Ending drives with points, JJ McCarthy delivering on Harbaugh's happiest stat...there will be data and math
There have been multiple posts over the past few weeks tracking the draft status of JJ McCarthy against other QB's. This post is not one of those.
Throughout the season Jim Harbaugh praised JJ McCarthy for leading drives which put points on the board. While there are many factors which go into a teams ability to accomplish this goal (for example, do you have a Blake Corum, Michigan did!), I thought it was an interesting way to see how JJ performed and how other top QB's perform to get some form of reference.
Without much further ado, I'm going to get to the chart type things with all the data and math.
Here is how the top QB's performed this past season in the effort to end drives putting points on the board with a specific interest in TD's.
A note on how the numbers were compiled:
- I filtered out drives which were end of half run out the clock type drives because at that point the goal is not putting points on the board. This is only drives when the QB listed was leading the drive.
Summary:
- Jayden Daniels is clearly in his own class. Those numbers are incredible. You will soon see some caveats about that. He's still good, but there are a few warning signs.
- Overall, Harbaugh was right. Of the top QB's, JJ was the second best at ending drives scoring points. And he did a good job of making sure it was 7 points not 3 (second best at 78%).
- Even though the differences between the QB's don't seem big (except for Daniels) a small % difference adds up over the course of an NFL season. The latest stats I could find for NFL drives was an average of 10.4 per team per game. Over 18 games the 3% difference between JJ & Penix is 5.6 more drives ending in points on the season. In a league where so many games are close, that is a big impact and could easily translate to 2 or 3 more wins on the season.
To provide a reference point, the NFL average for scoring drives is 37.91% with TD's coming at 21.94%. It gets really hard at the next level to score.
To translate this info to what it meant in college:
You'll notice some of those QB's had 2+ more drives per game than JJ. Normalizing the data to 10 drives for each QB, you could expect them to average the following points/game:
Jayden Daniels: 44.2
JJ McCarthy: 34.9
Caleb Williams: 33.3
Michael Penix: 32.8
Drake Maye: 30.4 (I feel sorry for the desperate team who's going to draft this guy)
Moving on...
In college you have the opportunity to run up stats against garbage opponents, so I factored that in to see if it made a difference.
Note: I was pretty lenient on the definition of quality opponent. Basically, if it was a P5 team I called it quality. So yes, Rutgers, Vanderbilt, etc all counted.
- Now you can see the caveat about Jayden Daniels. While his numbers against real teams are still great. That variance is noteworthy. Against the bad teams on the schedule, Daniels had an amazing 21 drives end in 19 TD's and 1 FG.
- What really stood out was JJ performed the same against everyone. His numbers are not impacted by the opponent. A lot of that has to do with how Michigan called their games and again, JJ taking much of the second half off. While some may look at that and try to say he plays down to the competition, I think the opposite, the guy brings it every game and what he brings is good.
- For everyone with a hesitation about Caleb Williams, here's your chart. Again, it's a team game, but I struggle to see this guy performing well when all things are equal and harder, which is the epitome of the NFL.
Again translating this into what is the actual impact on a game. Here are the expected average points scored/game against good (better???) competition:
Jayden Daniels: 39.5 (- 4.7, points per game less than the total season data from above section)
JJ McCarthy: 34.3 (- 0.6)
Michael Penix: 31.4 (-1.4)
Caleb Williams: 31.0 (-2.3)
Drake Maye: 28.5 (-1.9)
Again, when the defenses got tougher, JJ kept performing. I really hope history remembers how good this guy was despite the individual stats not being like everyone else (except for 27-1, of course!)
My intent is not to put down any of these guys, but again, the warning flags on Jayden Daniels & Caleb Williams are standing out again. When these two played better competition, it resulted in almost a TD and FG less per game. (Daniels -4.7, Williams -2.3)
I was curious how this compared to last seasons top QB's drafted to get a better reference point.
Here's the chart:
JJ McCarthy was 2-0 vs. one of the guys on that list.
Here's what it looks like against "Good" competition:
Am I going to say JJ McCarthy will be NFL rookie of the year? That would be outrageous....but. HELLO!!!!!! The guy brings it when it matters most.
Another fun little thing which came about from doing this, is there were multiple QB's who played against Michigan. Would you like to see what a really good defense looks like??? I wanted to see.
There really isn't much analysis, insights, or notes needed for this. I really hope you enjoyed watching this defense the play past two years because they made some pretty good QB's plus Honda McCord look really bad.
TL;DR Summary: Harbaugh was right. JJ McCarthy puts points on the board, and he does it as well as anybody the past two years, and he does it when it matters most. Whoever drafts him will be very happy because I think JJ will translate very well to the NFL while there are some questions about the other QB's in this class being able to translate their game to the next level.
Great work - I always appreciate the effort that folks put into this.
Also, I got a kick out of the use of "tomato cans". I've obviously seen it before, and maybe I'm just loopy after a busy day of work, but it got a legitimate vocal laugh out of me.
Now you made me curious about the etymology of the term. If you can trust my googly research, "tomato can" was first popularized in boxing to describe an unworthy opponent; someone who was easy to pound and make bleed, like a cheap tomato can.
thank you. interesting.
who will be our QB next year? i know i'm not alone that i fear we will have a huge drop at the QB position and, thus, our overall performance. hopefully those fears will prove to be unfounded.
My hope is that we change our style. If we try to run the 2022-2023 offense without JJ, Corum, and a dozen upperclassmen NFL-level OL…we’re gonna have a bad time.
Many many other teams seem to be able to scheme an offense around players’ strengths. Nebraska ran the triple option, ffs. This is going to be a good year to gauge how adaptable Sherrone Moore is.
all true, but you will have someone with less of an arm and probably much less of a capacity to process what the D is doing in real-time. so you run? go back and forth from gap to zone? then pass? i think we have to be creative and mix it up more than prior years. its just that we're going from about a '9' on total offense to probably a 6.5-7.5 on offense. hard to make that up with just play calling.
I don't have a point of view on changing style, but if whatever style they choose works well and the defense performs anywhere near reasonable expectations, Sherrone Moore is going to have some serious momentum going for him.
He was adaptable enough to pound PSU 32 consecutive times to win the game. Who, outside of the service academies, would have the guts to do that.
March 14th, 2024 at 11:39 PM ^
Could Tuttle be Rudock 2.0? He seems more dynamic than Cade and sets the floor for the QB room.
March 15th, 2024 at 11:56 AM ^
Do we have to call him Tutle then? Title Tutle? Or maybe it's Tuttttle?
Great analysis. A good QB makes the entire offense better, not just the passing game.
The most interesting thing to me is HOW each of these QBs arrived at their stats. Very different styles represented. Penix is the pocket passer-iest pocket passer who ever passed from the pocket - if I’m a GM, I’m comping him to Stroud.
JJ was the field general for the most complex running scheme in the country. The Alabama game in particular was PhD-level football, and he had to call all of it. And look at the results: an effective offense. There were zero gimmicks used at Michigan this year, we earned every point.
Great data crunching! Thank you
Is the TD stat for Caleb Williams (86%) a misprint? The ordering of results, and the comment about JJ being #2 leads me to believe the proper TD % for Williams was 76%.
52% of Caleb William's drives ended in a scores (either a TD or FG). Of those 52%, 86% were TD's.
Williams finished a higher percentage of his scoring drives with TD's, but had a lower percentage of drives on which his team scored.
I did not watch a ton of USC games, but what I did see tells me Caleb Williams often relied on broken plays and outstanding skill position players to score TD's. What he was not good at was longer sustained drives which, if stopped, ended with the team in field goal range. This is one of the red flags I would have about him in the NFL. There are far fewer broken plays in which you can find a wide open receiver who can waltz into the end zone after the catch.
He will get his broken plays in his rookie season, may be all 17 games or just the first half. The NFL defenses will then tighten. His Offensive coaches will tighten, and he will lose his mojo.
This is the magic of the Lincoln Riley offense. What happened to USC this year is what happens to Riley teams across the trails of history. Just watch his Oklahoma teams - put them up against a playoff team in the playoffs and they promptly pooped their pants.
Looked at in the light of Day this signals good tidings making other programs green with envy. Being engaged is already making him huskier.
Great stuff!
Thanks for this. The main caution I would bring is that we're using "JJ McCarthy" as a proxy for "JJ McCarthy, with Corum/Edwards/Mullings as backs, Loveland/Barner/Brederson as TE, Wilson/Johnson/Morgan as WR and an OL fit for the Zombie Apocalypse". One way to visualize that is to recall Cade McNamara in 2021 against OSU: his non-end-of-half drives went TD/INT/punt/punt/TD and then TD/TD/TD/TD.
The higher he gets drafted, the less likely he is to have that kind of high-level support. Ideal for him would be to go lower but to a team with a decent-to-good-OL that needs a QB to take the next step.
This is not meant to denigrate JJ in any way. It's merely a reminder of how difficult it is in football analysis to untangle all the synergies and dependencies, and how much it's a team game.
March 14th, 2024 at 11:47 PM ^
Agreed, scoring is ultimately about the whole team. I think I put that caveat in there somewhere. Even special teams and the defense have an impact on the likelihood of a drive ending in points.
One note I would make, the NFL has a lot of parity. The difference between a poor O-Line and good O-Line in the NFL is not like college where a bad O-Line is visibly worse to the untrained eye. Same with skill positions. The worst receiving core in the NFL is still at a level significantly above any college team. This is where the little percent better comes in. The NFL is a QB driven league. If that one guy can make a 3% difference along with everyone else, it can really show up in the W-L record of an NFL team.
March 15th, 2024 at 11:57 AM ^
So what you're saying is that OSU ran off the best quarterback they had at scoring points or touchdowns against Michigan.
I like that.
Yep. Kind of funny. Overall, he was about on par with Jalen Milroe, who Alabama is happy to have coming back. And he performed better against Michigan than CJ Stroud. Yet, they were fine with Stroud coming back and ran McCord out of town.
That is one crazy program they run down there.
March 15th, 2024 at 11:58 AM ^
Spreadsheets?
SPREADSHEETS!!!
March 15th, 2024 at 12:54 PM ^
One evening I told my wife I wanted to slip between the spreadsheets with her, apparently we had very different expectations for how the evening would end.
The JJ statistics raise an interesting point. Why did he not score more against tomato cans. And why does this seem to me to be a Michigan thing, predating Harbaugh. Certainly this was true under Lloydin the 90s, even with Brady.
I think there are multiple reasons. I'd have to really dig through some data to validate this, but these are my impressions:
- Michigan tends to use weaker opponents as a way to work on a part of their game they want to test or develop without revealing too much playbook. I think other teams will often rep core parts of their playbook against weaker opponents to perfect what they do best. Ryan Day has said this in interviews. I don't mind Michigan's philosophy on this although it can be frustrating watching other teams lighting up the score board while Michigan seems to underwhelm.
- While we may complain about the schedule, Michigan's cupcakes are typically not quite as horrible as other team's cupcakes. OSU typically has at least one team like Youngstown St. on their schedule. The SEC plays schools no one has even heard of. Michigan typically will play a MAC, or Mountain West or some other school like that. It makes a difference. Those types of programs have legit players sprinkled throughout their rosters and have real strength programs. Many of these other schools the SEC plays can barely field a team.
March 25th, 2024 at 11:37 PM ^
Agree with your first point but not so sure about the second. Sparty was just as horrible as anyone else’s cupcake. 😂😂😂
I think JJ would've easily duplicated JD if instead of going pro like most elite QBs do at 21 he got to spend next 2 years in that offense, with those receivers, and more importantly with that defense.
The most important factor to consider is JJ knew in plenty of games that the only way we lose is if he threw picks. Use Washington game as an example...once the game was 17-3 the game was over unless JJ turned the ball over. JJ was conditioned to check it down or 1 read tuck n run but do not try to fit into a tight window until you know it's game time like he did with Loveland 7 min left to ice the game. You don't make that throw in the 2nd quarter.
JJ was chasing a natty Jayden a Heisman 2 completely different mindsets.
The real question NFL team will ask themselves is how good will JJ be 3 years from now and how good were these other QB's at JJ's age?
My answer to your NFL question: JJ will be the best QB taken in this years draft when the various careers have ended.
I do not view this as a question worthy of debate.
Great data and presentation. Thanks for putting this together.
Another thing that Harbaugh also stressed, and JJ excelled at (except for the BGSU game), was avoiding drive killing turnovers. That's another thing that sets him apart from Williams (and I think Daniels).
Dude has really put it all together. I might have to start paying more attention to the NFL now.
Comments