Study of Draft Prospects with Stats

[jct]

I Love the work you've done here. It is a basis to start to figure out who the pats type guys are. Also we get a picture of sleepers. Thank You.

 

[Mainefan]

I, too, am amazed and grateful for your work, Pony. I hope BB and SP are reading it.

However, I do not have total confidence in your methodology. There are just too many factors that are not within a player's control, too many circumstances, including differing competitive levels, quality of one-on-one matches, minor injuries, weather, traction, home and away variations. I think you quickly assume equivalents that may not exist.

I don't mean to dismiss everything you've written. Quite the opposite. I think it is very provocative and very useful. In the case of some players, it is very revealing--especially in cases of low productivity.

But, like others, I wonder how some current NFL stars would rate in your measurements. Vrabel and Bruschi, for instance. Can studies such as yours reveal intelligence, instinct, the capacity for growth and, above all, heart? I don't think so. And yet these qualities are usually the difference between average performers and stars.

 

[PonyExpress]

I, too, am amazed and grateful for your work, Pony. I hope BB and SP are reading it.

However, I do not have total confidence in your methodology. There are just too many factors that are not within a player's control, too many circumstances, including differing competitive levels, quality of one-on-one matches, minor injuries, weather, traction, home and away variations. I think you quickly assume equivalents that may not exist.

I don't mean to dismiss everything you've written. Quite the opposite. I think it is very provocative and very useful. In the case of some players, it is very revealing--especially in cases of low productivity.

But, like others, I wonder how some current NFL stars would rate in your measurements. Vrabel and Bruschi, for instance. Can studies such as yours reveal intelligence, instinct, the capacity for growth and, above all, heart? I don't think so. And yet these qualities are usually the difference between average performers and stars.

Mainefan,
Your suspicion of numbers and the people who manipulate them is healthy and patriotic. But 40 times, heart, instinct, desire, vertical leap, strength, quickness and teamwork only have meaning when they are expressed on the field in the form of a play. That play for a defensive player can take the form of a variety of things: a tackle; a pressure; a penetration; tying up blockers; an interception, a pass break up, etc. many of which are registered by score keepers as statistics. Statistics themselves are reasonably objective records of the expression of heart teamwork and desire on the field of play. I think over a two year period the portrait of a player can be realized more or less through his expression in numbers. There are exceptions, but generally that is true. That doesn’t mean that highly productive players in college will excel in the NFL, but it makes us examine why they wouldn’t more closely. And it doesn’t mean that comparatively unproductive players won’t excel in the NFL, but it makes us question why they would. As far as Vrabel and Bruschi are concerned, have no fear: each of them grade out very well statistically compared with their peers, for the same reasons we admire them: they make plays.

 

[big mike]

It sounds like the statistics aren't telling you what you want them to. You say Carpenter's production was limited due to injury and "other issues". What are those "other issues"? This is why a study of production is useful, forcing you to ask these questions and separate fact from fiction.

Ok, I should have just said "injury" (I meant games directly missed due to the injury, and then playing injured, which should be expected to bring down a player's stats).

Rodney Harrison only had 16 tackles last season. Erik Coleman of the Jets had 122. Do you feel those stats adequately represent how good Harrison is in comparison to Coleman?

A statistical study is just a tool, helping identify questions that need to be asked and answered.

True. What I'm saying is that it needs to be acknowledged that such a study can only tell you so much, and is vulnerable to giving misleading impressions since there are many things a study doesn't take into account.

One CB may have very few INTs. That could make people think he's not very good - but what if he doesn't have many INTs because QBs almost never threw to the guy he was covering, since he does such a good job of blanketing receivers (and the CB on the other side of the field is weak)? How is that accounted for in the statistics?

If Carpenter is so talented, and I'm not saying he isn't, why did he only make 7 tackles for loss in two years against the run while playing in the Big 10? That question needs to be answered by carefully watching ALL his game tape. Why did he accumulate so few tackles? My methodology is certainly open to criticism and I can be accused of making some arbitrary decisions. But attacking my methodology does not absolve Carpenter from having to answer some questions about his play. That is why these studies can be useful if their limitations are understood.

I'm not attacking your study. I'm saying there are problems with ranking players based solely on statistics.

 

[big mike]

PonyExpress,

I applaud your hard work and the results you have attained. My doubt regarding the validity of the counclusions implied by your stats is similar to my doubts of the "football scientist" on ESPN. Football is a game where teams have entirely different defensive systems with entirely different responsibilities for players who seemingly play the same position. That's why comparing 2 guys who are considered ILBs may seem like an apples-to-apples comparison, but when you dig down under the covers, it really isn't.

Certainly, a player's statistics are an extremely important part of his "resume" when considering where he should be drafted. But those statistics should be considered in the context of the team's defensive scheme, surrounding players, and even opponents. For example, a great CB will not have great stats if the CB on his other side stinks, or if his DL stinks against the run and can't rush the passer, or if they play a third of their games against option teams.

Exactly. Well said.

 

[PonyExpress]

Big Mike,
I believe you are missing the point here. I never claimed this survey had anything to do with talent, only with production. My rationale is that athletic talent only matters when it is expressed as production. Furthermore, durability of a college player is itself a kind of athletic talent and is always factored in by pro teams when scouting college prospects. After all, they pay an NFL player whether or not he is injured. As a side note, Carpenter was not the only player among those surveyed to play injured and miss time, and all of them played an extrememly physical position. As far as Rodney Harrison and Eric Coleman, the Rodney of 2003-4 was among the best players in the NFL, and that's how we remember him. The Rodney of '05 suffered a devastating injury and it is unclear if he'll even play this year (our prayers are with him). Eric Coleman is probably an average safety (I haven't studied it) and an average, healthy safety is bound to be more productive than a Hall-of-Famer on the injured reserve.

 

[Pisa]

Enjoyed the effort to document production

It's awfully hard to count on numbers for the truth when we don't know what responsibility the player was given. A couple of years ago the Rams drafted Robert Thomas, a small quick linebacker who would attack seams in the opponents line. He was sometimes overpowered and sometimes hit the wrong hole. However, he also ended up with a tremendous number of tackles behind the LOS due to his style of play.

You list Huff as a CB here even though he played 35 of 38 games at safety.

Also the tackle positions are not equal. A guy like Watson at Michigan plays on the shoulder of the center and gets double teamed every play. The other tackle is much wider and has many more opportunities to make sacks and tackles behind the LOS. That does not mean Watson stinks; Michigan may want him to create a pile in the middle of the field. I am very suspicous when I see a nose tackle with numerous tackles. He may be failing to carry out his primary mission.

Carpenter is hard to evaluate whether you use numbers or combine type numbers. He played SLB, played down DE on passing downs and was injured for the last part of the season.

Also did the players appear in the same number of plays. Some coaches stick faithfully to a rotation system which limits the number of opportunities for everybody. Teams with smaller rosters may have their stars out there for 75% of the plays.

 

[PonyExpress]

Pisa,
Those are all very fair criticisms. As far as Gabe Watson is concerned, how was his role different from Haloti Ngata? Haloti Ngata was continuously double teamed and still ended up as the most productive defensive tackle in the survey, while Watson finished almost last. I don’t think Carpenter was as unique as some people suggest. Wilkinson of GT also was tasked with some pass rushing responsibilities, as was Demeco Ryans among others. I explained earlier how I treated his injuries (remember he only actually missed two games) along with the injuries of other players. As far as Huff goes, although he started games as a safety he was given many more man to man coverage responsibilities than other safeties, which accounted for his Div I leading 28 pass break-ups over the last two years, comparable with the elite CBs. In comparison, Donte Whitner had 8.

 

[Mainefan]

Thanks for your response Pony. As I said, I hope SB and BB are reading this. At the very least, it will provide them with some hard questions to ask players who have big names but little production.

 

[Flying Fungi]

bump......

 

[PonyExpress]

Just to illustrate some value to this production study...
Last year some people said they were amazed that Lofa Tatupu turned out to be so good. "How did we miss him?" they asked. "What a steal!" Look at his measurables... he was slow, a 4.8 40 for a LBer.
But these are his production numbers over his last two seasons at USC, 2003 and 2004:
Total tackles: 202
Tackles for Loss: 25
Sacks: 9
INTs: 7
Forced Fumbles: 4

Lining up his statistics with this year's class, he would have come out tied with AJ Hawk for the #1 production ranking. So Lofa Tatupu's excellent production at the pro level should not have been a surprise. And his sub-par height, weight, speed etc proved irrelevant.

 

[bucky]

What about Rod Davis? He had pretty good college stats, didn't he?

 

[PonyExpress]

What about Rod Davis? He had pretty good college stats, didn't he?

Good point Bucky. But Rod Davis didn't play in a BCS conference; southern Miss is in Conference USA. Also, I never claimed that all productive college players become productive pros; Only that most productive pros were productive college players.

 

[drpatriot]

Just to illustrate some value to this production study...
Last year some people said they were amazed that Lofa Tatupu turned out to be so good. "How did we miss him?" they asked. "What a steal!" Look at his measurables... he was slow, a 4.8 40 for a LBer.
But these are his production numbers over his last two seasons at USC, 2003 and 2004:
Total tackles: 202
Tackles for Loss: 25
Sacks: 9
INTs: 7
Forced Fumbles: 4

Lining up his statistics with this year's class, he would have come out tied with AJ Hawk for the #1 production ranking. So Lofa Tatupu's excellent production at the pro level should not have been a surprise. And his sub-par height, weight, speed etc proved irrelevant.

This is very interesting. Where would Shawne Merriman rank?

 

[PonyExpress]

Merriman's two year numbers were:
140 total tackles
27 Tackles for Loss
17 Sacks
5 passes broken up
3 forced fumbles
Compared with this years crop of DEs, Merriman would have finished 6th out of 19 in production behind Kiwi, Lawson, Williams, Tapp and Hali.

 

[drpatriot]

Merriman's two year numbers were:
140 total tackles
27 Tackles for Loss
17 Sacks
5 passes broken up
3 forced fumbles
Compared with this years crop of DEs, Merriman would have finished 6th out of 19 in production behind Kiwi, Lawson, Williams, Tapp and Hali.

Now THATS interesting. The question here is where all five of these players are better than Merriman (doubtful) or whether 3-4 linebackers are more difficult to measure due to the position switch (more likely).

 

[PonyExpress]

Now THATS interesting. The question here is where all five of these players are better than Merriman (doubtful) or whether 3-4 linebackers are more difficult to measure due to the position switch (more likely).

I still think 6th is very respectable. I phrased it badly in my last post; Merriman wasn't 6th out of 19, he was 6th out of all DEs in the nation. Pretty damn respectable. On the other hand, Michael Haynes the DE out of Penn State taken by the Bears with the 14th pick of the 1st rd in 2003 (the Pats pick; we traded up one slot with the Bears to take Ty Warren) had college production similar to Tamba Hali, and Haynes has been a bust. It just underscores the point that not all productive college players become productive pros, but almost all productive pros were productive college players.

 

[PonyExpress]

Just to continue the Merriman point, when you have great college production and also great athletic measurables you have a potential great pro. I think teams get into trouble with players like Wimbley who have marginal production and good measurables. The team that drafts him in rd 1 is courting disaster. IOW, I see him going to the Browns.

 

[Box_O_Rocks]

Just to continue the Merriman point, when you have great college production and also great athletic measurables you have a potential great pro. I think teams get into trouble with players like Wimbley who have marginal production and good measurables. The team that drafts him in rd 1 is courting disaster. IOW, I see him going to the Browns.

Nice fantasy, but I'm assuming the Browns are made of sterner stuff under RAC and Savage. If one of the elite players or Ngata doesn't drop to them to tempt them then Lawson is gone and Wimbley can go to some team less astute in their judgements.

 

[PonyExpress]

How good is Mario Williams? Is he really as good as Julius Peppers was? Production wise they are almost exactly the same, except in pass defense. Peppers' (who also finished 2nd among DEs in the NFL) strength is not only his ability to rush the passer, but to drop into coverage, or chase down a receiver and break up and even intercept passes. This shows how athletic he is.
Over his last two years in college Peppers had 127 total tackles, 43 tackles for loss, 25 sacks, 11 batted balls and 4 forced fumbels.
Mario Williams had 119 total tackles, 39 tackles for loss, 21 sacks, 4 batted balls and 1 forced fumble. Peppers also had 4 ints. Whatever they tested as at the combine, Peppers was much more athletic on the field. He had a better motor and used his speed against the pass in all phases.