FanPost

Making sense of the numbers - the inherent flaws in explosion number and lateral agility

If you don't know that the Denver Broncos are in dire need of some talent at the DT position, then you probably are a Raiders fan whose greatest lifetime accomplishment was completing seventh grade.

I did a fanpost about a month ago looking at the top 6 DTs in this draft and comparing them to the top 5 from 2011. To continue my unhealthy obsession with the defensive tackle position, I expanded my scope to analyze the top 20 defensive tackles in this years draft. I also started looking back at combine results from previous years to see if things like Explosion Number (EN) and Lateral Agility (LA) could be used to predict NFL success. If I had more time, I would have continue to go back in time and see how predictive (or not) the combine results are. I found some flaws both the EN and the LA (inherent mathematical flaws - the issue of applicability will be left to others to debate).

Keep in mind that this is solely looking at combine results which really should not be used in a vacuum. In-game statistics and game-film MUST be used to temper the assessment of any college player. I am not advocating that combine results alone should be used. I am, however, advocating that if you are going to use EN and LA to evaluate the relative merits of player A and player B, then you should modify them in the way that I have.

Make the jump with me to see what I found...

A great deal of this data was found at nfldraftscout.com which (at least to me) is a treasure trove of historical combine data.

http://www.nfldraftscout.com/ratings/players.php?genpos=DT&draftyear=2012&sortorder=tsxpos&order=ASC

Explosion number tries to quantify how much upper and lower body strength a trench player has. The formula is

Bench reps + vertical (in inches) + broad jump (in feet) = Explosion Number

Generally a number of 70 or greater means that you are a great athlete. There are a couple of problems with trying to compare trench players with this formula.

The first problem relates to using bench reps at 225. Shorter arms mean that the bar has to travel a shorter distance for each rep. Hence, two players with the same level of strength but different arm lengths are going to put in different numbers in terms of reps on the bench. If you remember your physics, work = force x distance. Moving the same mass an extra 3 inches up for each rep results in significantly more work for longer armed players. So longer armed players get hurt in explosion number calculations. Longer arms are a plus for offensive and defensive lineman – two prime examples are on the Broncos, Doom and Clady. To correct this problem with explosion number, I have normalized the reps to the average arm length of the DTs in the study. So a player with 36" arms who puts up 35 reps gets a bigger benefit to his EN than a player with 32" arms who puts up the same number of reps.

2012dtnormbench_medium

So what does this mean when we look at the top 20 DTs in this draft (there were 22 taken in 2011)? Let’s look at the bench press numbers: Poe’s 44 reps don’t look quite so impressive because his 32 arms aided him. However, his normalized bench number of 39 reps is still damn good (and still the best results of any DT in this draft). Simliarly Ta’Amu’s 35 reps get normalized to 31 reps because of his 32 inch arms. On the other side, Hicks gets a boost because of his massive 35 and 1/8th inch arms; his reps normalize from 26 to 33. The other two guys in this draft who got a boost by factoring in arm length were Cox and Brockers. Cox, who has 34.5" arms, went from 30 to 36 reps and Brockers, who has 35" arms, went from 19 to 24 reps (which still shows that he is weaker in the upper body than a lot of the DTs in the draft). His normalized 24 reps were still lowest of the DTs. The devils advocate here can look back at Marcel Dareus combine results and say that bench reps (or upper body strength and endurance) are not a good way to measure talent. Dareus' arms were average length for the 2011 DTs and he only put up 24 reps, one of the lower values for the 2011 DTs. Dareus also didn't do the broad jump so his EN is meaningless. Dareus, rightfully so, felt his in-game stats and film proved his worth. Fairley did much the same thing when he declined to do the bench press at the 2011 combine.

The second problem with EN is that it doesn’t factor the player's weight into the vertical. A 350 lb guy with a 34" vertical means that the 350 lb player is much more explosive than a 290 lb guy with the same vertical. Again, it’s simple physics; more force is required to get a heavier mass the same distance above the earth than a lighter mass. A player who can generate more force is more explosive and should get a benefit to his EN. So I have normalized the vertical numbers (second column) to the average weight of the DTs in the study; bigger DTs with above average verticals get more credit than smaller DTs with above average verticals.

2012dtnormvert_medium

The two big "winners" here are the two largest DTs, Poe and Ta'Amu, both of whom got their verticals bumped up (by 4.0 and 3.7 respectively). Only two DTs had their verticals knocked down by more than one inch, Wolfe and Randall. They are two of the lightest DTs yet their verticals were not as far above average as the formula would expect.

So with both the bench and the vertical adjusted how did this affect the EN for the guys? I am calling the adjusted EN, the True Explosion Number or TEN.

2012dtsten_medium

The big "winners" are Hicks (+8.0), Brockers (+6.1) and Cox (+4.7). All three got most of their boost for the arm length normalization of bench press data. The biggest "loser" in this analysis is Mike Martin, whose TEN is 3.3 below his EN solely because of his short arms and he is slightly below average weight for this year's crop of DT (311). This should show that my study was not biased, because I really like Martin and hope the Broncos get him. As you will see below, his change of direction (quickness) is phenomonal.

From his combine numbers, Fletcher Cox does not appear to have a very powerful lower body (broad jump and vert). This reflects negatively both on his EN and his TEN.

The other side of the coin is quickness off the line, initial first step and ability to move in the trenches. This has been quantified using LA by many draftniks, but there is a flaw with LA.

LA = 40 time – 20 yd shuttle time

For LA, a higher number is supposed to mean that you are quicker in the trenches. The problem I see with LA is that it rewards a DT too much for running a slow 40, because you subtract the shuttle time from the 40 time. I think a better measure of quickness in the trenches is what I am calling the "Change of Direction" number, or CoD.

CoD = 20 yd shuttle + 3-cone drill

The 20-yard shuttle is designed to test lateral speed and coordination. The player starts in a three-point stance. When the whistle blows, the players run five yards to one side, touching the yard line. They then sprints 10 yards in the other direction and again touch the yard line, at which point they sprint back to the yard line they started from. So this test involves two full stop/180-pivot/start manuevers. The 3-cone drill is similar. Players start in a three-point stance in front of three cones that are set up in an L shape, with each cone five yards apart. They then sprint five yards to one cone, sprint back to the starting cone, and head back to the second cone where they run around it and cut right to the third cone. The players then run a circle around the third cone from the inside to the outside and run around the second cone before returning to the first cone. Both drills test foot quickness and the ability to start, stop, shift your weight and change direction.

Comp01_20shuttle_medium

Comp01_3cone_medium

Similarly to CoD, you could just subtract the 20 yard shuttle time from the 3-cone time, but I like the additive method better. Either addition or subtraction would be a better measure of lateral agility than the current formula for LA, because the spread of values are greater and the ability to differentiate A from B becomes easier. For CoD, a lower value is better.

2012dtscod_medium

For a guy like Fletcher Cox who ran a really fast 40, the LA is useless and makes him look like he has no quickness relative to Powell and Martin who both have really high LA values. However, if we compare CoD values, Cox looks really good (as does Martin). It is interesting to note that if you went solely by EN and LA, Fletcher Cox wouldn't be sniffing the first round, but if you look at TEN and CoD he looks better.

Ok, so how does this modification predict (if at all) performance in the NFL? If you look back at 2011 and apply the same adjustments you get this information of the defensive tackles who were drafted.

2011dtsnormbench_medium

2011dtsnormvert_medium

2011dtsten_medium

2011dtscod_medium

The modified data shows that Wilkerson and Taylor were undervalued while Nevis and Casey were overvalued by EN. In terms of CoD, it's interesting to note that only 3 DTs had number below 12 in 2011, while 5 guys have numbers below 12 this year with Martin getting an astounding 11.44 (Fairley was the best in 2011 with 11.70, Suh was tops in 2010 with 11.64). Wilkerson and Taylor had pretty impressive stat lines as rookies in the NFL

Taylor - 16 games, 59 tackles, 4.0 sacks, 1 PD, 1 FF

Wilkerson - 16 games, 49 tackles, 3.0 sacks, 2 PD, 1 FF, 1 safety

Nevis appeared in 5 games totaling 19 tackles and 1 PD

Casey was impressive with 53 tackles, 2.5 sacks, 1 PD, 1 FF, 1 FR.

In 2011 guys that were undervalued by LA but that CoD showed have value were Luiget, Fairley and Casey. Fairley and Luiget's rookie NFL stats:

Fairley - 10 games, 15 tackles, 1.0 sacks (hurt for a good portion of the year). 0 starts.

Liuget - 13 games, 19 tackles, 1.0 sacks, 2 PD, 1 FF

Of course you could make the argument that you don't need to be very quick to play well in the tranches Phil Taylor is an example of a rookie DT from last season (who was taken in the first round) that was able to collapse the pocket based on size and strength (6'3" 334 lbs). That being said, Taylor's TEN was the second best on any of the DTs that performed all three drills (Fairley, Dareus, Paea and Ellis all skipped one or more).

So what about the DTs drafted in 2010? Does this help separate the wheat from the chaff?

2010dtsten_medium

2010cod_medium

My modifications predict that EN and LA undervalued Al Woods, Linval Joseph and Terrence Cody while those two overvalued Geno Atkins, D'Athony Smith and Jeffrey Owens. Here are some NFL stats for those 6 guys (assuming that the value of NT/DT can be measured in tackles, sacks and PD - which it really can't)

Woods (4th round) - 11 games, 19 tackles, 0.5 sacks, 1 game started

Linval Joseph (2nd round) - full time started for the Giants in 2011, 52 tackles, 2.0 sacks, 5 PD. Only played 6 games as a rookie (8 tackles)

Terrence Cody (2nd round) - full time starter for the Ravens in 2011. 49 tackles, 1 PD. Played sparingly as a rookie. Currently listed as 370 lbs.

Geno Atkins (4th round) - as situational rush DT in 2010 recorded 3.0 sacks. Full-time starter in 2011 for the Bengals: 50 tackles, 7.5 sacks (lead all NFL DTs), 3 PD, 2 FF.

D'Anthony Smith (3rd round) - has yet to accumulate any NFL stats

Jeffrey Owens (7th) - appeared in 1 game for the Eagles in two seasons.

So there it is - three years worth of combine results for the DTs. The combine results were not a great predictor of success, but poor results in TEN and CoD seemed to have a good correlation to guys performing poorly in the NFL.

Thanks for hanging in there. There are a lot of numbers in this fanpost.

This is a Fan-Created Comment on MileHighReport.com. The opinion here is not necessarily shared by the editorial staff of MHR

X
Log In Sign Up

forgot?
Log In Sign Up

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

Join Mile High Report

You must be a member of Mile High Report to participate.

We have our own Community Guidelines at Mile High Report. You should read them.

Join Mile High Report

You must be a member of Mile High Report to participate.

We have our own Community Guidelines at Mile High Report. You should read them.

Spinner.vc97ec6e

Authenticating

Great!

Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.

tracking_pixel_9341_tracker