Draft Countdown Forums

Draft Countdown Forums (http://www.draftcountdown.com/forum/index.php)
-   2014 NFL Draft Forum (http://www.draftcountdown.com/forum/forumdisplay.php?f=47)
-   -   Moneyball (http://www.draftcountdown.com/forum/showthread.php?t=52494)

BoiseSt39 04-24-2012 09:08 AM

Moneyball
 
I'm reading the book "Moneyball" and I'm sure many of you have read the book or at least seen the film.

I'm wondering if there are any parallels between MLB scouting fallacies and NFL scouting fallacies.

In the book, the A's GM realizes that batting average and fielding ability is dramatically overvalued and that on-base percentage and walks are dramatically undervalued.

Has there been any study into football statistics?
Is bench press overvalued and yards after contact undervalued in a RB?
Is there an over emphasis on height with QBs?
How overvalued is the 40 time vs. agility drills?

Shane P. Hallam 04-24-2012 09:12 AM

Studies have been done, most results are very inconsistent and inconclusive. With football being dependent on multiple positions to gain you statistics, it is difficult to evaluate.

Closest it has been is pass rush, sacks. 20 is the magic number for sacks in college that dictate NFL success, but even that is iffy.

Big Bird 04-24-2012 09:19 AM

Boise State fan, starting a moneyball thread, in which he asks if there is an over emphasis on height with quarterbacks.

Hmm, I wonder who he was thinking of when he started this...

keylime_5 04-24-2012 09:25 AM

The height thing with QBs is more about history than anything. In the old days it wasn't as big of a deal b/c the OLinemen were shorter and there were fewer people in the world who played football compared to today. In the last 25ish years, if you are under 6'2" you almost certainly are not gonna be a franchise QB.

I think it's a little tougher to find things like that that are undervalued in football b/c baseball is a game where statistics don't lie and you can tell everything you need to know about how good or bad a player is doing with stats. There are so many positions in football that don't even recrod statistics or only fill roles in a scheme that it's hard to tell much about a player with stats. Same with measurables- there are certain things that are important, but everyone knows that it doesn't mean you'll be a bad player. Usually though, small and slow doesn't do well in the NFL at any non-kicker position.

BoiseSt39 04-24-2012 09:27 AM

Quote:

Originally Posted by Big Bird (Post 2963514)
Boise State fan, starting a moneyball thread, in which he asks if there is an over emphasis on height with quarterbacks.

Hmm, I wonder who he was thinking of when he started this...

No mystery there.

Lil Quip 04-24-2012 09:46 AM

As has been alluded to, the NFL really needs to develop stats for certain positions before we can evaluate their deficiencies and make up new ones.

That being said, QB kicker and punter are only positions in football that makes sense to do that kind of rewriting of statistics.

In regards to the draft, I think there is more opportunity though. I think the best correlation to Moneyball is the 40. Its weight has been diminished to a large degree. One aside, one thing I have always wondered is why teams don't have prospects run in pads. It hurts times and it would take time to get a large enough sample, but I think it would make the 40 a more telling parameter.

The way the league works, I think it would take a grass roots movement of a team developing its own drills to complement the generic ones. If a team develops a DB drill that is a better measure of success than those in practice, other teams would be quick to jump on it.

CDCB14 04-24-2012 09:53 AM

Quote:

Originally Posted by Shane P. Hallam (Post 2963507)
Studies have been done, most results are very inconsistent and inconclusive. With football being dependent on multiple positions to gain you statistics, it is difficult to evaluate.

Closest it has been is pass rush, sacks. 20 is the magic number for sacks in college that dictate NFL success, but even that is iffy.

Wait what do you mean? If a player hits 20 sacks in a season they usually do well in the NFL? Or do you mean as a career?

Bixby (Thumper) 04-24-2012 10:01 AM

Too much emphasis on off the field problems.

Shane P. Hallam 04-24-2012 10:02 AM

Quote:

Originally Posted by CDCB14 (Post 2963547)
Wait what do you mean? If a player hits 20 sacks in a season they usually do well in the NFL? Or do you mean as a career?

In a career. It's one of those general rules that aren't all that accurate.

DeepThreat 04-24-2012 10:02 AM

Quote:

Originally Posted by Bixby (Post 2963554)
Too much emphasis on off the field problems.

I feel like this is probably pretty accurate when it comes to arrests but not issues like laziness.

scottyboy 04-24-2012 10:06 AM

Quote:

Originally Posted by Big Bird (Post 2963514)
Boise State fan, starting a moneyball thread, in which he asks if there is an over emphasis on height with quarterbacks.

Hmm, I wonder who he was thinking of when he started this...


CJSchneider 04-24-2012 10:13 AM

As far as statistical fallacies and anomalies go, I think QB rating is of little use.

WCH 04-24-2012 10:37 AM

There's been a lot of work done in this area, and most of it either doesn't work very well (yet) or uses flawed methodology. Advanced statistical analysis of football is still just barely past the "Bill James publishing books in his garage" stage, and it's even worse at the college level.

There are a lot of problems if you even want to attempt this stuff. We can't do things like calculate adjusted +/- scores with any consistency, because we don't know when players are in the game unless we chart it ourselves. Play-by-Play can be hard to come by at the college level, and are almost nonexistent for smaller schools or games before about 2003. Public combine data is difficult to come by once you go much beyond 10 years, and it's only been the past few years that conventional college statistics have been easily available going back much further than a decade.

I've done a lot of research in this area and I'll try to post more this evening, but here are a few links from Football Outsiders:

SackSEER (Pass Rushers): http://www.footballoutsiders.com/nfl...-forecast-2012
Lewin Career Forecast (QBs): http://www.footballoutsiders.com/nfl...-forecast-2012
Playmaker Score (WRs): http://www.footballoutsiders.com/fo-...maker-score-20
POE and Speed Score (RBs): http://www.footballoutsiders.com/var...d-draftability

As you'll see, this is stuff is generally much better at identifying busts than it is hitting on successes. Even then, it's not great.

It's spectacular that people are trying to do this at all, but my biggest complaint is that a lot of this work just isn't methodologically sound.

Rosebud 04-24-2012 10:40 AM

Didn't JPP do absolutely terrible in all of those bust-identifying advanced stastics?

WCH 04-24-2012 10:46 AM

Quote:

Originally Posted by Rosebud (Post 2963603)
Didn't JPP do absolutely terrible in all of those bust-identifying advanced stastics?

He was projected to be a spectacular failure. He had bad combine numbers and didn't have very much college production.

Needless to say, scouts don't have to worry about being replaced by computers any time soon.

MI_Buckeye 04-24-2012 04:35 PM

The biggest problem with using a metrics-based system to project future success for prospects in football is that the game is so different at the NFL level, and there is such a wide variation of schemes and fundamentals.

Baseball, no matter what level you play at, is virtually the same worldwide. Sure, the players might be bigger, stronger and more skilled, but the basics remain the same. In football, a player can go from a Wing-T in high school, to a read option-based offense in college, to a Coryell-style attack in the NFL. The difference in the skill set required to succeed in these offenses are so profound, it diminishes the importance of statistical production at the previous level.

Also, the game of baseball is very conducive to statistical analysis. Everything they do can be measured. Football statistics do not reveal nearly as much about a player's value or ability. Think about it like this, the most statistically relevant position in the NFL is running back. That's why running backs dominate the early roounds of nearly every fantasy football draft. Does that mean running backs are the most important players? Clearly not.

TuckNole 04-25-2012 01:23 PM

Statistical analysis in football is very good...

for helping thing about strategy, (and or predicting wins / losses if you are a gambler) but still isn't great at scouting.

I read football outsiders religiously, but they are still tweaking stuff every year as they discover more and more. I find the process more interesting than anything as it helps you think about things you hadn't before. Like this year for pass rushers. Turns out passes batted down is almost as relevant as sacks for predicting future pass rushing ability.

cajuncorey 04-25-2012 01:59 PM

The Oakland raiders believe 40 time applicable to every position even offensive line.

descendency 04-25-2012 02:40 PM

I've never seen anything close to what SABRmetrics can do in baseball. Most of the statistics I see are ones related to strategy (not punting, not kicking field goals, etc)

NorrinRadd12 04-25-2012 02:49 PM

They should add power clean and squat to the combine and make them do the drills, especially the 40, in full pads.

MightyMitts 04-25-2012 02:53 PM

Like what was mentioned earlier, statistical analysis is best used to downgrade prospects because they either lack the (seemingly) necessary physical ability or productivity. Determining what ability, skills, and productivity for a position is the hard part.

In the last few years, when a pass rusher, for instance, "out performs" poor agility and quickness numbers it's usually because they are tall, strong, and have very long arms. Jason Pierre-Paul, Carlos Dunlap, and Aldon Smith fit into this group. In 2012 this may also apply to Chandler Jones and Frank Alexander.

whatadai 04-25-2012 08:03 PM

Quote:

Originally Posted by Shane P. Hallam (Post 2963556)
In a career. It's one of those general rules that aren't all that accurate.

What if they played for 4 years and only got 5 eaxh season?


All times are GMT -5. The time now is 10:46 AM.

Powered by vBulletin® Version 3.6.4
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.