OT: Moneyball & Stargazing
If you haven't watched Moneyball, the basic premise is that in the early 2000s, the Oakland As were a small team with a smaller budget that would get their recruits raided by bigger more well-funded teams like the Yankees
In the film, Jason Giambi is one of these players that goes off to the Yankees, leaving Oakland's cupboard bare after nearly winning the world series. There's a scene where a player analyst named Pete says something to the effect of "you shouldn't be trying to buy the best players, you should be buying wins, and if you want to buy wins you have to buy runs." So Oakland adopts an approach of replacing Giambi's On Base Percentage (OBP) with three lower-ranked players who have glaring weaknesses in their game BUT get on base quite a bit. The film simplifies how this actually played out, but ultimately this statistical approach was adopted and integrated into virtually every team in the MLB to date and is largely attributed to the Red Sox breaking their curse.
Does the parallel logic of "if you want to buy championships, you have to buy touchdowns, and if you want to buy touchdowns you have to buy first downs" hold up?
I understand that this fictionalized account of baseball history is wildly different from the world of NCAA football; however, there's some common threads here:
1. It's seemingly more challenging for Michigan to recruit blue chip players, even after two wildly successful seasons.
2. There's a lot of stargazing that seems to happen in football recruiting (at least in my limited view of the fan side) and well-rated recruits fall all over the spectrum of "played up to or above expectations" and "was a bust."
So, how much do you think this framework is applied in college football recruiting today in your experience? And do you think it would provide value to use this approach in recruiting (if it isn't already).
Moneyball from a football perspective is moving those chains and stopping the run. Which very much puts our coach up there.
The opposite would probably be recruiting 5 star WRs that would make Matt Millen salivate. Even our own fans applaud the passing yardage CJ Stroud put up on his way to 27 and 23 points. Those yards between the 30s being a lot like an intentional walk.
Ryan Day and OSU are very much the Yankees racking up huge offensive games vs mediocre or bad teams.
The coaches made it very clear before both of those games that OSU was going to move the ball. They were going to get yards. The key was to keep them out of the endzone. How many "4 point plays" was Michigan going to make? It worked wonderfully.
I don't think the Moneyball strategy applies to football, unfortunately. There isn't a single statistic or a set of statistics you can point to that say player X is dramatically undervalued. Baseball is a game where you can look at a box score and get an almost 100 percent accurate perspective of what took place. In football, you have to check the tape to see how well a player actually performed.
The best example of this is CJ Stroud himself. His stats in both games were fantastic, and yet they struggled to get to 20 points in each. In football, you need to marry great strategy to great athletes. That's why Alabama and Georgia (and somewhat OSU) have been so dominant. They have fantastic strategy and player development, but they also have some of the best athletes in the country.
In order to beat those teams, you need both. Or you need a few lucky bounces, a la TCU vs Michigan.
I think this is correct. To get the most out of a Moneyball theory you need lots of stats over a long period. It works over the grind of a baseball season but not for 12 or 16 games.
Yeah, that's not how it works.
The big difference with baseball and other sports is that there are only 27 outcomes compared to hundreds or thousands (or more) with each play. It makes data analysis for a single player much more definitive and easier.
I’m not sure that’s 100% accurate. They did it to Georgia who had an excellent defense. They were a missed FG on the final play of the game away from winning the national championship. That was with their top two wr’s (possibly the top two in the country), top two te’s, and top three rb’s out of the game due to injury.
Harrison Jr played most of that game and was the main reason they had the double-digit lead. I don’t think they scored after he went out.
You mean exactly what teams would let Rich Rod do. “Look at the yards Michigan racked up” cool look at the lack of points
Football is a very complicated game with a simple premise at its core: get first downs. Control the ball = keep the other teams basketball on grass offense off the field = control the clock.
And all of that boils down to make sure...make absolutely sure...you have a great offensive line. You know, like a two time Joe Moore award winning offensive line.
Moneyballing largely means being cost efficient on wins (doing better than you should based on payroll) while never actually winning championships.
Boston won by combining moneyball with actually buying the best players.
I don't know if that's entirely fair --- over a 162-season, being cost efficient almost always pays off (large enough sample size), but the playoffs of course are 3-,5-,and 7-game series. Oakland had some bad playoff luck after some excellent regular seasons in the 2000s.
I sound like an old fart MLB traditionalist, but I hate the wildcard. The Dodgers and Braves were, by far, the best teams from April to September in the NL last year. Yet the NLCS was Padres and Phillies. It felt more random than merit-based.
I don't know if that's entirely fair --- over a 162-season, being cost efficient almost always pays off (large enough sample size), but the playoffs of course are 3-,5-,and 7-game series. Oakland had some bad playoff luck after some excellent regular seasons in the 2000s.
I feel like you're both right. Yes, cost efficiency is always beneficial, but I think the original point is that efficient, but limited, players can only take you so far. It's no different than football. Our moneyball-like approach has gotten us two straight B1G titles and playoff appearances, but we haven't had the top end talent across the board to challenge UGA. That's where I think the Red Sox comment came in. Moneyballing got them in position, but it was the high end talent that got them over the top...
Kind of like the Tampa Bay Rays too. They're a small market team moneyballing their way to the top for the past 10 seasons, but never winning a World Series.
That's not at all what it means.
Then take the time to explain it
A little knowledge is a dangerous thing. Moneyballing is about understanding what the market currently over/under values and exploiting said inefficiencies.
It's tougher to do in college football because players can't be bought and sold. Well, they can kinda be bought. But they can't be sold. Maybe someday.
So "moneyballing" is done in a more indirect fashion. Sometimes you've got coaches who understand a particular recruiting market better than the wider market, so you apply more resources there. Like Don Brown and northeast US.
We're about to do it here with something mentioned on today's Roundtable podcast: greatly embiggening the roster to hoard more lottery tickets. The way we - and other big programs - did it back in the day before schollie limits. Using a resource UM has in abundance - money - to buy wins. Bama has used sketchy tactics to do this as well - sign a bunch of players, let them sink or swim for several months on campus and discard the failures quickly to stay under NCAA limits.
Limits which, as today's podcast mentions, seem to be fading away.
Good analysis. Also, in Moneyball, the inefficiency stemmed from the outdated lore of scouts that drove decision-making as opposed to analytics. It's somewhat different in college football, where the "analytics" are inefficient, being driven by the rating services with their biases and the outsized influence of the camp system. As a result, quality players can be under-rated because they came to football later or did not get exposure at camps and thus have their star rating suppressed. So Michigan has succeeded in getting players who "fly under the radar" of the rating system because they started football later (Ojabo) or didn't do the camps (Ronnie Bell). And Michigan exploits this situation by scouting (almost the reverse of Moneyball). But one way the analytics kick in is with respect to desired physical attributes. When Michigan sees the physical attributes, it goes against the rating system and pursues the player.
One major difference between baseball and football is the much more important team aspect, in which character and personality play a more important role. In baseball, it doesn't really matter if the guy is a self-absorbed jerk if he gets on base a lot. In football, the only players who perform individually are kickers and punters. Every other player must perform as part of a team and their ability to contribute to the "teamness" of the team plays an important role in success. Michigan clearly views this as an exploitable way to outperform recruiting rankings.
I'd argue Epstein used moneyball tactics to assemble the best team he could buy. The same as Bean, but with a substantially bigger budget.
Bean used market inefficiencies to his advantage. They were largely gone by Epstein's time, but you could (and can currently) use them to identify players' market value.
Coaching matters, and 'croots matter. Theyre both important, but I think the last two seasons have proved that the right coaching combo matters more than 'crooting. So many people on here said that Michigan shouldnt expect to beat OSU and that 10-3 was the best we could hope for.
Look at how Clemson has taken a step back since they lost some key assistants. Texas and Florida continue to struggle despite always recruiting well. OSU is recruiting their asses off still, but they obviously miss Urban's influence in late November. Ole Miss under Hugh Freeze had some of the best talent money could buy but couldnt put it all together.
The real kicker is when you get a combo of the best recruits and the best coaching like Alabama for so long and Georgia, scary. Or the 2000s USC teams.
Right now we are kind of like Tressel's 2000s OSU teams, run heavy that can play with anybody, but they really became dangerous once they started spreading the field out with Troy Smith. Hopefully that happens this year.
Right on the money. In college football, coaching is king and recruiting is queen.
As for the "moneyball" approach when it comes to roster building, it's certainly possible to build a short-term contender for a brief window if you hit it big with three stars... however, that model has never proven to be sustainable for building a long-term annual contender. 5 stars don't always hit or go supernova, but time has proven that 5 stars hit at a much higher rate than 3 stars.
Look at MSU under Dantonio... sometimes your 3-star recruits turn out to be Kirk Cousins and Le'Veon Bell and you can win a conference championship and make a run to playoffs... but the majority of the time they turn out to be Brian Lewerke and Madre London.
The thing to watch for me is how the transfer portal affects this formula. A school like Michigan — which is likely never going to recruit at a top-5 level for NIL reasons but should finish top 10-12 annually — can supplement its 5-star shortcomings by cherry-picking the three-star gems that other schools uncover. You don't need to recruit a 5-star interior lineman if you can instead swindle an Olu Oluwatimi off Virginia... or how USC stole Jordan Addison from Pitt... or how Lincoln Reilly regularly competes at a higher level than he should by picking off transfer QB after transfer QB (Mayfield-Murray-Hurts-Williams)
With the transfer portal, there's now a safety net for these top schools, where – if they whiff on the scouting and development side of things at a key position or two – they can salvage and bolster those positions via the portal... and I wonder if that makes the scouting and development approach more realistic for long-term contention.
Recruiting is VASTLY more important than coaching. Where are those Nick Saban national titles at Michigan State, his Super Bowls for the Dolphins? Saban's an excellent football coach, he's not winning national titles at Rutgers, UNLESS he is able to completely change the recruiting structure there.
I don't think you have to look further than the biggest snake-oil salesman in the country in Franklin, Jimbo, Kiffin, Freeze, etc to see that good coaching trumps good 'crooting every time.
And... how much of the ability to recruit is simply being a good coach in the first place? In your hypothetical, if Saban were to indeed to move to Rutgers, of course he wouldn't sign a top-5 class right away... but I could EASILY see Scarlet Knight Saban pull a top 20-25 class immediately with no issues and – after a winning season or two – pull together a top-10 class and then start contending.
You truly think any of those teams have what Bama has in terms of talent? They aren't even close.
Do you think we have the talent Bama has?
Well, Saban did sign a few really good guys and future NFL picks at MSU, and they were on probation when he first got there from Perles.
Plus he won LSU its first national title in 50 years.
If Saban had been at MSU long enough he most likely would have had a lot more success, it's not like Bill Snyder at KSU, Frank Beamer at VT or Barry Alverez at Wisconsin won instantly either at schools that historically weren't very good.
And those guys won because they started to recruit much better players. They didn't win because they had once in a lifetime coaching. KSU went all-in on the JUCO route. Virginia Tech had the 2nd best strength program in the country (Nebraska #1) in the 90s (i.e. player development).
So why have Virginia Tech and Kansas State fallen off in recent years then? Kansas State finally did well again last year.
Is your argument that coaching makes no difference vs players? If thats the case how did we go from 5-7 to 10-3 in one year with virtually the same team from Hoke to Harbaugh?
In regard to Clemson, they didn't just lose assistants but their top 2 coordinators. Brent Venables is an absolute stud of a coordinator too.
Of course it also helps having DeShaun Watson and Trevor Lawrence back to back. Can't deny that.
Yes, and when Dabo made "loyalty" decisions when it came to replacing his two long-time coordinators, they declined as a team. I live near Clemson and thus around die-hard Clemson fans. They grew to assume that any 5 star QB was a guarantee (Michigan fans know better). Their recent experience has pointed out to them that DeShaun and Trevor were positive outliers who were key to their championships. Just as JJ is essential to our title hopes this year.
In my opinion, if you want to buy wins, you have to buy yards. And yards in every phase, including and especially special teams because hidden yards are often ignored and yet are hugely impactful. Pretty much the only thing that derails total yards--including net kicking game yards--is settling for FGs (especially chip shot FGs) in the red zone. That is more impacted by play calling in my opinion.
So, recruit great kickers, return men, sure tacklers, and backs/receivers who are great YA(contact and or catch) guys, plus hire an OC who excels in the red zone.
I'm not sure I would agree with "buying yards", I would say "buy points". We've had teams sustain long drives but not be able to put the ball in the endzone which ended up costing us.
Also I remember the abortion of a game against Notre Dame in 2014 where we outgained them but lost 31-0...and people were trying to justify that outgaining them was a half win basically.
Buy the most resources efficient points - Touchdowns.
As I laid out at the top of the thread in a comment, yards don't necessarily equate to points. OSU has proved that the last two seasons. Scoring touchdowns, not field goals, is key to offense. Preventing touchdowns is key to defense. There is no way to money ball those two things as far as I know.
Over a large enough sample-size, yards are points.
A typical FBS college football season has 900 games: figure 12 drives for each team. That's 21,600 drives. Over a sample that large, the correlation between yards and points is very high. This correlation happens every single year.
It's a tricky business. One can and arguably should optimize for the aggregate ("buying yards"), but football (unlike baseball) isn't a sport of the aggregate. Ultimately there are only 12 games, and even then only a few of those games are the ones that really matter.
Having lived through RichRod, yards will never be points to me again.
In my opinion, if you want to buy wins, you have to buy yards.
The problem with this, though, is that yards don't always equate to points. RichRod had, statistically, pretty good offenses here, at least in that he put up a lot of yards. But he didn't put up a lot of points, and the points he did put up came mostly at the expense of bad defenses; we could barely score against defenses with a pulse. Take 2010 for example. We were #10 in yards per play, but only #25 in points per game. OSU against us the past two years is another example. They put up a lot of yards, but only 27 and 23 points.
Harbaugh the past couple years has been the opposite of RichRod, in that we've ranked much higher in points per game than we have in yards (both total yards, and yards per play). In other words, RichRod had an inefficient offense (much fewer points per yard), whereas Harbaugh's offenses have been much more efficient.
Not that yards aren't important, especially as you say hidden yards (which might be one of the secrets to Harbaugh's "efficient" offense), but I feel like points and efficiency are more critical than yards...
I disagree that yards are not points. To continue the baseball/moneyball analogy, a walk or a single is not a "run" but put enough of them together, and you will score more points over the long haul.
You (and a lot of other people responding) didn't read my post thoroughly enough to get the point. I'm talking about ALL the yards gained in a football game. A 45 yard punt with no return is 45 yards gained. I'm not just talking about offense and defense.
So, with respect, your argument about Rich Rod, yards/play, etc. is not really applicable here.
EDIT: Sorry, I didn't realize the audio was so badly out of sync.
Off-sync or not Chuck this is pretty cool and each response corresponds with their coaching style.
Key to wins; 1) Recruiting 2) Coaching.... Just look at schools with top 10 recruiting classes over the last 10 years. All of them are in the top 15 for total wins over that span.
I like how Jim's response is almost the exact opposite of Ryan Day. In fact, Ryan Day is the outlier among all of them.
Me watching Donovan run wild in Ohio:
Thank you Dennis!
That other guy was so lazy!
I think back to the survey from the coaches last year where they were asked what's the most overrated stat in football, and Harbaugh said, "anything besides points."
I do think about this when it comes to stargazing, because either the stars need to specify that they're talking about NFL potential, or the stars need t reflect the fact that there's a real argument that TCU recruited better than Texas over the last four years.
By the way, Moneyball is one of my favorite movies of all time
First of all, "even after two wildly successful seasons" isn't really true. Wildly successful for Michigan fans - yes. "Wildly successful" for the elite programs of the country, not so much. We beat OSU and won the Big Ten twice, but that's it. Again, I'm a fan, I'm as excited as all of us on this blog, but "wildly successful" is actually winning a championship, heck even playing for one comes close.
Having said that though, I think the football program has done that in many ways already. I don't have the links, and I CBA to look them up, but when they do the recruiting class in review once the class has graduated and you see Michigan re-ranked from the 10-20 class at time of signing to 3-10 at graduation, I think this speaks for itself.
I believe, for the most part, Michigan gets probably 70-80% of the recruits they want to fit the way they want to play. Sure they miss our but even Saban and Kirby miss out too, albeit more like the 90-95% range.
To give one example, I think the way Michigan recruits receivers fits into it. When Harbaugh came here originally they brought in huge classes of top, star receivers. They've since stopped doing that because they want receivers to play the way Jim wants them to play and that's block first, catch second and you're not going to bring in high stars that only want to look like Randy Moss. It seems to be working for them.
Yes...... and No.
All teams other than UGA, 'Bama, and OSU have to rely on finding hidden gems (value) to augment their talent base - so that already exists in some form.
The fundamental problem is that baseball is (at it's core) a one on one contest between pitcher and batter. We have seen during the Hoke years (and early-era Harbaugh OL coaching) where a plus play can be ruined or drastically reduced based on one miscue by another player regardless of the caliber of the "skill player"
My first thought for an example was one of the many missed block IDs by Hoke's highly rated OL groups . - but instead i want to use a positive play ------ CJ's 1st TD in the 'shoe last year.
- A non-JJ QB doesn't hit that route under duress -- (Cade turfs it - sorry Don Thomas)
- A slightly worse OL effort or a failed RB block doesn't give that split second of extra time for the throw
- On OSU -- If 5-star talent Jack Sawyer had hustled and not assumed the CB would make the tackle CK never gets to the end zone and the whole game changes.
- On the other hand if Marvin Harrison played for M maybe he would have been more open and not depended on the gaffe by Sawyer?
TLDR - Too many moving parts to be an effective overall strategy.
Very well put. Football and Baseball are fundamentally different games, thus Moneyball will look different for them. I do think it's possible to use Moneyball in NCAA Football, but would be temporary and not a long term success story.