Fascinating: Moneyball going to the NFL

[ShotgunDawg]

Yesterday, I believe a fascinating thing happened in the NFL when the most dysfunctional franchise in the NFL hired Paul DePodesta to head up their organization. I realize that the NFL & college football has been using some form of analytics for some time now, but I believe the hire of DePodesta by the Browns signals that this thing is about to be taken the next level & will eventually offer the Browns some sort of competitive advantage on the field.

My question is: since MSU is already the Oakland As of the SEC, can we lead the college football charge in this area? In baseball, the analytical revolution was not led by the Yankees, Red Sox, or Dodgers, but rather by small market teams that were forced into a desperate search for an edge. In many ways I see MSU as the perfect college football circumstance for the growth of advanced analytics in college football. We have money, limited resources, a passionate fan base that wants to win, & a need to find more efficient & better ways to do things. We are never going to beat Alabama or LSU in recruiting, so the only way to consistently beat them is to be smarter & more efficient than them.

I'm not sure what already exists in college football, in regards to this, but I'd love to see MSU Football form a Research & Development Analytics Department of about 3 or 4, extremely smart, potentially IVY League educated personnel, that would be paid 6 figure incomes & charged with finding new, innovative ways for the MSU Football team to be more efficient, run more statistically probable plays, put the correct players on the field in optimal situations & positions, & push the envelop in evaluating high school football prospects & increasing the hit rate by recruiting the correct players.

Currently, almost all of this responsibility is charged to the coaching staff, but no matter how smart the coaching staff is, they are hardly the best group of people to be performing this task. Football coaches don't have economic or statistics degrees & they don't have time to spend hours & hours on R&D. I believe this an area that MSU should embrace & lead the college football revolution in. Why? Because we need to.

Here's a good article:

http://mmqb.si.com/mmqb/2016/01/06/nfl-cleveland-browns-paul-depodesta-jimmy-haslam

[thf24]

Interesting concept for sure. The main problem I see with applying it to us, or most college teams, is that we can't go out and get free agents as needed to satisfy the model. I'd be interested to see it applied to something like personnel combinations though.

[ShotgunDawg]

Interesting concept for sure. The main problem I see with applying it to us, or most college teams, is that we can't go out and get free agents as needed to satisfy the model. I'd be interested to see it applied to something like personnel combinations though.

Correct, we can't get free agents, but that's only a sliver of what this is all about.

For example: MLB teams are using shifts & NBA teams are using still frame pictures from above the court at different points in the game, to statically figure out which players create the most space or which combination of players creates the optimal lineup against certain teams.

The point is that the need for advanced analytics goes far beyond just math or just acquiring players. Figuring out which plays and players are statistically optimal in certain situations against certain teams, could be invaluable, & that's only one example. The opportunities here are endless & I believe MSU should lead the charge. However, you have to be committed to it & go out & hire legit people to be a part of this department.

MSU needs to start being a leader instead of a follower. The only way we are going to become a national power, without cheating, is if we become committed to pushing the envelope & getting ahead of other programs. The last thing I want to see is the Browns start improving, due to this, & Alabama leads the charge in the anaytical relm, & we are left just chasing Alabama again. Let's get ahead of this & lead the charge.

[Irondawg]

I don't how you could do it recruiting without coming up with some kind of psych evaluation. I think it would be near impossible to predict who will grow a few inches or be able to carry more weight without losing speed and most important who wants to become the best football player they can and won't get distracted by other stuff.

It's a neat idea but not sure exactly how it would apply to recruiting.

Now on the flip side doing a psych evaluation and trying to determine if you can come up with some reference points on who is most likely to fulfill potential vs flame out would be extremely interesting.

[ShotgunDawg]

I don't how you could do it recruiting without coming up with some kind of psych evaluation. I think it would be near impossible to predict who will grow a few inches or be able to carry more weight without losing speed and most important who wants to become the best football player they can and won't get distracted by other stuff.

It's a neat idea but not sure exactly how it would apply to recruiting.

Now on the flip side doing a psych evaluation and trying to determine if you can come up with some reference points on who is most likely to fulfill potential vs flame out would be extremely interesting.

Psych evaluations & eye tests are already done in MLB. That's a given part of it, but I'm sure that extremely smart people with all day to think about this stuff, could come up with even more ways of increasing the hit rate on high school prospects. MLB team already do this everyday.

Point is, this is why you hire an R&D analytics department full of extremely smart people. Their job is think outside of the box, examine every facet of the program & game, & come up with innovative, proprietary ways of becoming more efficient & optimizing the effectiveness of each player, coach, play, etc in the program. The opportunities here are endless.

There are endless amounts of IVY league like, extremely intelligent people that are graduating & want to work in sports, but can't get a job. You could find highly qualified, brilliant people to do this. You'd just have to play well.

I'm sure we'll just do what MSU has always done and just wait for Alabama to do it and let us know if it works. Conservative is our middle name

[BrunswickDawg]

To some degree we already are. If you know that a big hinderance to the program is attracting 5 star talent - or free agents, statistically going for the top 3 stars or players who aren't well scouted (small school kids) gives your a better chance of finding players who give you a high return on your investment without outstripping low resources (recruiting budgets).

[ShotgunDawg]

To some degree we already are. If you know that a big hinderance to the program is attracting 5 star talent - or free agents, statistically going for the top 3 stars or players who aren't well scouted (small school kids) gives your a better chance of finding players who give you a high return on your investment without outstripping low resources (recruiting budgets).

Correct, we are doing some of this already, but not near to the level that it could be done. We have staffers that search out for under the radar talent, but I'm talking about hiring sabremetricians & highly educated people to form a think tank that could optimize the efficiency & personnel use throughout the entire football program. I'm not talking about hiring GAs or ex high school coaches. I'm talking about hiring people that have offers from Google, Apple, & Microsoft, but want to work in sports.

Some program is going to do this very very soon in college football & they will get an edge. Do we want to follow or lead it?

[Beaver]

I'm not opposed to the idea. Shoot, I'd love to do something like that. The main difference between Moneyball and football are the number of data points. It's usually easier to see trends in the data when there are lots of data points. Also, college teams change so much from year to year -- along with coaches and other staff, it would be hard to form any kind of consistent database. You'd basically be starting from scratch year after year. Plus in the SEC, if you try something for a couple games and it doesn't work, your season is over... In baseball, you'd still have 160 other games to try other things.

If MSU were to do something like that, I'd start with hiring a couple guys or even using a couple smart GA's and have them collect data from teams all over the country during the course of a season, and spend the off-season analyzing the data and see what patterns they can find in playcalling, recruiting/freshman PT, game management, and physical traits of different teams that make up a successful (relatively speaking) team.

[Dawg Corps]

I'm going to stay out of the discussion on how useful/applicable this concept would be to NCAAF. However, I'd like to interject the fact that just because someone has a degree from an Ivy League school doesn't make him/her a better analyst or statistician than anyone else. It's all about having the judgement to properly apply whatever data you've collected based on how significant your results are...

[PassInterference]

Does this mean the Browns will go for it every 4th down? That's supposed to be the statistical thing to do.

[BulldogDX55]

I like the idea, but here is the problem: Do you see Dan Mullen doing what a bunch of stats nerds tell him to do, playcalling wise?

Either he goes along with it, because winning is what he loves most, or he shuts it down because ego. You'd have to make sure Dan is on board first and foremost.

[gravedigger]

Yesterday, I believe a fascinating thing happened in the NFL when the most dysfunctional franchise in the NFL hired Paul DePodesta to head up their organization. I realize that the NFL & college football has been using some form of analytics for some time now, but I believe the hire of DePodesta by the Browns signals that this thing is about to be taken the next level & will eventually offer the Browns some sort of competitive advantage on the field.

My question is: since MSU is already the Oakland As of the SEC, can we lead the college football charge in this area? In baseball, the analytical revolution was not led by the Yankees, Red Sox, or Dodgers, but rather by small market teams that were forced into a desperate search for an edge. In many ways I see MSU as the perfect college football circumstance for the growth of advanced analytics in college football. We have money, limited resources, a passionate fan base that wants to win, & a need to find more efficient & better ways to do things. We are never going to beat Alabama or LSU in recruiting, so the only way to consistently beat them is to be smarter & more efficient than them.

I'm not sure what already exists in college football, in regards to this, but I'd love to see MSU Football form a Research & Development Analytics Department of about 3 or 4, extremely smart, potentially IVY League educated personnel, that would be paid 6 figure incomes & charged with finding new, innovative ways for the MSU Football team to be more efficient, run more statistically probable plays, put the correct players on the field in optimal situations & positions, & push the envelop in evaluating high school football prospects & increasing the hit rate by recruiting the correct players.

Currently, almost all of this responsibility is charged to the coaching staff, but no matter how smart the coaching staff is, they are hardly the best group of people to be performing this task. Football coaches don't have economic or statistics degrees & they don't have time to spend hours & hours on R&D. I believe this an area that MSU should embrace & lead the college football revolution in. Why? Because we need to.

Here's a good article:

http://mmqb.si.com/mmqb/2016/01/06/nfl-cleveland-browns-paul-depodesta-jimmy-haslam

If it can be translated to recruiting, and I think it can, the result would be a team signing a bunch of 'diamonds in the rough'. I'm not opposed to that at all. For instance, if correctly done, it would find the Dak Prescott's and focus on them rather than the bidding wars that go on for the Shea Pattersons. It would actually kill recruiting rankings for us as a team. The thing is, I think Mullen does this to a degree already. The kids leadership ability and demonstration of it in High School and in camps is a trait he prioritizes as much or more than anything. I think that is one reason he's been as successful as he's been and probably why Freeze has had some trouble. As someone who analyzes data all day every day, deciphering between relevant and irrelevant statistics would be a tough job.

[BrunswickDawg]

If it can be translated to recruiting, and I think it can, the result would be a team signing a bunch of 'diamonds in the rough'. I'm not opposed to that at all. For instance, if correctly done, it would find the Dak Prescott's and focus on them rather than the bidding wars that go on for the Shea Pattersons. It would actually kill recruiting rankings for us as a team. The thing is, I think Mullen does this to a degree already. The kids leadership ability and demonstration of it in High School and in camps is a trait he prioritizes as much or more than anything. I think that is one reason he's been as successful as he's been and probably why Freeze has had some trouble. As someone who analyzes data all day every day, deciphering between relevant and irrelevant statistics would be a tough job.

In season, I think it can help more on the predictive scouting side. Analysis of opponents tendencies comes to mind - What % do they run play X on 3rd down? What is their most effective play against X defensive set? How often do they go to player X in situation Y? We could apply it to our own tendencies too to see if we are making predictive calls. All coaches/teams develop habits that can be exploited - which is why Dak on 3rd and short became an ineffective play for us.

[Hypnodawg]

I'm not opposed to the idea. Shoot, I'd love to do something like that. The main difference between Moneyball and football are the number of data points. It's usually easier to see trends in the data when there are lots of data points. Also, college teams change so much from year to year -- along with coaches and other staff, it would be hard to form any kind of consistent database. You'd basically be starting from scratch year after year. Plus in the SEC, if you try something for a couple games and it doesn't work, your season is over... In baseball, you'd still have 160 other games to try other things.

If MSU were to do something like that, I'd start with hiring a couple guys or even using a couple smart GA's and have them collect data from teams all over the country during the course of a season, and spend the off-season analyzing the data and see what patterns they can find in playcalling, recruiting/freshman PT, game management, and physical traits of different teams that make up a successful (relatively speaking) team.

I actually talked with Mullen about this (and more) a couple years back. He seemed interested, its my fault for not following up.

"The main difference between Moneyball and football are the number of data points." This was his main problem with most of what I was suggesting.

I've done a good bit more research on the topic since then. As I have frequently discovered, it is difficult to have an original idea, and there is probably a website dedicated to it. That said, the people working this are still missing some huge pieces.

[ShotgunDawg]

I'm going to stay out of the discussion on how useful/applicable this concept would be to NCAAF. However, I'd like to interject the fact that just because someone has a degree from an Ivy League school doesn't make him/her a better analyst or statistician than anyone else. It's all about having the judgement to properly apply whatever data you've collected based on how significant your results are...

Completely agree.

Just said "IVY League" to qualify what type of candidates I'd like to hire.

[smootness]

Analytics can be used in football to determine the percentage of certain plays/decisions, but I don't think it will ever be involved in player evaluation and prediction to the extent it is currently used in baseball and now basketball. Baseball is a very mathematical sport. Each hitter faces the same scenarios and pitchers and can be evaluated apart from any teammates. Basketball is more difficult, but it is still ultimately comes down to one guy shooting the ball at the basket. Football is different; it is the most dependent on teammates and what the defense is giving you.

I just don't see it ever being as effective as it has proven to be in baseball. I don't think the Browns know what they're doing.

[Dawg61]

Everybody has a plan till they get popped in the mouth. Sabermetrics in the NFL? Haha

[BB30]

Analytics can be used in football to determine the percentage of certain plays/decisions, but I don't think it will ever be involved in player evaluation and prediction to the extent it is currently used in baseball and now basketball. Baseball is a very mathematical sport. Each hitter faces the same scenarios and pitchers and can be evaluated apart from any teammates. Basketball is more difficult, but it is still ultimately comes down to one guy shooting the ball at the basket. Football is different; it is the most dependent on teammates and what the defense is giving you.

I just don't see it ever being as effective as it has proven to be in baseball. I don't think the Browns know what they're doing.

Being a Browns fan I can assure you this is most likely the case.

[Blackout]

Analytics can be used in football to determine the percentage of certain plays/decisions, but I don't think it will ever be involved in player evaluation and prediction to the extent it is currently used in baseball and now basketball. Baseball is a very mathematical sport. Each hitter faces the same scenarios and pitchers and can be evaluated apart from any teammates. Basketball is more difficult, but it is still ultimately comes down to one guy shooting the ball at the basket. Football is different; it is the most dependent on teammates and what the defense is giving you.

I just don't see it ever being as effective as it has proven to be in baseball. I don't think the Browns know what they're doing.

Yes

[sleepy dawg]

I know some studies have been done on this before. One study I saw a few years ago showed you should never punt or kick field goals, except in extremely rare scenarios.