Money Ball

[GRT8]

I have a background in scouting in my younger days, but have changed careers since for family reasons. I always have had a strong opinion of how to analyze players,teams, needs and have been around people that thought that way......it isn't right or wrong......just my way..............DON"T put so much emphasis on stats.Don't ignore it, but don't be obsessed with it. Tom Brady would never be a first round pick by stats in college. But if the team drafting him had aquired more picks, they may find a Brady in their selections.

When i see people looking at stats and "tendancies" i applaud their research and talent......but what are stats that are available in so many forms allover the internet,good for......debate....the obvious......we all do it. I do. A better analysis of habits or tendancies is when they are behind...inside the 20 yard line/ in rain/ double teamed...game situations vs.stats. only avail on film, which some are not able to view,unless have resources.

A better tool IMO is to look for natural football talent,smarts, playmaking ability and football intelligence....that stats don't always provide.A good read is Moneyball. This is just to add to so many people's great analyzing skils in here......and find your stragedy for the GM in us all.


In football, the outcome of every play includes contributions by 22 different players, making a single player's stats, such as yards per carry, completions or tackles somewhat harder to value. But just as much of the groundbreaking statistical analysis in baseball was first done by those who didn't work for any major league teams, there are new football metrics being compiled by some league outsiders as well.
As general managers and scouting staffs huddle up to try and find the best player, they'll pore over players' 40-yard dash times and stats collected during college football games against opponents who aren't of the caliber of even the weakest of the NFL teams.


Relying on such stats is one reason that there have been so many high-profile bad draft picks. But it doesn't have to be that way

I am a firm believer in trading down ( if avail) to gather picks and find contributers...

 

[VT Dolphan]

Good post, I agree completely

 

[O'townPhinFan]

I have a background in scouting in my younger days, but have changed careers since for family reasons. I always have had a strong opinion of how to analyze players,teams, needs and have been around people that thought that way......it isn't right or wrong......just my way..............DON"T put so much emphasis on stats.Don't ignore it, but don't be obsessed with it. Tom Brady would never be a first round pick by stats in college. But if the team drafting him had aquired more picks, they may find a Brady in their selections.

When i see people looking at stats and "tendancies" i applaud their research and talent......but what are stats that are available in so many forms allover the internet,good for......debate....the obvious......we all do it. I do. A better analysis of habits or tendancies is when they are behind...inside the 20 yard line/ in rain/ double teamed...game situations vs.stats. only avail on film, which some are not able to view,unless have resources.

A better tool IMO is to look for natural football talent,smarts, playmaking ability and football intelligence....that stats don't always provide.A good read is Moneyball. This is just to add to so many people's great analyzing skils in here......and find your stragedy for the GM in us all.


In football, the outcome of every play includes contributions by 22 different players, making a single player's stats, such as yards per carry, completions or tackles somewhat harder to value. But just as much of the groundbreaking statistical analysis in baseball was first done by those who didn't work for any major league teams, there are new football metrics being compiled by some league outsiders as well.
As general managers and scouting staffs huddle up to try and find the best player, they'll pore over players' 40-yard dash times and stats collected during college football games against opponents who aren't of the caliber of even the weakest of the NFL teams.


Relying on such stats is one reason that there have been so many high-profile bad draft picks. But it doesn't have to be that way

I am a firm believer in trading down ( if avail) to gather picks and find contributers...

I don't think I could have said it better muhself!!!

Good post man!!!

 

[PhinsRock]

Very good post, but be careful - making sense isn't popular around this forum.

 

[ckparrothead]

I often find that the people that spurn statistics are those that do not know how to use them properly or creatively.

Statistics are a context builder. Consistent correlations are meaningful. Was it a coincidence that in the 22-factor athleticism models that I developed last season, both Mark Anderson and Marques Colston graded out very, very highly?

No. Not really. People forget the origin of statistics. They are amalgamated factual evidence. Evidence is what we all use to make any evaluation, be it visual or statistical.

In the amalgamation of the factual evidence into a construct that is both meaningful and portable, a lot is left behind. No statistician would ever claim that statistics encapsulate everything. Nor would any statistician claim perfect correlation between any statistic and NFL value.

However, it is well to keep in mind that perceptions lie just as easily as statistics lie. Perceptions are faulty; they are subject to emotional bias, and subject to imperfect memory bias.

People do not have infallible memory banks like a computer. Most operate on impression memory, condensing an event down into an impression of the event. Drawing up an evaluation based on an array of impression memories in a manner that is consistent is nearly impossible. That process can be just as flawed as statistics-mongering.

I use statistics as a context builder, and place meaning on consistency within statistics. I have done very well evaluating my own observations as well as the observations of others in that manner.

 

[SoDakDolfan]

The only thing I would add to your list is an incredibly strong 'will to succeed.' IF a player has that determination they can do a lot even if they don't meet all the normal 'stat' requirements. Exhibit A would be Zach Thomas.

 

[caneaddict]

Hey GRT8-

Have you received my PM's? I've been trying to reach you with no response?

Please PM back.

 

[GRT8]

I often find that the people that spurn statistics are those that do not know how to use them properly or creatively.

Statistics are a context builder. Consistent correlations are meaningful. Was it a coincidence that in the 22-factor athleticism models that I developed last season, both Mark Anderson and Marques Colston graded out very, very highly?

No. Not really. People forget the origin of statistics. They are amalgamated factual evidence. Evidence is what we all use to make any evaluation, be it visual or statistical.

In the amalgamation of the factual evidence into a construct that is both meaningful and portable, a lot is left behind. No statistician would ever claim that statistics encapsulate everything. Nor would any statistician claim perfect correlation between any statistic and NFL value.

However, it is well to keep in mind that perceptions lie just as easily as statistics lie. Perceptions are faulty; they are subject to emotional bias, and subject to imperfect memory bias.

People do not have infallible memory banks like a computer. Most operate on impression memory, condensing an event down into an impression of the event. Drawing up an evaluation based on an array of impression memories in a manner that is consistent is nearly impossible. That process can be just as flawed as statistics-mongering.

I use statistics as a context builder, and place meaning on consistency within statistics. I have done very well evaluating my own observations as well as the observations of others in that manner.

I really think you are getting most of the points, but no professional is going to grade someone by feeling or emotion or impression of an event. I believe that football intelligence and natural instincts far outweigh statistics.It isn't people forget the origin of stats, they forget that stas are not the origin of early evaluation in a career...too many players don't come with stats,but ability. Not every position is guided by stats. You have a good system that you use......you" are not the example" no one singled you out. keepupthe good work.

 

[wazzy]

Hey GRT8-

Have you received my PM's? I've been trying to reach you with no response?

Please PM back.

Great name cane......... With all this said I have a question for the original poster!

Who do you think of watching all the qb's coming into the draft is the best fit for Miami?

 

[GRT8]

I truely cannot say at this point as I haven't looked at them all. I really don't think we will go Qb this draft......but i will post it if I get time to evaluate them.

 

[Dj Shoj]

AND ...........that is why we should draft GINN with the #9 PICK.

 

[TractorTraylor]

I often find that the people that spurn statistics are those that do not know how to use them properly or creatively.

Statistics are a context builder. Consistent correlations are meaningful. Was it a coincidence that in the 22-factor athleticism models that I developed last season, both Mark Anderson and Marques Colston graded out very, very highly?

No. Not really. People forget the origin of statistics. They are amalgamated factual evidence. Evidence is what we all use to make any evaluation, be it visual or statistical.

In the amalgamation of the factual evidence into a construct that is both meaningful and portable, a lot is left behind. No statistician would ever claim that statistics encapsulate everything. Nor would any statistician claim perfect correlation between any statistic and NFL value.

However, it is well to keep in mind that perceptions lie just as easily as statistics lie. Perceptions are faulty; they are subject to emotional bias, and subject to imperfect memory bias.

People do not have infallible memory banks like a computer. Most operate on impression memory, condensing an event down into an impression of the event. Drawing up an evaluation based on an array of impression memories in a manner that is consistent is nearly impossible. That process can be just as flawed as statistics-mongering.

I use statistics as a context builder, and place meaning on consistency within statistics. I have done very well evaluating my own observations as well as the observations of others in that manner.

I am currently reading Moneyball and to me, it is super interesting. CK, Id was wondering if youd share those athleticism models with me, just to see what youre doing. I find this area very intriguing but my question for you is, since your models had success last year, what do they say is the proper pick for us and who should we watch as a sleeper pick (ala Colston)? Hope to hear your answer.

 

[IluvJuMiami]

Great thread. Great response fom CK. Finheaven is on fire today. By the way, who's the SOB who gave the thread anything lower than 5 stars? You'd figure after the simple "lets get Samuels" thread or "I hate Daunte", something like this would be highly appreciated.

Edit: Sorry CK, I obviously ment you.

 

[ckparrothead]

I am currently reading Moneyball and to me, it is super interesting. CK, Id was wondering if youd share those athleticism models with me, just to see what youre doing. I find this area very intriguing but my question for you is, since your models had success last year, what do they say is the proper pick for us and who should we watch as a sleeper pick (ala Colston)? Hope to hear your answer.

The model, like everything else, is just another context-builder.

I think the fact that guys like Marques Colston, Brandon Marshall, and Mark Anderson stood out strongly within the model speaks highly of it but they weren't the only ones.

For instance, Mile Austin really stood out, and all he became was a pretty good return man for the Dallas Cowboys.

Then again, he was generally regarded prior to the athleticism tests to be an undrafted prospect lucky to make a roster and his becoming a good kick returner (a position where athleticism is at a premium) may also speak well for the model.

Anyhow, I can't model anything until the Combine has come and gone.

What I can tell you is that the model weighs about 22 factors, from about 15 or so measurements, and is a percent-over-adjusted-median model such that you are able to exert more control over the significance of variations within the talent pool of certain measurements.

For instance, there are guys out there in the LB pool that may be able to throw the 225 lbs bar up 40+ times, and other guys that only throw it up 20 times. If the median is 20 times then you've got a 0.0% grade for the one guy and a 100.0% grade for the other guy. That's all fine and good but if you stacked (added) that next to, let's say, a hand measurement where the median hand size is probably going to be like 9.75 inches and the biggest ANY person's hand is going to be is 11 inches (huge freaking hands) then in essence the world-beating bench press mark ends up inherently like 8 times more weighted than the world-beating hand measurement.

Is that fair? Obviously many would argue that bench press is more important than hand size. I would agree with them. But 8 times more important? I would prefer to adjust the medians such that the standard deviations within the talent pools for each measurement are all the same, which I feel gives you a more apples-to-apples method of addition...and from THERE you can sort of decide for yourself how much to weight each of the stats.

I ended up with decile grades. Now obviously, none of it had ANYTHING to do with ANY kind of field grading. Like I said, it is a context builder. Marry field production and instincts with a solid athleticism score and you usually don't go wrong unless something sneaks up from the periphery (like, for instance, the guy going to jail).

 

[TampaFinsFan01]

The only thing I would add to your list is an incredibly strong 'will to succeed.' IF a player has that determination they can do a lot even if they don't meet all the normal 'stat' requirements. Exhibit A would be Zach Thomas.

Zach Thomas was a tackling maching at Texas Tech. He had 22 tackles in a bowl game vs Air Force. Stats would have helped his cause. He was hurt by physical measurables.

I would only say to this debate, I think too much is made of workouts after the season is over.

I try to watch as much football as I can during the season and evaluate certain players from the games I watch. When I see a player go from a middle round grade after the season and then shoot up to a first round grade after the combine, I grow suspicious. Workouts should never trump previous performances, except in rare situations (injured player, small-school player, a WR at a running school ala Keith Jackson at OU). Most of the time a guy in the NFL isn't going to perform better than he did in college.

Exhibit A, Chad Jackson, the first round WR taken by the Patriots. After the season, he was probably about a 3rd round pick. He was good at Florida, certainly nothing special. Yet his 40 time at the combine was just stupid-fast. His agent used that momentum and got him picked higher than he probably should have been picked.