you understand this chart.
Basically, (a) Stephen is correct, that Michigan wasn't particularly trying for a success rate win, and (b) Michigan spent much of the first half playing what Brian called their "Ohio State" defense -- the one you'd use against McCord and (more to the point) Harrison, Jr. That left them open to quite a few 5-yard rushes on first down. The game plan was basically "you won't be able to do that on every play, so we'll let you try it on plays where it doesn't matter." PSU failed to convert a lot of third and shorts, and even their non-garbage-time touchdown drive was aided by two fourth-down conversions. Give them credit for making the plays, but literally everything had to go right.
The reason that the fancystats look the way that they do is that Michigan played Lloydball in the second half. "Welp, we've got a lead now, may as well run out the clock." This was partially to prove a point, I think, but also it was exactly this ^^ -- why take risks when PSU needs a ton of low-probability events just to try to catch up? PSU was unlikely to be able to score from their 15. From the Michigan 40, though, they'd have had a chance.
November 14th, 2023 at 1:35 PM ^
Michigan wasn't particularly trying for a success rate win
You can both run the clock down while also trying to move the ball. Michigan ran a ton of plays that were going to top out at 3 yards if everything went perfectly. Things don't go perfectly so they got 1 yard instead. This idea that Michigan wasn't trying to be successful is either an indictment of the coaches (we only have certain plays for 50 yards or we're out of ideas) or just plain dumb.
why take risks when PSU needs a ton of low-probability events just to try to catch up?
Again, you can avoid passing the ball and putting JJ at risk while also not running straight into boxes full of PSU dudes running hair on fire at the RBs, overwhelming the OL, and tackling the RB at 1 yard. Last year against PSU Michigan threw the ball more (which again, didn't need to do that here), but the running game was constantly putting PSU in a blender. They delivered Corum and Edwards to the Safeties at 7 yards untouched multiple times. In this game, they delivered Corum and Edwards to the LoS and then they met a LB and a Safety.
It's ok to just run, run, run, and run some more. But Michigan's 2nd half approach wasn't actually good at running the clock down. Their 3 drives that I've talked about ad nauseam averaged 4 plays and 2 1/2 minutes per drive. If they were looking to burn clock, maybe getting more than 1 first down would have removed a full PSU possession. Otherwise, just line up with 9 OL and lurch forward 2 yards at a time because that's the effective outcome they experienced anyways.
November 14th, 2023 at 12:30 PM ^
And a tie doesn't win you the game so had PSU scored, Michigan would have to mount a drive to get into scoring range, something that they didn't seem likely to do in the 2nd half until PSU (stupidly) went for it inside their own 30 and turned it over and then the defense had to not just get a stop, but push Michigan back a few yards to make the FG difficult. Which led to the DE dipping inside to get a TFL and Corum breaking it. If Michigan was capable of "turning it on" to put together a drive and score again, even just via the ground game, why sit on the ball and run into -1/-2 boxes for 2 YPC?
Regardless, the point is, I agree that protecting JJ from the pass rush should have been priority, but that doesn't mean you sit on the ball and run up the gut with no chance of success. I would have to rewatch every single offensive snap (maybe later) but Michigan wasn't bleeding clock in any meaningful way in the 2nd half. Again, after that FG they had 3 drives before PSU's TOD. 2 of which were 3 and out and bled a total of 4 minutes. That's not really smart play calling or impressive road grating. That's being Iowa, who we all ridicule.
Your last analogy isn't really accurate. Michigan wasn't holding the ball above PSU's head out of reach. It was more like giving the ball back to the 8 year old and assuming you won't slip he tries to run around you. Will you slip and pull a hammy? Probably not. But if you do... uh you're going to need to stand back up and score yourself to put that little punk in his place. You can be a little aggressive (again, not passing, just don't run for nothing) and still burn clock.
November 14th, 2023 at 2:34 PM ^
I agree. Think of OSU 1996; Shawn Springs falls so Tai Streets scores for I think only touchdown of the game for either team. This happened against us sometimes, which I will not name.
November 14th, 2023 at 11:25 AM ^
Michigan lit a bunch of downs on fire by running into -1 or even -2 boxes. They could have blocked the play perfectly, each OL and TE wins their match up and the run would still only go for 1 or 2 yards because Corum/Edwards isn't going to consistently make 1 or 2 guys miss in a phone booth. Plus then you had a PSU DL or 2 win his block or at least draw even mucking the whole thing up even more.
Again, not every play will go for all of the yards between the LoS and the goal line, but it's just a fact of Michigan football that they run plays that, at best, will get them 4 yards but average 2.689.
Look at the drives between the FG and Corum's TD. 12 total plays covering something like 40 yards. The offense last year against PSU didn't come close to what this approach was. Last year they were heavy on live reads and outside runs to get PSU in conflict and stretch their assignments. This year it was bully ball with a bunch of inside stuff.
November 14th, 2023 at 11:45 AM ^
Agree with most of the others. This chart basically reflects Michigan's strategy, which was contingent on the correct assumption that the risk of a game-changing TO on offense from a play like a strip-sack was greater than the risk of giving Drew Allar multiple chances to drive the field.
They took a 14-3 lead and then were never in a position where they couldn't just play to the score, and so they played to the score. Given the situation and the strength of PSU's team, it was a fine choice. It meant that the numbers got a bit gamed, especially with the late drive.
Michigan usually does fine on this chart. I strongly suspect things will look different for OSU, and even if it doesn't go how we planned, it will be for other reasons.
November 14th, 2023 at 11:51 AM ^
Net success rate doesn't include explosives. I like SP+ post game win expectancy as it captures all the variables. I think the Net Success rate is useful when talking about varying degrees of 100% win expectancy (as Michigan likely was for the first 9 games of the season). I haven't seen the post game win expectancy this year because I am no longer on TwitX.
Does anyone have that for the UM-PSU game?
November 14th, 2023 at 1:22 PM ^
IIRC we were 87%, first game not at 100% the entire season.
November 14th, 2023 at 12:01 PM ^
Many times fancystats do not tell the tale of the game honestly, just as the final scoreboard can twist the actual narrative of the game.
Michigan WAS dominating PSU throughout the game. Indeed, if that interception had stood, Michigan had another short field with a likely fg, and possible TD, which would have made the score 27-9 or 31-9. And, to top that off, PSU had to rely on the stupid unsportsmanlike penalty that only resulted from the interception being overturned, along with another penalty, all while Michigan was playing back to limit any big plays, to go down the field and score that last TD.
The game wasn’t nearly as close as the final score ended up.
November 14th, 2023 at 12:31 PM ^
I don't think PSU necessarily shut down the pass as much as Sherrone didn't want to risk a strip sack since we having so much trouble blocking the DEs (JJ's completion % was somethin like 88%). He could have called some more protections to give the tackles some help, but when you know PSU can't score, keep pounding the rock and don't ever give PSU good field position (seemed like kind of an Iowa game plan to me).
November 14th, 2023 at 12:37 PM ^
As some famous guys once said "You play to WIN THE GAME."
Once UM was up, there was no need to pass, since the UM D vs PSU O matchup was so much in our favor. If UM had to pass, I am sure they could have but why bring out plays when we can save them for OSU?
The funny thing is UM could have passe out of heavy two (release a TE or two) but again, no need. The game was a virtual lock once UM was up, they went heavy, and they ran the ball.
November 14th, 2023 at 1:00 PM ^
I think this is an interesting stat, but what is left out is the score differential. You are more likely to run into a stacked box, avoiding negative yardage, when you are up by more than 8 points against an offense that has a freshman QB that has had problems throwing it. The key was getting up 14-3, then 17-9 and then the strategy was "pain".
November 14th, 2023 at 4:32 PM ^
Success rate is hard to rack up when you attempt 8 passes all game and run the ball 32 consecutive times and your opponent knows you’re running it.
November 14th, 2023 at 5:22 PM ^
Yeah, the big thing here is that success rate only looks at success rate, not at how much above or below the threshold you are. It's one reason that SP+ from Bill Connelly also looks at points per play, which looks at explosiveness. Michigan was a little behind Penn State in overall success rate (though both had a lot of three and outs), but Michigan blew Penn State away in points generated, which shows up on the expected points added chart, pasted in from Gameonpaper.com
