Definitely Frankenstein. Even with all the speed in space in the last game Mason played FB/TE about 30% of all offensive snaps (per Brian's UFR) and they had 2 or fewer WRs on about 30% of all offensive snaps.
What is most promising from the past few games is that their "Frankenstein" approach isn't just a combination of the two within a single game-- they seem willing to adjust and emphasize whichever approach is most appropriate each game. After Notre Dame everyone was talking about how much the offense resembled last year's offense, and then after the MSU game everyone is now talking about speed in space.
Question about S&P+: where do people find 2019 S&P+ ratings now that Bill Connelly is at ESPN? There are no 2019 ratings on Football Outsiders or SBNation, and I can't seem to find them on ESPN's website either.
Haven't listened to the podcast yet so maybe I'm missing a crucial detail to the argument. Were Brian and Sam debating whether Al is a good playcaller or whether you should trust a mediocre playcaller to provide good analysis of play-calling? Because if it's the latter, then whether Brian would be a better playcaller than Al is immaterial in my opinion. To use an example from education (which also applies to coaching), being good at X doesn't mean you'll be good at teaching X (or coaching X). So if Al isn't even good at X (play-calling), why would we assume that he is a good teacher/analyst/coach of X (play-calling)? There is a lot more evidence that Brian is good at play-calling analysis than that Al is good a play-calling analysis.
Add to all of that the fact that MSU's non-conference schedule is actually pretty hard for a middle of the road Power 5 team, and it's hard to see them winning more than 7 games a year for the next 4-6 years. They've got BYU on the road, home-and-homes against Miami (YTM), Boise State, and Boston College, plus home games against WKU and WMU.
It really wouldn't be surprising to see them go .500 or worse the next 3 years:
2020: Miami (YTM), @ Iowa, Michigan, Ohio State, Minnesota, @ Penn State (plus lose-able games @ BYU and @ Indiana)
2021: @ Miami (YTM), @ Michigan, Penn State, @ Ohio State (plus lose-able vs. Nebraska, WKU, @ Purdue, and vs. Indiana)
2022: @ Boise State, Ohio State, Wisconsin, @ Penn State, Minnesota (plus lose-able games vs. WMU and @ Indiana)
I can't remember what game it was but in one of the defensive UFRs Brian suggested that star DTs is the low-key most important attribute of elite defenses. Mazi/Hinton has a much higher ceiling than Kemp/Dwumfour, so you might think that by the end of the year the DTs are an upgrade (whether Kemp or Dwumfour come back or not). With Hutchinson and Paye also back, any pass-rush contributions from the youngsters (Vilain, Ojabo, etc.) would just be a bonus. Problem is next year has Washington, Wisconsin, Penn State, MSU, and Minnesota in the first 7 games, so the DL is going to be tested early.
Deep bomb to Nico, bubble screen to DPJ, rinse, repeat.
But in all seriousness, unless the weather is bad use the pass to set up the run. Yeah yeah everyone says use the run to set up PA passes, but here if you hit Nico deep once or twice, and connect on a few successful screen passes, that will open up things for the run game.
I think which MSU we see will depend entirely on the first 3-4 possessions. If Michigan builds a 10+ point lead early, I could see MSU folding. But if we have a string of three and outs or turnovers and it's a tight game early, then the "win it for Mork" narrative could really take hold.
He’s taking negs because he spent his entire post negging Michigan. Give Dantonio and MSU whatever respekt you want— I have no doubt MSU will be up and ready for this game. Just no need to bash Harbaugh in the process.
That fits with my expectations as well, but I think it's possible they pour on a TD or two at the end to open it up to 45-13, like they've done against Illinois and ND the past few weeks.
The nice thing about the past 3 games though is they showed an ability to respond to adversity and have now scored TDs on 4 of the last 6 drives following an opponent TD (excluding the TD at the end of the ND game where Michigan's next drive milked 3:45 of clock until they took a knee). So if you count a 3:45 game-ending drive as a "win", Michigan accomplished what they wanted to accomplish (TD or end the game) on 5 of the last 7 drives following an opponent TD. The only exceptions are the first two PSU TDs.
That's a relatively small sample size, but it's a nice antidote to the BPONE that has pervaded Michigan football fandom for the past few years, where an opponent TD felt like the dam was about to break.
EDIT: Also wanted to add these numbers here for context. Michigan gained 108 yards on 11 plays on the two drives following ND TDs, 210 yards on 31 plays on the four drives following PSU TDs, and 79 yards on 15 plays on the three drives following Illinois TDs. That's almost 7 yards/play on drives following opponent TDs.
Re: Michigan's struggles implementing a coherent offense
Having so many guys banged up (Runyan, DPJ, Black, Charbonnet, Wilson, other OL whose names are escaping me, etc.) in preseason and the first few games must have hurt their ability to fully install the new offense. I can't remember a season in recent memory where there were so many guys on the two-deep who were banged up even before the season started.
I think they ride Dantonio until he retires, no matter the outcome. Who could they possibly hire amidst all the scandal and players leaving? Narduzzi barely has a winning record in the ACC-- he'd get destroyed in the Big Ten East. And they don't have the money, recruiting, or facilities to attract any of the top-5 (or event top-10) coaching options with no connection to MSU.
It's so hard to tell what's going to happen here, especially when you think about how much Michigan's performance changes over the course of a season. It's nice to see so many bullets in the chamber-- it feels like you just need to find 1-2 good players and 1-2 decent players out of Wagner/Johns/Castleton/Brooks/DDJ/Bajema/Nunez. The caveat being that Teske and Simpson aren't really capable of playing a secondary position so lineup flexibility is likely to be low, and "hits" on players at the 2-4 spots would be much more welcome than if Castleton, Brooks, or DDJ were the "hits".
I agree that having 3 wildcards and emphasizing strength of schedule would be a good fix. Last year the 3 wildcards would have been ND, Georgia, and Michigan (according to CFP rankings). In 2017, the wildcards would have been Alabama, Wisconsin, and Auburn. In 2016, the wildcards would have been Penn State, Michigan, and Wisconsin, although I doubt the CFP committee would have 4 teams from one conference in the playoff, so maybe you end up with USC or Florida State instead.
Interesting to think that in an 8-team playoff scenario Michigan would have been a frontrunner to get a wildcard with 2 losses in two of the past four seasons. Also interesting to think about some of the more surprising teams that would have made it (at least using the final CFP rankings from those seasons): Baylor, TCU, and Mississippi State in 2014, Iowa and Stanford in 2015, Penn State and Michigan in 2016, Wisconsin, Auburn, and USC in 2017, maybe UCF in 2018?
They tried that approach in the second half against Army with the same players, and it didn't work. The difference is they only called two run plays all game against Army whereas here they had a full complement of counters and counters to the counters that Notre Dame couldn't respond to. Oh and they actually ran it with the QB a few times.
If you're talking about Michigan's offense then I totally agree and hope it bodes well for this Saturday and the rest of the season. Need to figure out how to punch it in from the 3 yard line though. 7 points on 8 goal-line plays isn't great.
I know there were lots of mistakes and the defense could have played better, but I chalk a lot of these issues up to "good offenses find ways to make you pay no matter what". If you go back and look at the ~70 games in the past 5 years where a S&P+ top-10 offense played an S&P+ top-10 defense, the team with the top-10 offense scores 29 points on average. So when an elite offense goes up against an elite defense, the elite offense still puts points on the board. The fact that our offense only put up 7 points in the first half is way more concerning to me than the fact that Don Brown's defense got torched a few times.
Edit: I'll go a step further here. If you look at the ~120 games where an S&P+ top-25 offense went up against an S&P+ top-10 defense, the top-25 offense wins that battle most of the time. Top-25 offenses score at least 21 points 67% of the time they go up against an elite top-10 defense. Harbaugh's issue is not Don Brown, and it's not the defense. It's the offense. Harbaugh offenses have scored at least 21 points only 3 times in 10 tries.
Yeah you kind need a metric like "How upset would you be if your team lost to this team?" or a slightly less BPONE version like "How important is winning this game to your team's season?"
I would be happy with 2-2 against Penn State, Notre Dame, MSU, and OSU, as long as it also included wins over Maryland and Indiana. Losses to both MSU and OSU would hurt, but would also give us two wins over top-10 opponents which is almost as rare as wins over MSU/OSU.
I didn't watch the pregame and had the sound down on the announcers so it's very possible that I missed that. I certainly don't recall him out there. I can amend my comment to "Once Lavert Hill gets back I hope they don't go back to rotating evenly".
No offense to Vincent Gray, but why is he out there so often? Lavert Hill and Ambry Thomas make the occasional mistake, but it doesn't feel like they get picked on the same way that Gray does. Illinois' most consistent offensive play was to isolate Gray on the outside and throw a fade.
Agreed that the learning curve is longer and steeper than I would have hoped for. But to me this analysis shows that Michigan is successfully running some subset of spread plays 7-9 times per game with decent success. So I'm not sure why you couldn't just run those plays 12-14 times per game instead. If every week they made sure they were able to effectively run 1-2 RPOs, 1-2 arc read plays, 1-2 waggle plays, and the speed option play they've already used multiple times, they would still be able to utilize spread plays way more often than they currently are. I mean you don't want to use the same play too frequently, but you should be able to run it at least twice before the defense adjusts.
I mean those are possible, but I don't think likely. The spread stuff that we're talking about is basic stuff, and it's not like they're not running it at all, they're just not running it enough. And when you don't use that stuff when you can't move the ball and you're tied with Army in the 4th quarter, getting blasted by Wisconsin, or in a one score game with Iowa in the 4th quarter, what are you really saving that stuff for?
I also am skeptical that they aren't running it because it exposes a weakness. The numbers shown above suggest that they are actually quite good at running spread plays. What weakness are you exposing if you're getting 8+ yards/play on spread passes and getting 5.5+ yards/play on QB runs or RB carries that followed a QB run? Michigan's averaging almost 1.5 yards/play MORE on spread-style passes and QB runs than when they try to run non-spread stuff.
I really hope the answer is injury, fumbles, and transition costs led them to avoid spread stuff, but that healthier QB, slowly improving confidence in Shea not fumbling, and more overall comfort in the offense will lead to improvements over the next few weeks. Good news is Illinois' defense sucks so they have another week to fine tune the offense. Bad news is Penn State is the 15th ranked defense according to FPI and has held three (admittedly mediocre) Power 5 offenses in Pitt, Maryland, and Purdue to a combined 17 points.
Certainly seems like Illinois has faced a weaker schedule than Rutgers. Rutgers has faced BC (#46 offense according to FPI), Iowa (#50 offense), Maryland (#60 offense), and us (#93 offense-- woof). Illinois has faced Minnesota (#22 offense), Nebraska (#70 offense), EMU (#87 offense), UConn (#129), and Akron (#130). Minnesota is better than any offense Rutgers has faced, but I'd have to imagine Illinois's stats are boosted by playing 40% of their games against the two worst offenses in the country.
Completely agree with this take. Adapting to the skills of your players is an under-rated ability in coaches. For example, if Patterson won't pull on reads and you actually want him to pull, why not just call a designed QB run? Worst case scenario is you eat a 5 yard loss, but that's unlikely if the other team doesn't think Patterson will pull it.
I'm not sure I've ever seen such a discrepancy between the interior and outside DL. 297 lbs average on the interior, 245 lbs average on the outside. Give me some arc and stretch plays please.
Mostly a sample size thing at this point. Once you exclude penalties and goal-line plays, there are only ~290 offensive plays. That might sound like a lot, but when you start to separate out plays based on 2, 3, or 4 different variables, there are lots of bins I'm looking at that have only 10-15 plays in them. Probably need 2-3 more games before you could start thinking about significance testing.
Every 10 plays or so is consistent with how I looked at QB runs. Often enough to make the defense think about it every time there's a mesh point, but not too often.
That's a very good question, and a good point that Illinois might not be the team we want to emulate haha.
What percentage of the time do you think a team should be running RPOs?
It's not as easy to find play-calling data for other teams, and Brian almost never does a UFR for Ohio State, but I could at least look at the option play frequency for the past few Penn State teams.
Yeah I'm not sure which of those options it is either. I also think the combination of QB injuries and QB fumbles probably jointly reduced the overall number of QB runs. Transition costs could also be hurting the use of RPOs. One of the most disappointing things about the Iowa game to me was that I didn't see any of the new stuff they tested out in the Rutgers game. The waggle plays and stretch plays were reasonably effective in the Rutgers game (Rutgers caveat of course) but they didn't even really try them in the Iowa game (UFR only shows 1 waggle and 1 stretch play). It felt like that was Michigan trying to add spread elements to replace RPOs and read-options that attack the perimeter of the field, and then they ran almost none of that. So I have no idea what the Iowa gameplan was, or how it related to past gameplans.
What makes you say that they are not good? Collins is averaging 13.5 yards/target with a 68% catch rate. Even if you toss out fades, posts, and corner routes, he's still averaging 12.3 yards/target and an 82% catch rate on drags, ins, outs, and hitches.
Collins, Bell, Black, and DPJ as a group are averaging 9.3 yards/target with a 56% catch rate. Those numbers are elite and I can promise you there aren't more than 3 other programs in the country with a group of WRs with those numbers.
Why be worried about time of possession after a game we just won 10-3? Huddling is not the answer. Getting to the line quickly, then slowly reading the defense and adjusting play-calls based on the defensive formation (like attacking soft coverage with screens) like Ohio State does is the answer.
Wow. I haven't finished this yet but I count 36 minuses in coverage for the DBs and LBs. Can we have a game-plan that involves quick throws to Collins, Black, Bell, and DPJ that attack the edges of this old-school defense please?
Great stuff as always! I've been playing around with some analyses of the UFR data relative to national databases of WRs and yeah, passes to our "platoon of NFL WRs" is something we need to keep doing at a high rate. Like if you extrapolate our WR stats to a full season, the yards/target, catch rate, and success rate for those 4 players (Collins, Black, DPJ, and Bell) would put all of them in the top-40 for WRs in the country, with Collins and Black both top-10 WRs. That's an embarrassment of riches that needs to be taken full advantage of.
This is a fascinating take. I have no idea if this is actually true, but it strikes me as plausible that there might be other unintended legal consequences to allowing NIL rights. I am still in favor, but given the immensity of what this change would do to college sports, it's fair to expect lots of unintended consequences.
Question that's been bugging me and since I've now seen this figure a few times I'll go ahead and ask: What's the "COA" in the competitive imbalance figure? There's a vertical line and to the left of it says "Recruiting before COA adopted" and to the right it says "Recruiting after COA adopted".
Maybe someone smarter than me knows the answer to this, but if you are worried about Shea getting hurt, why not have him keep it 3-4 times per game and slide or run out of bounds the second someone got close? Teams are overplaying the RB so much I'd have to imagine that they would still be able to get 3-4 yards on almost every play.
I will take the under on that one for sure. Rutgers gave back some tickets (I think?) and the student section won't show up for a noon kick-off when the team is playing well.
Damn that is a lot of injuries on offense. It's only Week 5 of the season and almost half of the two deep has missed at least a half of football (Patterson, McCaffrey, Charbonnet, Wilson, Runyan, Stueber, McKeon, DPJ, Black). No excuses for the O, but that has to be hard to gel a new offensive system when almost half of your players are hurt.
The 2-4-5 that Wisconsin runs has OLBs that are 235 and 221 lbs, with the two DL 293 and 279 lbs. The ILBs are 224, 231, and 232 lbs. Army's 4-0-4 base defense ran out a front seven that was 280, 280, 275, and 255 lbs, with LBs that were 235, 235, and 240 lbs. The average front four and front seven are pretty comparable, even if Army has bigger guys and a NT they can trot out for obvious run downs.
Agreed with both of these takes. The most frustrating part of the past two games was the missed opportunities. And more generally, I think the solution to BPONE is having a functional offense that is fun to watch. It'd be nice to win some big games for sure, but it would be so much better if they were making big plays and it didn't feel like being down by 10 in the 3rd quarter was a death sentence.
Patterson and McCaffrey actually did a pretty good job on RPO reads. They were correct on 85% of RPO reads (according to FanNamedOzzy's diary posts), but only correct on 55% of read option run plays. Accuracy and timing were the bigger issues on RPOs. They were 2 for 5 on RPO passes, with at least two of those being errant throws.
I'd adjust this to say you should focus on stats when the game is still competitive. The MTSU game was a 2-score game until the very end of the 3rd quarter. The Army game was competitive the whole way. Plenty of teams manage to still have a successful offense even when they're up by two scores, and I'm not sure why you would crawl into a shell when you're up by 13 with 20+ minutes left in the game. Sure you burn clock by running the ball, but that only works if you're getting first downs and stringing together long drives.
Looks like I'm about to have a conversation with myself since the board seems to be focused on other (Wisconsin-focused) threads. To add to this, on zone read run plays Michigan averaged 3.7 yards/play on correct QB reads and 3.8 yards/play on wrong QB reads. And that's not counting the 3 doomed play calls where the QB had no good choices (which gained 0, 1, and -11 yards). Including the 3 doomed plays, Michigan only averaged 2.2 yards/play on these plays. This just reiterates the Neck Sharpies post in showing that Army had this play dialed in.
FanNamedOzzy posted a two-part diary breaking down the various option plays in the Army game. I believe there were 13 called RPOs (at least according to those diaries). They ran something like 76 plays, so about 17% of their plays were RPOs. You're right that RPO teams don't get to choose whether a play is a run or a pass, but Bill C's point is that running on standard downs is less successful because the defense knows it's coming, and presumably is set up to stop the run. If that's the case you would expect an RPO call to skew towards pass, since the defender being read is more likely to be in a position to stop the run.
Recent Comments
Definitely Frankenstein. Even with all the speed in space in the last game Mason played FB/TE about 30% of all offensive snaps (per Brian's UFR) and they had 2 or fewer WRs on about 30% of all offensive snaps.
What is most promising from the past few games is that their "Frankenstein" approach isn't just a combination of the two within a single game-- they seem willing to adjust and emphasize whichever approach is most appropriate each game. After Notre Dame everyone was talking about how much the offense resembled last year's offense, and then after the MSU game everyone is now talking about speed in space.
Question about S&P+: where do people find 2019 S&P+ ratings now that Bill Connelly is at ESPN? There are no 2019 ratings on Football Outsiders or SBNation, and I can't seem to find them on ESPN's website either.
Haven't listened to the podcast yet so maybe I'm missing a crucial detail to the argument. Were Brian and Sam debating whether Al is a good playcaller or whether you should trust a mediocre playcaller to provide good analysis of play-calling? Because if it's the latter, then whether Brian would be a better playcaller than Al is immaterial in my opinion. To use an example from education (which also applies to coaching), being good at X doesn't mean you'll be good at teaching X (or coaching X). So if Al isn't even good at X (play-calling), why would we assume that he is a good teacher/analyst/coach of X (play-calling)? There is a lot more evidence that Brian is good at play-calling analysis than that Al is good a play-calling analysis.
Add to all of that the fact that MSU's non-conference schedule is actually pretty hard for a middle of the road Power 5 team, and it's hard to see them winning more than 7 games a year for the next 4-6 years. They've got BYU on the road, home-and-homes against Miami (YTM), Boise State, and Boston College, plus home games against WKU and WMU.
It really wouldn't be surprising to see them go .500 or worse the next 3 years:
2020: Miami (YTM), @ Iowa, Michigan, Ohio State, Minnesota, @ Penn State (plus lose-able games @ BYU and @ Indiana)
2021: @ Miami (YTM), @ Michigan, Penn State, @ Ohio State (plus lose-able vs. Nebraska, WKU, @ Purdue, and vs. Indiana)
2022: @ Boise State, Ohio State, Wisconsin, @ Penn State, Minnesota (plus lose-able games vs. WMU and @ Indiana)
I can't remember what game it was but in one of the defensive UFRs Brian suggested that star DTs is the low-key most important attribute of elite defenses. Mazi/Hinton has a much higher ceiling than Kemp/Dwumfour, so you might think that by the end of the year the DTs are an upgrade (whether Kemp or Dwumfour come back or not). With Hutchinson and Paye also back, any pass-rush contributions from the youngsters (Vilain, Ojabo, etc.) would just be a bonus. Problem is next year has Washington, Wisconsin, Penn State, MSU, and Minnesota in the first 7 games, so the DL is going to be tested early.
Couldn't have been that blatant if they were able to get away with it for years?
Deep bomb to Nico, bubble screen to DPJ, rinse, repeat.
But in all seriousness, unless the weather is bad use the pass to set up the run. Yeah yeah everyone says use the run to set up PA passes, but here if you hit Nico deep once or twice, and connect on a few successful screen passes, that will open up things for the run game.
I think…
Re: your last point
I think which MSU we see will depend entirely on the first 3-4 possessions. If Michigan builds a 10+ point lead early, I could see MSU folding. But if we have a string of three and outs or turnovers and it's a tight game early, then the "win it for Mork" narrative could really take hold.
He’s taking negs because he spent his entire post negging Michigan. Give Dantonio and MSU whatever respekt you want— I have no doubt MSU will be up and ready for this game. Just no need to bash Harbaugh in the process.
Any list that has so many Michigan and Notre Dame games and doesn’t include the first Under the Lights game is a joke.
That fits with my expectations as well, but I think it's possible they pour on a TD or two at the end to open it up to 45-13, like they've done against Illinois and ND the past few weeks.
The nice thing about the past 3 games though is they showed an ability to respond to adversity and have now scored TDs on 4 of the last 6 drives following an opponent TD (excluding the TD at the end of the ND game where Michigan's next drive milked 3:45 of clock until they took a knee). So if you count a 3:45 game-ending drive as a "win", Michigan accomplished what they wanted to accomplish (TD or end the game) on 5 of the last 7 drives following an opponent TD. The only exceptions are the first two PSU TDs.
That's a relatively small sample size, but it's a nice antidote to the BPONE that has pervaded Michigan football fandom for the past few years, where an opponent TD felt like the dam was about to break.
EDIT: Also wanted to add these numbers here for context. Michigan gained 108 yards on 11 plays on the two drives following ND TDs, 210 yards on 31 plays on the four drives following PSU TDs, and 79 yards on 15 plays on the three drives following Illinois TDs. That's almost 7 yards/play on drives following opponent TDs.
Re: Michigan's struggles implementing a coherent offense
Having so many guys banged up (Runyan, DPJ, Black, Charbonnet, Wilson, other OL whose names are escaping me, etc.) in preseason and the first few games must have hurt their ability to fully install the new offense. I can't remember a season in recent memory where there were so many guys on the two-deep who were banged up even before the season started.
I think they ride Dantonio until he retires, no matter the outcome. Who could they possibly hire amidst all the scandal and players leaving? Narduzzi barely has a winning record in the ACC-- he'd get destroyed in the Big Ten East. And they don't have the money, recruiting, or facilities to attract any of the top-5 (or event top-10) coaching options with no connection to MSU.
It's so hard to tell what's going to happen here, especially when you think about how much Michigan's performance changes over the course of a season. It's nice to see so many bullets in the chamber-- it feels like you just need to find 1-2 good players and 1-2 decent players out of Wagner/Johns/Castleton/Brooks/DDJ/Bajema/Nunez. The caveat being that Teske and Simpson aren't really capable of playing a secondary position so lineup flexibility is likely to be low, and "hits" on players at the 2-4 spots would be much more welcome than if Castleton, Brooks, or DDJ were the "hits".
I agree that having 3 wildcards and emphasizing strength of schedule would be a good fix. Last year the 3 wildcards would have been ND, Georgia, and Michigan (according to CFP rankings). In 2017, the wildcards would have been Alabama, Wisconsin, and Auburn. In 2016, the wildcards would have been Penn State, Michigan, and Wisconsin, although I doubt the CFP committee would have 4 teams from one conference in the playoff, so maybe you end up with USC or Florida State instead.
Interesting to think that in an 8-team playoff scenario Michigan would have been a frontrunner to get a wildcard with 2 losses in two of the past four seasons. Also interesting to think about some of the more surprising teams that would have made it (at least using the final CFP rankings from those seasons): Baylor, TCU, and Mississippi State in 2014, Iowa and Stanford in 2015, Penn State and Michigan in 2016, Wisconsin, Auburn, and USC in 2017, maybe UCF in 2018?
They tried that approach in the second half against Army with the same players, and it didn't work. The difference is they only called two run plays all game against Army whereas here they had a full complement of counters and counters to the counters that Notre Dame couldn't respond to. Oh and they actually ran it with the QB a few times.
If you're talking about Michigan's offense then I totally agree and hope it bodes well for this Saturday and the rest of the season. Need to figure out how to punch it in from the 3 yard line though. 7 points on 8 goal-line plays isn't great.
I know there were lots of mistakes and the defense could have played better, but I chalk a lot of these issues up to "good offenses find ways to make you pay no matter what". If you go back and look at the ~70 games in the past 5 years where a S&P+ top-10 offense played an S&P+ top-10 defense, the team with the top-10 offense scores 29 points on average. So when an elite offense goes up against an elite defense, the elite offense still puts points on the board. The fact that our offense only put up 7 points in the first half is way more concerning to me than the fact that Don Brown's defense got torched a few times.
Edit: I'll go a step further here. If you look at the ~120 games where an S&P+ top-25 offense went up against an S&P+ top-10 defense, the top-25 offense wins that battle most of the time. Top-25 offenses score at least 21 points 67% of the time they go up against an elite top-10 defense. Harbaugh's issue is not Don Brown, and it's not the defense. It's the offense. Harbaugh offenses have scored at least 21 points only 3 times in 10 tries.
The graphics for that yellow one look like some high schooler made them.
Yeah you kind need a metric like "How upset would you be if your team lost to this team?" or a slightly less BPONE version like "How important is winning this game to your team's season?"
I would be happy with 2-2 against Penn State, Notre Dame, MSU, and OSU, as long as it also included wins over Maryland and Indiana. Losses to both MSU and OSU would hurt, but would also give us two wins over top-10 opponents which is almost as rare as wins over MSU/OSU.
I didn't watch the pregame and had the sound down on the announcers so it's very possible that I missed that. I certainly don't recall him out there. I can amend my comment to "Once Lavert Hill gets back I hope they don't go back to rotating evenly".
No offense to Vincent Gray, but why is he out there so often? Lavert Hill and Ambry Thomas make the occasional mistake, but it doesn't feel like they get picked on the same way that Gray does. Illinois' most consistent offensive play was to isolate Gray on the outside and throw a fade.
Agreed that the learning curve is longer and steeper than I would have hoped for. But to me this analysis shows that Michigan is successfully running some subset of spread plays 7-9 times per game with decent success. So I'm not sure why you couldn't just run those plays 12-14 times per game instead. If every week they made sure they were able to effectively run 1-2 RPOs, 1-2 arc read plays, 1-2 waggle plays, and the speed option play they've already used multiple times, they would still be able to utilize spread plays way more often than they currently are. I mean you don't want to use the same play too frequently, but you should be able to run it at least twice before the defense adjusts.
I mean those are possible, but I don't think likely. The spread stuff that we're talking about is basic stuff, and it's not like they're not running it at all, they're just not running it enough. And when you don't use that stuff when you can't move the ball and you're tied with Army in the 4th quarter, getting blasted by Wisconsin, or in a one score game with Iowa in the 4th quarter, what are you really saving that stuff for?
I also am skeptical that they aren't running it because it exposes a weakness. The numbers shown above suggest that they are actually quite good at running spread plays. What weakness are you exposing if you're getting 8+ yards/play on spread passes and getting 5.5+ yards/play on QB runs or RB carries that followed a QB run? Michigan's averaging almost 1.5 yards/play MORE on spread-style passes and QB runs than when they try to run non-spread stuff.
I really hope the answer is injury, fumbles, and transition costs led them to avoid spread stuff, but that healthier QB, slowly improving confidence in Shea not fumbling, and more overall comfort in the offense will lead to improvements over the next few weeks. Good news is Illinois' defense sucks so they have another week to fine tune the offense. Bad news is Penn State is the 15th ranked defense according to FPI and has held three (admittedly mediocre) Power 5 offenses in Pitt, Maryland, and Purdue to a combined 17 points.
That beard might be worse than the used car salesman look.
Certainly seems like Illinois has faced a weaker schedule than Rutgers. Rutgers has faced BC (#46 offense according to FPI), Iowa (#50 offense), Maryland (#60 offense), and us (#93 offense-- woof). Illinois has faced Minnesota (#22 offense), Nebraska (#70 offense), EMU (#87 offense), UConn (#129), and Akron (#130). Minnesota is better than any offense Rutgers has faced, but I'd have to imagine Illinois's stats are boosted by playing 40% of their games against the two worst offenses in the country.
Completely agree with this take. Adapting to the skills of your players is an under-rated ability in coaches. For example, if Patterson won't pull on reads and you actually want him to pull, why not just call a designed QB run? Worst case scenario is you eat a 5 yard loss, but that's unlikely if the other team doesn't think Patterson will pull it.
I'm not sure I've ever seen such a discrepancy between the interior and outside DL. 297 lbs average on the interior, 245 lbs average on the outside. Give me some arc and stretch plays please.
Mostly a sample size thing at this point. Once you exclude penalties and goal-line plays, there are only ~290 offensive plays. That might sound like a lot, but when you start to separate out plays based on 2, 3, or 4 different variables, there are lots of bins I'm looking at that have only 10-15 plays in them. Probably need 2-3 more games before you could start thinking about significance testing.
Every 10 plays or so is consistent with how I looked at QB runs. Often enough to make the defense think about it every time there's a mesh point, but not too often.
That's a very good question, and a good point that Illinois might not be the team we want to emulate haha.
What percentage of the time do you think a team should be running RPOs?
It's not as easy to find play-calling data for other teams, and Brian almost never does a UFR for Ohio State, but I could at least look at the option play frequency for the past few Penn State teams.
I hope you're right
You're so right! They don't have 2 TEs or a FB on the field so Harbaugh clearly has no input in the offensive game plan. /s
Yeah I'm not sure which of those options it is either. I also think the combination of QB injuries and QB fumbles probably jointly reduced the overall number of QB runs. Transition costs could also be hurting the use of RPOs. One of the most disappointing things about the Iowa game to me was that I didn't see any of the new stuff they tested out in the Rutgers game. The waggle plays and stretch plays were reasonably effective in the Rutgers game (Rutgers caveat of course) but they didn't even really try them in the Iowa game (UFR only shows 1 waggle and 1 stretch play). It felt like that was Michigan trying to add spread elements to replace RPOs and read-options that attack the perimeter of the field, and then they ran almost none of that. So I have no idea what the Iowa gameplan was, or how it related to past gameplans.
What makes you say that they are not good? Collins is averaging 13.5 yards/target with a 68% catch rate. Even if you toss out fades, posts, and corner routes, he's still averaging 12.3 yards/target and an 82% catch rate on drags, ins, outs, and hitches.
Collins, Bell, Black, and DPJ as a group are averaging 9.3 yards/target with a 56% catch rate. Those numbers are elite and I can promise you there aren't more than 3 other programs in the country with a group of WRs with those numbers.
Why be worried about time of possession after a game we just won 10-3? Huddling is not the answer. Getting to the line quickly, then slowly reading the defense and adjusting play-calls based on the defensive formation (like attacking soft coverage with screens) like Ohio State does is the answer.
Wow. I haven't finished this yet but I count 36 minuses in coverage for the DBs and LBs. Can we have a game-plan that involves quick throws to Collins, Black, Bell, and DPJ that attack the edges of this old-school defense please?
Great stuff as always! I've been playing around with some analyses of the UFR data relative to national databases of WRs and yeah, passes to our "platoon of NFL WRs" is something we need to keep doing at a high rate. Like if you extrapolate our WR stats to a full season, the yards/target, catch rate, and success rate for those 4 players (Collins, Black, DPJ, and Bell) would put all of them in the top-40 for WRs in the country, with Collins and Black both top-10 WRs. That's an embarrassment of riches that needs to be taken full advantage of.
This is a fascinating take. I have no idea if this is actually true, but it strikes me as plausible that there might be other unintended legal consequences to allowing NIL rights. I am still in favor, but given the immensity of what this change would do to college sports, it's fair to expect lots of unintended consequences.
Question that's been bugging me and since I've now seen this figure a few times I'll go ahead and ask: What's the "COA" in the competitive imbalance figure? There's a vertical line and to the left of it says "Recruiting before COA adopted" and to the right it says "Recruiting after COA adopted".
Maybe someone smarter than me knows the answer to this, but if you are worried about Shea getting hurt, why not have him keep it 3-4 times per game and slide or run out of bounds the second someone got close? Teams are overplaying the RB so much I'd have to imagine that they would still be able to get 3-4 yards on almost every play.
I will take the under on that one for sure. Rutgers gave back some tickets (I think?) and the student section won't show up for a noon kick-off when the team is playing well.
Damn that is a lot of injuries on offense. It's only Week 5 of the season and almost half of the two deep has missed at least a half of football (Patterson, McCaffrey, Charbonnet, Wilson, Runyan, Stueber, McKeon, DPJ, Black). No excuses for the O, but that has to be hard to gel a new offensive system when almost half of your players are hurt.
The 2-4-5 that Wisconsin runs has OLBs that are 235 and 221 lbs, with the two DL 293 and 279 lbs. The ILBs are 224, 231, and 232 lbs. Army's 4-0-4 base defense ran out a front seven that was 280, 280, 275, and 255 lbs, with LBs that were 235, 235, and 240 lbs. The average front four and front seven are pretty comparable, even if Army has bigger guys and a NT they can trot out for obvious run downs.
Agreed with both of these takes. The most frustrating part of the past two games was the missed opportunities. And more generally, I think the solution to BPONE is having a functional offense that is fun to watch. It'd be nice to win some big games for sure, but it would be so much better if they were making big plays and it didn't feel like being down by 10 in the 3rd quarter was a death sentence.
Patterson and McCaffrey actually did a pretty good job on RPO reads. They were correct on 85% of RPO reads (according to FanNamedOzzy's diary posts), but only correct on 55% of read option run plays. Accuracy and timing were the bigger issues on RPOs. They were 2 for 5 on RPO passes, with at least two of those being errant throws.
I'd adjust this to say you should focus on stats when the game is still competitive. The MTSU game was a 2-score game until the very end of the 3rd quarter. The Army game was competitive the whole way. Plenty of teams manage to still have a successful offense even when they're up by two scores, and I'm not sure why you would crawl into a shell when you're up by 13 with 20+ minutes left in the game. Sure you burn clock by running the ball, but that only works if you're getting first downs and stringing together long drives.
Looks like I'm about to have a conversation with myself since the board seems to be focused on other (Wisconsin-focused) threads. To add to this, on zone read run plays Michigan averaged 3.7 yards/play on correct QB reads and 3.8 yards/play on wrong QB reads. And that's not counting the 3 doomed play calls where the QB had no good choices (which gained 0, 1, and -11 yards). Including the 3 doomed plays, Michigan only averaged 2.2 yards/play on these plays. This just reiterates the Neck Sharpies post in showing that Army had this play dialed in.
FanNamedOzzy posted a two-part diary breaking down the various option plays in the Army game. I believe there were 13 called RPOs (at least according to those diaries). They ran something like 76 plays, so about 17% of their plays were RPOs. You're right that RPO teams don't get to choose whether a play is a run or a pass, but Bill C's point is that running on standard downs is less successful because the defense knows it's coming, and presumably is set up to stop the run. If that's the case you would expect an RPO call to skew towards pass, since the defender being read is more likely to be in a position to stop the run.