Writers of Pro Football Prospectus 2008

10 Apr 2008

FO on ESPN: Combine Stats Not Created Equal

This week's FO feature on ESPN is me talking about running backs at the combine and what their performance there has to do with NFL success. Obviously, we'll discuss this more in the book.

Posted by: Bill Barnwell on 10 Apr 2008

54 comments, Last at 14 Apr 2008, 11:53pm by Mr Shush


by Criswell (not verified) :: Thu, 04/10/2008 - 11:23am

Very interesting. I wonder how the adjusted 40 times of the other top running backs in the draft compare to McFadden's.

by parker (not verified) :: Thu, 04/10/2008 - 11:25am

Somebody go back and tell Dallas not to draft Emmitt Smith in the first round.

by parker (not verified) :: Thu, 04/10/2008 - 11:27am

Just curious: Did you guys try it with an adjustment for height as well. After all 68in and 210 is alot different than 74in and 210.

by MJK (not verified) :: Thu, 04/10/2008 - 11:31am

Nice research. My thoughts here aren't criticisms--these are exactly the sorts of articles that I love from FO. But good research does get me wondering about things...

A question--the 40 time is raised to the fourth power? So the metric that correlates best with ultimate performance is actually an inverse QUARTIC relationship? That's a very strong function--it implies that, for backs of the same weight, being just slightly faster could lead to a dramatic increase in adjusted 40-score.

Which makes the problem highlighted by Westbrook even worse. 40-times are notoriously inconsistent, depending on the timer and how fast a guy felt on a given day. With such a strong relationship, I wonder if the noise from the test data is swamping actual trends... If a player's test has a little bit of error, teams could draw just the wrong conclusion...

(Remember with Westbrook, also, that he is not used like a traditional RB. He's used a little more like Reggie Bush ought to be used--a kind of hybrid between an every-down back and a 3rd down receiving back. How many RB's come close to being the leading reciever on their team?)

One final thing I wonder is how much the 40-time affects a player's success not because it indicates his actual skill, but because it influences his draft position, and teams are less likely to give up on a high draft pick than a low one. In other words, if player A has a fantastic 40-time and ends up being taken in the late 1st round instead of the high 3rd round where he may have been projected before the combine, then the coach will feel pressure to play him as the starting back. Even if he struggles, he will continue to get a lot of carries (which was one of your metrics, and which plays into DPAR), and may eventually prove himself. On the other hand, if the same guy has a bad 40-time, he may slip to the fifth round or later, and if he struggles, the coaches will have no compuction about benching or cutting a low-drafted guy.

by MJK (not verified) :: Thu, 04/10/2008 - 11:33am

Good point, parker.

Maybe adjusting for BMI (which is a formula that adjusts weight for height) could give an even stronger correlation.

by Mr Shush (not verified) :: Thu, 04/10/2008 - 11:41am

"McFadden's adjusted 40 score is a superb 120, 10th best in the past decade. Complaints about his weight or body mass index (BMI) being too low can be ignored, since neither has any sort of relationship with NFL success."

Given the PFR blog entry linked on my name, I'm slightly surprised to find BMI had no relationship with NFL success. How did you check that? Did you try using BMI in place of weight in creating your adjusted 40 time measure? And is it possible that a connection along the lines of "very low BMI makes sustained success as a full-time starter almost impossible, but otherwise BMI is irrelevant" could be lost in a regression analysis (which I assume is what you did)? Interesting article, though.

by Joey Jo-Jo Junior Shabbadu (not verified) :: Thu, 04/10/2008 - 11:55am

MJK - you have a good point regarding the impact of "noise" in the data due to manual timing. I guess that's unavoidable given the current data source. However, in this case, it appears that an inverse quartic is necessary to show the relevance of 40 times.

It's safe to say that pretty much every draftable RB runs between a 4.3 and a 4.7. But nobody would say that the speed of a back that runs 4.3 is 9% better than the speed of a back that runs 4.7, as the raw numbers would imply; rather, a guy who runs 4.3 is WORLDS better.

To illustrate a real-life example, McFadden's excellent 4.33 time, raised to the 4th power, is 351.5, whereas Westbrook's mediocre 4.57 time, raised to the 4th, is 428.6. The 22% difference between these numbers seems fair given how these numbers are typically treated in the scouting process.

by Ted Max (not verified) :: Thu, 04/10/2008 - 12:18pm

Interesting stuff.

By the way, a correlation of .45 means that more than 75% of the reason why running backs are successful is NOT due to whatever 40-run times indicate. So 40 times are a great deal in terms of giving you 25% of the explanation with a simple test. But they are far from the whole story, which most of us would agree with.

One question: Are any of these things "partial" correlations (meaning, controlled for the other measures)? It'd be nice to know if, for instance, the general athleticism shown by leaping drops out when you control for 40-times, since it might indicate more or less the same thing.

by ebongreen (not verified) :: Thu, 04/10/2008 - 12:26pm

I'm not surprised at the finding that speed matter, or that raw speed needs to be modified by a player's weight. I'm a little surprised that it's not a player's MASS that's being discussed rather than his weight, this being a site for science & math nerds.

I don't think BMI is going to get you that far as a correlation goes, because most athletes have very strange BMI relative to the population as a whole. What this boils down to, in my eyes, is a player's "explosiveness index": how many newtons (kg*m/sec) can he produce in a short amount of time? In physics terms, how much "work" can he do each second?

I'd be interested to see how that figure matches up with the adjusted 40 score mentioned in the article; I think it's a little easier to digest. :-)

by Will B. (not verified) :: Thu, 04/10/2008 - 12:36pm

The Dolphins should move Ted Ginn to halfback. He has good 40 times, and BMI doesn't matter.

by john (not verified) :: Thu, 04/10/2008 - 12:46pm

Very interesting analysis, but one question (and this is not intended as a criticism in any way): Why not use pro day data?

It would seem that the only significant difference between the drills prospects run at the Combine and the drills prospects run at pro days is the venue. Granted, this is not necessarily an insignificant difference, as pro day prospects are no doubt more comfortable and relaxed at home and there may be a faster track surface. But a 40 yard dash is a 40 yard dash, whether it's at the Combine or somewhere else, isn't it? I'm not sure pro day data is so tainted that it had to be excluded from your analysis.

Actually, it would be interesting to run the pro day numbers in the same way and see if they support your findings from the Combine, which I would imagine they will.

by John Morgan (not verified) :: Thu, 04/10/2008 - 12:56pm

Extending on what Ted Max said, a .45 correlation is especially not very strong when said group is already being evaluated and ranked because of their speed/size. A player who is both large and fast is likely to be drafted higher and get more opportunities to accumulate yards, carries and DPAR.

Also, 3 of your 6 examples of players who did poorly in this metric and failed as running backs, failed for every reason but their speed/size. William Green and Chris Perry were both successful rushers. Green's DPAR wasn't very good, but neither was Cleveland's run blocking. "Weedhead" William Green couldn't keep it together off the field. His career ended because of injuries. Perry's career has also been wracked with injuries. Maurice Clarett currently resides in Toledo Correctional Institution. Unless the metric measures vulnerability to injury, lawless and the probability one's wife might stab them in the back, I don't think these three make good examples.

All in all, this seems a little shaky to me. Naturally, being big and fast are good traits for a running back, but does this metric provide insight beyond common knowledge, common sense?

by Will B. (not verified) :: Thu, 04/10/2008 - 1:13pm

Well, it does tell us that being agile isn't particularly important (unless your are Barry Sanders).

by mattman (not verified) :: Thu, 04/10/2008 - 1:15pm

An interesting study, but kind of disappointing that it focused solely on McFadden. Are you going to post adjusted 40 scores for other running backs in the draft? And was vertical jump factored into the ratings at all?

by Joey Jo-Jo Junior Shabbadu (not verified) :: Thu, 04/10/2008 - 1:16pm

...or that we're using the wrong tests to try to measure agility.

by Bill Barnwell :: Thu, 04/10/2008 - 1:28pm


McFadden 120.05
Stewart 116.68
Mendenhall 114.76
Charles 108.68
Jones 103.7
Johnson 121.91
Rice 99.69

#2: You must've missed the paragraph where I talked about Westbrook. Or you're a bad troll.

#3/other BMI: I did try running it with BMI in the numerator as opposed to weight. Across the board, it bore less of a relationship.

#4: I think that's a very valid point -- inherently, we're going to get some issues like that when we're basing some sort of projection upon five seconds of a human being's life. When you talk to people at the combine who are evaluating people, what they talk about when it comes to interviewing a player is making sure he's prepared -- not necessarily with the answer that he gives being very important, but that he is ready for whatever you're going to ask and knows how to do so in the proper fashion. I think the same thing is true, to an extent, for something like 40 time.

There's also a valid point that what we might be tracking here is simply who teams like best. That being said, even when we're looking strictly at a pool of draft picks (as opposed to all combine attendees), or at Day One picks, or just first-round picks, the adjusted 40 has a stronger relationship across the board than the actual 40. Factoring in that with the success of late-round guys with high adjusted 40's like Ryan Grant make me feel pretty confident that there's something more to this than just identifying who teams like.

#11: Because they're different animals. In the research that we've done with a limited sample of Pro Days, guys average four-hundredths of a second faster 40 times on Pro Days than they do at the Combine, but the variance across the sample is so high as to make that obviously just noise. Some guys are four-tenths of a second faster, some are four-tenths of a second slower. They're run on different tracks, with different speeds, and have totally different scenarios. The combine is the equaliser, and it's a much more comfortable thing to qualify numbers produced in the same environment. Otherwise, we have to control for how players do at each individual Pro Day for that college relative to other players from that college, and then across the nation...the work might be worth it for an NFL AGM, but not for me.

#12: William Green had numbers comparable to his backup or teammates each year he played, which suggests that he wasn't worthy of a first-round pick (who would, theoretically, outplay his backup who'd been selected in later rounds). Unless you want to argue that the Browns offensive line was so shitty as to neutralize all running back play, which I can't prove or disprove.

It's entirely possible that one of the things we could be tracking in the failures here is an inability to keep up with the NFL workload.

Clarett was cut before he went to jail. One had nothing to do with the other.

by Ted Max (not verified) :: Thu, 04/10/2008 - 1:29pm

#12 makes a lot of sense to me. 0.45 certainly isn't an especially high coefficient, but it depends on what you are comparing it to. It's much better than nothing, for instance, which might be all that scouts' "gut instincts" get you.
And yes, it makes no sense to talk about speed as a predictor and cite examples where speed wasn't a factor. Unless we assume that if Clarette had been fast enough, he wouldn't have been a nutjob.
#14 and #15 are also interesting. It could be that leaping or shuttle runs (as measured at the combine) don't really matter, or it could be that the effects of 40 time are swamping them. That's why it'd be nice to have one big analysis (like a multiple regression) that includes all of the relevant factors at once, controlled for each other.
Still, while any analysis can be improved or extended, ANY analysis is better than "gut instincts" about player with nothing to back them up.
So, thanks for the article.

by Oh, Mathematics (not verified) :: Thu, 04/10/2008 - 1:58pm

Re: 9

Considering that all NFL games are played on planet Earth, I don't think the difference between mass and weight will give any more insight.

by TomHat (not verified) :: Thu, 04/10/2008 - 2:00pm

very interesting read. I always thought the combine numbers meant nothing. turns out that one of the numbers matters and none of the others do.

Also, Re Smith: Clearly even a single outlier makes the entire correlation totally worthless.

by justme (not verified) :: Thu, 04/10/2008 - 2:30pm

Very much enjoyed the article, sir - looking forward to more in the book.

by Pat (not verified) :: Thu, 04/10/2008 - 2:56pm

(Remember with Westbrook, also, that he is not used like a traditional RB.

In previous years, sure. But Westbrook led the league in rushing DPAR last year and came darn close to leading the league in rushing last year. Safe to say he's being used like a traditional and a non-traditional RB now.

by johnt (not verified) :: Thu, 04/10/2008 - 3:26pm

0.45 isn't high if you're performing an experiment where you legitimately expect the outcome to be predicted by the independent variable. In the real world, there are psychologists and economists who would kill to get their hands on a 0.45. Remember, if you've truly got a high R value you're basically saying there's no room for chance, it's almost impossible for a guy with great physicial attributes not to have a successful career, which is obviously absurd.

by Nick (not verified) :: Thu, 04/10/2008 - 3:48pm

Re: 17

One of the reasons Clarett was so slow was that he was out of shape after missing multiple years of football. He had posted faster (non-combine, obviously) 40 times prior to entering college. His rivals profile listed him at 230 lbs and a 4.5 40, which gives him an adjusted 40 score of about 112. Compare that to his combine time of 4.72 time (while weighing 234 lbs) and his adjusted score drops to about 95.

In Clarett's case at least, his being a nutjob led to a bad 40 score, and also led to his being a pretty huge disaster as an NFL player.

by bravehoptoad (not verified) :: Thu, 04/10/2008 - 4:10pm

Considering that all NFL games are played on planet Earth, I don’t think the difference between mass and weight will give any more insight.

You're forgetting that certain games are now played in Europe.

by MJK (not verified) :: Thu, 04/10/2008 - 4:20pm

I think it's important to remember that the message of this article doesn't seem to be that teams should base all their draft decisions on adjusted 40-times, but rather, of all the tests done at the combine, the 40 is probably the most relevant (for RB), and, when you adjust the results for weight, it gives laymen (rather than scouts) a metric they can use to predict a guy's performance.

But I also buy the argument that at least some of the predictive power of the 40-time is due to the fact that the 40-time is a rough way of measuring how well a player prepared, not just their natural athletic ability. It makes sense that a player, like Clarett, who lacks the dedication to be a good NFL player would also lack the dedication to prepare for the combine, and hence his bad 40-time may be correlated to his likliehood of being a bust, for reasons other than his athletic ability.

Remember, the article is arguing that 40-time correlates with success, not that it is a causitive factor or even the thing it ostensibly measures (sprinting speed) is a causitive factor. Correlation is not causation.

by Joe T. (not verified) :: Thu, 04/10/2008 - 4:21pm

You know what I'm curious to see? How much wonderlic scores are indicative of future success at the "heady" positions.

Good article. The next logical step is, why are 40 times an indicator of success? I saw some decent explanations in the columns thread. Also, does this hold true for WRs and DBs, the real burners on a team?

by cjfarls (not verified) :: Thu, 04/10/2008 - 4:30pm

Truly excellent... This is why I love FO.

When paired with the Lewin QB projection system, we're finally getting draft analysis that is more based on real data than hype and the "talking heads".

I can't wait for a few years from now when we can test these predictions against the "mainstream" picks. I'm guessing that like where Slate's assessed season predictions and in which DVOA, etc. has outperformed the "experts", these tools will be valuable in creating better draft assessments than the "talking heads" as well.

One question for Bill - Did you experiement at all with limiting the sample size to account for NFL scouts judgement? A.k.a. use the Lewin "first 2 rounds" adjustment (or first 3, etc.)? Is there something there that might improve the fit of this model?

by langsty (not verified) :: Thu, 04/10/2008 - 5:11pm

4: "(Remember with Westbrook, also, that he is not used like a traditional RB. He’s used a little more like Reggie Bush ought to be used–a kind of hybrid between an every-down back and a 3rd down receiving back. How many RB’s come close to being the leading reciever on their team?)"

This kinda dovetails with what I was thinking about when reading the portion on BMI/weight not making a difference - Westbrook is actually a very effective inside runner who can also be used as a satellite/motion type of guy in the passing game. Bush is very much the latter, not so much the former, and I think a lot of it is that he lacks the sizable lower body thickness of a Westbrook. BW has always had a huge ass and thighs and fits a lot more naturally into the feature back role.

How this all relates to McFadden, I have no idea. I can't recall ever seeing an athlete in football with his exact size, skillset and bodytype. I tend to think he'll be succesful on the next level, but his chicken-legs would worry me a little.

by Tom D (not verified) :: Thu, 04/10/2008 - 6:00pm

Re 24:

I laughed for a good 5 seconds, bravo sir.

by VS (not verified) :: Thu, 04/10/2008 - 6:33pm

Sorry guys, you're still my favorite writers but I don't like this one at all.

By using rushes, rushing yards, and DPAR, volume is basically the only measurement of supposed success. The 40 time has been the de facto indicator of draftability recently, so it's no surprise that players with better 40 times are going to be given more opportunities to rush the ball. Teams have never been very eager to admit they've made a mistake on an early pick, no matter how much better the backup may be. To me, this just confirms that high draft picks get more playing time than late picks...not a tremendous surprise.

In my mind, the fact that vertical leap appeared with such a correlation despite teams basically ignoring it on draft day would be the result worth investigating further.

PS: Any way I can work for you guys this summer :)

by Jason (not verified) :: Thu, 04/10/2008 - 7:33pm

Where is the list for the 9 other players this past decade with a number higher than McFadden?

by Bob in Jax (not verified) :: Thu, 04/10/2008 - 8:00pm

Excellent work. I would tend to echo #9 (ebongreen) that "explosiveness" is a good thing in football (or most athletic endeavors, really) that is mostly related to 40 time. Most sprinters know that the race is really won and lost in the start when the distance is so short, and explosiveness would produce a faster start when used properly (hence, preparation).

I also want to congratulate the other commenters for the sustained excellence of the discourse in this comment thread, and most of the other threads as well. Every so often, I forget and scan the comments on Fox Sports, or Yahoo sports, or some other such, and then remember that I like this site for more than just the excellent articles. The discussions are enlightening in and of themselves (and funny -- e.g. #24), and always a good read.

Keep it up, all.

by chip (not verified) :: Thu, 04/10/2008 - 10:40pm

#16 Bill - Are you really saying that Johnson is the best back in the draft or is that a typo? The metric appears to be listed in descending order, yet Johnson is in there sixth with a 121.9.

For as much as people bitch about the talking heads, the list (ex-Johnson) tends to match up with most draft pundits. Now, if the 121.9 is right, that is something to talk about.

by Jon (not verified) :: Thu, 04/10/2008 - 10:43pm

I'm a little surprised by these results. 40 times certainly matter for RBs, you want him to have that extra gear if he breaks free. Intuitively, I'd expect some of the other events to be more important.

The Giants for instance, they seem to really heavily weigh the 3-cone drill. Bradshaw has a disappointing 40-time (and also fell for other reasons), but he had one of the best 3-cone times last year. The technique for running a 40-yard dash doesn't necessarily translate perfectly to in game speed.

by Jon (not verified) :: Thu, 04/10/2008 - 10:46pm

33: Johnson certainly is very underrated because he hasn't really played in the spotlight that much. I caught ECU's bowl game, and he was just running circles around Boise State. He reminds me a lot of Jerious Norwood - who was stuck on a bad team, but had tremendous production in the SEC, and great workout numbers. The one worry about Johnson is that some teams have red-flagged his knee I believe.

by Bill Barnwell :: Thu, 04/10/2008 - 10:59pm

NO. I'm not saying Chris Johnson is the best back in the draft because he has a higher speed score than McFadden. There's much more to projecting performance than a speed score. What I'm saying is that they both have the pedigree of successful NFL backs.

Top ten speed scores:

- Brandon Jacobs (123.5)
- Kevin Jones (123.86)
- AJ Harris (123.36)
- Jackie Battle (123.14)
- Chris Henry (122.73)
- Justin Fargas (122.33)
- Chris Johnson (121.91)
- Fred Staton (121.44)
- Ronnie Brown (121)
- Darren McFadden (120.05)

by chip (not verified) :: Thu, 04/10/2008 - 11:12pm

#36 I should have finished my thought... "best back in the draft on this metric".

Interestingly, Matt Forte posts a 112.2. Sounds like a 1st round pedigree at a late second round price. Maybe I shouldn't be so suspect of Jerry Angelo's interest in Forte.

by Lou (not verified) :: Thu, 04/10/2008 - 11:46pm

This research is really fantastic. I heard an interview with Bill Polian a couple years ago before the draft where someone asked him whether combine stats where an important factor in making draft decisions, and his answer lead me to believe that they had done some sort of regression analysis between those numbers and players success in the nfl.
Its really awesome to see FO look at this. I imagine the same analysis could be done for WRs, don't know how you'd do it for positions that don't have decent stats though. As someone else mentioned it'd be cool to see the wonderlic incorporated.

by VS (not verified) :: Fri, 04/11/2008 - 12:48am

You could just look at the combine stats compared to games started or % of team's plays played(if that is even available which I'm pretty sure it's not but it'd be cool) for the rest of the positions.
I'd be very interested in seeing how different this would look if you just used the playing time metric for the RBs as well.
Can I ask why you picked # of rushes as a measure of success? That seems like it would already be very accounted for based on both the yards and DPAR.

by Purds (not verified) :: Fri, 04/11/2008 - 9:55am

I want to speak to the question posed about judging a person on one 5-second moment in his life, his 40 time at the combine. While in general I agree that we don't want to thin-slice like this for all, I would offer that preparation for that 5-second moment of a RB's life can highly mimic the preparation he'll put into everything: meetings, practices, learning plays, etc.

This second point I make with this background: I coach track and am a former collegiate hurdler. Although an athlete can have an "off day" or an "on day," in reality track performances are one of the least "lucky" performances in sport. Track athletes get no lucky bounce of a ball off a rim, no hitting a tree and landing in the fairway, no wind blowing the ball out of the park. In many ways, track is the truth. Did you prepare? Are you focused? Do you have talent?

So, while certainly some guys have 40 times at the combine that won't predict their future success, I would attribute that discrepancy to other factors, not so much a "good day" or a "bad day" on the track. In relative terms, very, very few athletes perform on the track outside of the limits created by their talent, their preparation, and their focus. And even fewer perform outside of those limitations on big meet days, which the combine would be for the RB. (Of course, injuries and illnesses can be mitigating factors.)

by Jimmy (not verified) :: Fri, 04/11/2008 - 10:26am

If you asked me to make an intuitive guess which combine stat would correlate best with NFL performance by a running back I would have said the 20 yard time over the 40. I have no idea wether this information is released. Backs only get the chance to run 40 yards in a straight line maybe once a game.

by Herm? (not verified) :: Fri, 04/11/2008 - 11:30am

41 Jimmy
you could take your point a step further. Not many backs have a chance to run 20 yards. Even when lining up 8 yards deep, it is very rare to run that far....and even at that, not all plays are predicated upon running as fast as you can from a standing or 3-point start.
It would seem measuring accelleration and top speed are the 2 main goals here. The 40 gives you the best possible score (time) without having to invest in major technical / error reducing equipment.
But it is the NFL, they put so much money into high draft choices, they would benefit from it, and I'd like to see that investment. Maybe someone can hold a bake sale for them.

by RandomAbuse (not verified) :: Fri, 04/11/2008 - 11:38am


If I remember correctly Kirwan at NFL.com ran an article where he produced an "agility index" by adjusting 3 cone drill times for speed (indicated by 40 times).

I'm not sure this sort of analysis would be very effective for other positions though, there are too many different types of WR (pure speed vs excellent route runners) or corners (zone vs man) for it to be easily broken down into one simple metric. I am willing to be proven wrong though.

by Rich Conley (not verified) :: Fri, 04/11/2008 - 12:35pm

". In relative terms, very, very few athletes perform on the track outside of the limits created by their talent, their preparation, and their focus."

Except at the combine, and pro-days, where times are measured in a way that isn't close to repeatable.

by Purds (not verified) :: Fri, 04/11/2008 - 2:00pm

Rich, you're right in many ways about that, although the NFL is getting a little better with the introduction of one type of electronic timing (not the same as a track start, but at least objective). Problem is that not all reported times are the electronic ones. See the linked story.

by Herm? (not verified) :: Fri, 04/11/2008 - 2:44pm

I agree with your points, Purds, but I cautiously think there are exceptions to mention that affect results:

One is Biorhythmic cycle, as there are peaks and valleys that can affect your times/distances enough to affect a scout's decision.
Most athletes should but don't train to a schedule that allows them to peak at the most important times.

Also, there needs to be consistency of conditions and consistency of measure: same temperature, wind, surface, grade, etc...and same method of timing. There is a fantastic margin of error if you begin timing with a stop watch at the sight of smoke or first motion rather than using a linked timing system.

You also know that by nature, 40 times (and track results) are like the SAT's...recruits walk around telling their best performance as if it were representative of their typical performance, when in fact they may never repeat it again.

If I were to invest in a potential draft choice, I'd want to see multiple results, taken over a few weeks or even months, with the consistency I mention above.

And that doesn't even account for how we've mentioned that the 40 (even with weight factor) may not be the very best way to measure every position.

Sorry to get so wordy over a few 100ths of a second.

by Sophandros (not verified) :: Sat, 04/12/2008 - 9:11am

37: Which is why, when combined with his production, I'm at a loss as to why Forte's name isn't mentioned more often among the top backs in the draft.

After Tuesday, we should have a better idea about whether he may be a target in the early second round by his hometown team.

by Mr Shush (not verified) :: Sat, 04/12/2008 - 2:52pm

The position I would most like to see this analysis done for is DE (including guys who get drafted as 3-4 OLBs). Sacks would make a reasonable complement to starts as a counting stat, and I think there is evidence (Merriman, Ware, Williams) that great combine performance at DE is predictive of NFL success.

by chip (not verified) :: Sat, 04/12/2008 - 2:54pm

#47 Agreed. He'll go at the top of the second at the latest. The big three will go in front of him, but I bet he's fourth off the board. I hope Angelo has a crack at him in the second round - I'd rather have a QB, but Lovie would rather pound the rock to compliment his D instead of upgrading the passing game to the 21st century.

by Tom D (not verified) :: Sun, 04/13/2008 - 2:07am

Re 49:

"Lovie would rather pound the rock to compliment his D instead of upgrading the passing game to the 21st century."

This isn't entirely true. When Lovie first came to Chicago, he hired Terry Shea, the QB coach for the Chiefs, as his OC. He wanted to make the Bears the Rams north. Have a strike anywhere passing offense coupled with a turnover creating defense. Well, Shea turned out to be a complete idiot (calling 7 step drops games after it was obvious that neither the line nor the QBs could handle it, giving Thomas Jones the ball only 240 times when he was the only bright spot of the offense), and talent of our skill position players conspired to ruin his dream. So we became the 2000 Ravens instead.

by Disco Stu (not verified) :: Sun, 04/13/2008 - 5:36pm

Re #48- Mamula!!!

by Mr Shush (not verified) :: Mon, 04/14/2008 - 8:46am

Mamula wasn't as bad a player as most people think: he averaged nearly a sack every other game, and wasn't helped by the fact that he was drafted to play DE despite the fact that he was clearly better suited to being a 3-4 OLB. Obviously that's not what you're looking for when you take a pass-rusher in the top ten, but I'd characterise Mamula as a disappointment rather than a flat-out bust. His performance was, by analogy, closer to Eli Manning than to Ryan Leaf.

by Ben (not verified) :: Mon, 04/14/2008 - 3:47pm

48. My guess is for edge rushers you will find that KEI > 70 has a pretty strong correlation with success and the guys with the best VJ among those who hit the KEI mark are the safest bets.

by Mr Shush (not verified) :: Mon, 04/14/2008 - 11:53pm

"the guys with the best VJ among those who hit the KEI mark are the safest bets.

Hey, I like kitty as much as the next guy, but I've seen no evidence that would lead me to believe that it lead to better NFL performance. Indeed, the whole Ashlee Simpson thing appears to suggest the opposite. Unless kitty just leads to sacks, whoever's getting it.

Seriously, what is VJ? Oh, right, "Vertical Jump". Now I feel dumb. Yeah, ok, that reads like something sane now. To what extent are vertical leap and 40 time co-linear?