Connelly's returning production data: round 1
Connelly has put out his first round of returning production numbers. These account for transfers to date. Of course, there's still the spring transfer portal window so these data will change again.
As it stands, Michigan’s numbers are as follows:
- Overall: 36% (#108)
- Offense: 24% (#132)
- Defense: 47% (#108)
Note that my numbers differ from Connelly's because he doesn't account for ties.
Here’s what he says about returning production:
On average, teams returning at least 80% of production improve by about 6.4 adjusted points per game in the following season's SP+ ratings. That's a pretty significant jump! For a team ranked 25th in SP+ last year, adding 6.4 points to its rating would have bumped it to about 13th. And if we lower the bar to just 70% returning production -- a bar 25 teams currently clear -- that's been enough to boost teams by an average of 4.0 points since 2014.
On the other end, about 11% of teams (roughly 15.2 of 134) return under 50% of their production in a given season. That results in an average drop of about 5.7 adjusted points in SP+. For a team that was 25th last year, losing 5.7 points would drop it to 51st. And for the rare team that returns less than 40% of production, the outlook is generally dire: Only 2.6% of teams fall under 40%, but they fall by an average of 9.8 points.
There are no guarantees here, of course. Returning 87% of its production didn't stop BYU from underachieving in 2022 and returning 33% didn't prevent Ohio State from improving slightly in 2016.
I won't just copy and paste the numbers because I don't enjoy Bolivia. Instead I just made some graphs. They’re arranged by conference, because I thought that’d be the most interesting way to display the data.
Presented without comment.
OVERALL
OFFENSE
DEFENSE
Thoughts?
February 5th, 2024 at 10:56 PM ^
Sorry. Forgot to provide the LINK ($) for anyone that's interested.
He also said the first set of SP+ numbers for 2024 will come out next week. That should be fun. 😊
February 5th, 2024 at 11:12 PM ^
Hey Blue@LSU,
I hope things are well.
Please send me an email to [email protected]. There are some questions I want to ask about prior diary entries you've done.
(I'll ask a few questions on other topics to check if it's you or an impostor emailing me.)
Regards,
Frank
February 5th, 2024 at 11:41 PM ^
Will do, Frank. I'm just off to bed, but I'll send an email tomorrow.
(Just don't make the questions too hard for me 😊)
Edit: just sent you a message from my proton account.
February 5th, 2024 at 10:57 PM ^
How do you quantify "No one throws at Will Johnson"?
That defense is going to be elite. Sometimes you look at a mathematic result and realize it doesn't capture what's going on.
Obviously, a lot needs to come together on offense. But if it can score 20, it will win a lot of games.
February 6th, 2024 at 9:14 AM ^
I mean an elite defense with a decent offense sounds like a top ten team which is what Connelly's data is saying.
February 6th, 2024 at 9:52 AM ^
This kind of stuff is supposed to go beyond the counting stats, but it seems that he's relying on counting stats to measure "production."
February 6th, 2024 at 1:41 PM ^
Its not just stats like pass breakups though, I think it's based more on starts and percentage of possible snaps played
February 5th, 2024 at 11:00 PM ^
We may not be bringing back much production on either side of the ball, and we may not be bringing back many of our core coaches. But, well, that’s all I have to say about that.
February 5th, 2024 at 11:08 PM ^
Defense should be amazing unless the loss of Minter hurts us. Derrick Moore, Graham, Stewart, Benny, Grant, Hausmann, Johnson, Rod Moore, Makari Paige. Offense is more sketchy. Need a transfer qb there.
February 5th, 2024 at 11:13 PM ^
With all the coaching turmoil, I'd be absolutely thrilled if we can keep our roster intact, especially our D.
February 6th, 2024 at 7:58 AM ^
Wile E Coyote year 0. If there's a drop-off in staff, I doubt it's super evident on the field in 2024
February 5th, 2024 at 11:15 PM ^
Oh no, we suck again!
February 6th, 2024 at 1:07 AM ^
I couldn’t even get past the first set of percentages before this popped into my head lol.
February 5th, 2024 at 11:42 PM ^
This along with the possibility of losing the entire defensive staff means Sherrone has a hell of an uphill battle. But I think most of us knew that.
February 6th, 2024 at 12:03 AM ^
I'd be happy with 10-2
February 6th, 2024 at 12:29 AM ^
I'd hope so. Even if Harbaugh and the whole staff returned, 10-2 would be phenomenal. You're replacing your QB, star RB, two best WRs, and entire OL -- and that's just one unit.
February 6th, 2024 at 3:48 AM ^
10-2 is a playoff team
February 6th, 2024 at 1:48 AM ^
These numbers seem silly at best.
We entered the offseason with sneaky repeat potential. We could still enter the fall that way if we can nail the DC hire and keep the roster intact. It’s seems poised to be another year where no one looks unbeatable, and Ohio State will probably be the best team we play all year no matter who else we end up facing.
Orji may prove to be a better deep ball QB than JJ, but the short and intermediate passing game could be dreadful w/o some very creative scheming to help him out.
February 6th, 2024 at 9:11 AM ^
Idk, seems very reasonable to me, we take a step back missing our awesome graduating class but stay in the top ten.
February 6th, 2024 at 9:57 AM ^
Repeat? Like, repeat as National Champs?
I'm going to skip the internet style bluster and simply say:
I don't think that's realistic. Any "expectation" that Michigan is a substantial repeat possibility is, IMO, unfair to the players and coaches on the 2024 team.
For SM to take this staff and this roster to a NC would be an astounding feat of football coaching. Like, far more astounding than it takes to "just" run a good to excellent Michigan football program over time, and that's pretty damn good.
Rob
February 6th, 2024 at 2:55 AM ^
I don’t mind -4ppg next season as a result of this season.
February 6th, 2024 at 9:30 AM ^
If we scored -4 points per game over last year we only lose to Alabama.
February 6th, 2024 at 6:52 AM ^
Only 2.6% of teams fall under 40%, but they fall by an average of 9.8 points.
Hey that's us!
So if Michigan is 10 points worse, our SP+ score would fall from 31.3 to 21.3. This would put us ranked 8th if things shake out similarly to last year.
I'll take it (as long as the sky doesn't fall any further).
February 6th, 2024 at 11:48 AM ^
8th (10-2 pre playoff) would be a really impressive accomplishment IMO. FWIW, I also thought 10-3 was an impressive accomplishment when JH did it as well. :)
Rob
February 6th, 2024 at 3:00 PM ^
I played with some numbers (forced a linear model onto BillC's returning production percentages) and got the following for implied lines for next year:
- -19.5 vs Fresno State
- +3.5 vs Texas
- -26 vs Arkansas State
- -18.5 vs USC
- -18.5 vs Minnesota
- -12.5 at Washington
- -20.5 at Illinois
- -30 vs Michigan State
- +6.5 vs Oregon
- -28 at Indiana
- -18.5 vs Northwestern
- +7 at Ohio St
And some theoreticals all on a neutral field:
- +6.5 vs Penn St
- +12 vs Georgia
- -5 vs Bama
February 7th, 2024 at 1:38 PM ^
Some of these lines look pretty rational. 9-3 would be a pretty good outcome, I think..However...
I do think some of these lines understates the difference played by guys that were somewhere between all-time greats and really good players for Michigan and the guys that will replace them.. I know this is pretty obvious, but....
JJ McCarthy to (say) Alex Orji - The distance here could be really huge. Not hating on Orji, just saying, if JJ is "the best quarterback ever to play for Michigan", and Alex Orji is not that. Also, he is playing with a 100% new o-line. Only Loveland back as a receiver... Given these variables I think that some of the lines in those games above might be difficult to achieve, or even impossible.
Mike Sanristil to Jaden McBurrows - Again, we're moving from a guy who made huge difference making plays seemingly all season to a guy who is likely to be a good player. I urge you to imagine 2023 with McBurrows playing nickel instead of Mikey.
Junior Colsun to Barsham Junior Colsun was a tackling force this year, who really glued the back seven together. Barsham is likely to be a good player. I think the distance between these guys is not as HUGE as it is in the other 2 places, but pretty damn big.
Corum/Edwards to Edwards/Mullings - Again, Edwards and Mullings - Likely to be darn good backs. Corum: greatest running back in Michigan history. 2023: Running behind extremely cohesive O-line. 2024: Literally every starter is new.
Offense and Defense, the potential for Michigan to have moved from awesome to good is really there, without being nasty or overly negative about the replacement situation.
9-3 is a darn good outcome in this situation, IMO. I'd be really happy with 9-3. 10-2 would be damn impressive. Better than that, eg, winning ALL those 10 games, plus beating texas or ohio state... That's an astounding outcome that "expecting it" is just wrong and unfair to every player and coach involved.
Rob
February 6th, 2024 at 9:03 AM ^
I think Brian did a good job explaining the flaw in the system in his UV. The talent on D is going to be elite. Maybe not top 5, but pretty high.
Texas had a horrible pass defense..maybe they get better?
USC has a horrible everything on defense and a new QB.
Washington lost everyone and is in a rebuild.
Oregon might be the best team on the schedule.
OSU is gonna have a lot of pretty stats, but haven't improved on the DL or at LB.
If Michigan can grab a QB and a WR after spring ball, (even Tuttle getting a 6th year would likely be my preference at this point) 10-2 seems like the floor.
February 6th, 2024 at 9:58 AM ^
And a DC?
February 6th, 2024 at 9:10 AM ^
Might be the year Nebraska finally breaks ou!
February 6th, 2024 at 9:23 AM ^
- Overall: 36% (#108)
- Offense: 24% (#132)
- Defense: 47% (#108)
This is a typo correct? It didn't seem right that our returning offense is worse than our defense yet the over all numbers were the same. Not mathematically impossible but unlikely.
So according to Connelly it reads:
- Overall: 36% (#128)
- Offense: 24% (#132)
- Defense: 47% (#109)
Worse but makes more sense.
February 6th, 2024 at 9:53 AM ^
Yeah. #108 overall was a typo. It’s correct in the graph.
February 6th, 2024 at 9:44 AM ^
Seems odd that defense is so low. Graham and Grant are back on the DL, Moore, Paige and Johnson are back, and then quite a few guys who didn't technically start but played a ton of snaps (Stewart, Hausmann, Moore, etc.) are back.
Offense makes sense since Loveland is the only returning starter and outside of him and Edwards, its entirely new faces everywhere, but I'm just really surprised that defense isn't higher up.
February 6th, 2024 at 9:55 AM ^
returning 33% didn't prevent Ohio State from improving slightly in 2016.
Don't know what the SP+ numbers were like, but the 2016 OSU team was definitely not as good as the '15 team which was absolutely stacked and probably would have repeated as national champ, if not for a crazy monsoon game against Sparty.
February 6th, 2024 at 7:15 PM ^
'15 OSU messed around a bunch before turning it on vs us and ND. So their overall SP+ numbers aren't that great even if their team at their peak was elite.
February 6th, 2024 at 10:08 AM ^
So we should all go bet on the over for everything regarding Virginia Tech and Iowa State?
February 6th, 2024 at 11:34 AM ^
Yes. I highly recommend you do that!
February 6th, 2024 at 2:13 PM ^
This is where we really find out if the people saying our recent success is an indicator of future success (IE, we don't need to change anything because 3 BIG Championships and a National Championship) is correct.
February 6th, 2024 at 4:05 PM ^
Problem is, we've already massively changed.
February 6th, 2024 at 2:26 PM ^
Wait, why is Iowa listed as returning production on offense?