NewsBot
New Member
- Messages
- 111,281
- Reaction score
- 2,947
We try to identify the prototype players the Cowboys are looking for in the 2015 draft and then use similarity scores to find prospects that could match what the Cowboys may be looking for.
One of the harsh realities of draft season is that regardless of how much time we spend trying to get up to speed on the prospects in this year's draft, nobody will come anywhere close to the type of big picture NFL teams have.
Unless you're a professional scout, most of your information will reach you via some kind of media outlet, and Jack Bechta of the National Football Post explains the limitations that come with that.
What they do have is combine data, glimpses of players from all-star games, and some game highlight shots. Some may even have all the TV tape (not coaches tape) of all the games. Most get second hand reports from pro days that the media was excluded from and others have a direct conduit to scouts and even directors.
What they don’t have are the important pieces of the puzzle that have a huge impact on what decisions are ultimately made on draft day. The media lacks access to college injury files, Combine physical reports, first hand character reports from college coaches and teammates, and the whispers that come from college trainers and position and strength coaches who usually know more about the players than anyone.
But that doesn't necessarily invalidate all the work that draftniks across the world are putting into their draft preparation. While it's true that we'll never have the full picture that NFL teams do, if we think of that picture as a puzzle, then you don't need every piece of the puzzle to see what the picture looks like. If you assemble enough of the right pieces, you can get a good early idea of what picture you're looking at.
And that's exactly what we're trying to to here at BTB, offer pieces of the puzzle that in their entirety hopefully help you get a clearer overall picture of what the Cowboys could do in the draft. We looked at the college production of draft prospects, we're tracking pre-draft visits, we're looking at mock drafts, we're reviewing scouting profiles, and much more. Most recently, we looked at SPARQ, an aggregate metric for athleticism, that may have provided some extra insights that helped fill up the puzzle a little more.
Today we're going to try yet another approach, that may or may not yield one more piece of the puzzle.
When teams look at rookie prospects, one of the first things they look at is whether the player meets the physical prototype expected from the position. That includes height and weight, but also includes athletic markers such as the ones measured at the NFL Combine.
Bill Parcells explained the rationale for this quite succinctly in a TV interview a few years ago: For every position, there is a certain combination of physical measurables that has proven successful in the NFL, and deviating from this success model doesn't have high chances of success. In 2011, Gil Brandt codified some of those measurables in a set of target test results for Combine measurements.
But while those numbers are a great resource to evaluate a player's general fit to the NFL, every team's target measurables may differ slightly from the general NFL model.
Which is why today, we're going to make some educated guesses about what the Cowboys' prototypes may look like, and then see which players in the 2015 NFL draft match those prototypes most closely.
For the positions we'll look at today, we'll try to figure out which players could be prototypes for the Cowboys. And we'll combine that with the data from mockdraftable.com, whose similarity score we'll use to identify prospects in this year's draft who most closely match the Cowboys' presumed prototype.
Defensive Tackles
The Cowboys were very high on Aaron Donald last year, but the Rams picked Donald at No. 13 and the Cowboys had to look elsewhere for their first-round pick. Nevertheless, Donald brought an almost unheard of athleticism to the defensive tackle position, specifically the 3-technique position, that the Cowboys and other teams found highly intriguing. Here's Donald's player card from mockdraftable.com, which uses similarity scores to calculate which players are most comparable to Donald in terms of their measurables:
The data here suggests that four players in this year's draft class have similar measurables to Donald: Xavier Cooper, Grady Jarrett, Mario Edwards, and Louis Trinca-Pasat. Cooper, Jarrett, and Edwards are expected to to go somewhere between the second and third round, Trinca-Pasat is likely going to be a late-round pick up.
The Cowboys liked Donald for a number of things, his measurables among them. It follows that the Cowboys should have an interest in the four players above, and we already know that they've had a private workout with Mario Edwards.
And when you dig a little deeper and look at each of those four prospects in a little more detail, you'll see that they correspond to a very specific prototype of a defensive tackle, the quick-twitch, penetrating, fast get-off 3-technique. Here are the similarity score of five different prospects matched across four different NFL veterans
Player Similarity Score
Name Proj. Rd Aaron
Donald Geno
Atkins Sheldon
Richardson Amobi
Okoye
Xavier Cooper Donald Geno
Atkins Sheldon
Richardson Amobi
Okoye
2 70.5% 72.3% - - - -
Mario Edwards 2-3 68.0% - - 77.4% - -
Grady Jarrett 3 69.2% - - 75.6% 83.2%
Rakeem Nunez-Roches 4 - - - - 74.7% 82.8%
Louis Trinco-Pasat 6-7 66.5% - - - - - -
In the draft process, many people develop pet cats, but not all pet cats fit what the Cowboys are looking for. That first-round DT somebody mocked the Cowboys' way may look outstanding on paper, but that doesn't mean the Cowboys see that player the same way.
Of course, that doesn't mean that the method we're using here is any more accurate, but at least it's building off what we know the Cowboys liked at one point.
Linebackers
We know that the Cowboys were all set to pick LB Ryan Shazier with the first pick in last year's draft when the Steelers swooped in and grabbed Shazier from under the Cowboys' noses at No. 15. We also know that the Cowboys remain very high on Sean Lee, whom they had a first-round grade on in 2010.
Using the same approach as with the DTs, we can cross-check this year's linebacker class against Shazier and Lee, and this is what pops out:
Player Similarity Score
Name Proj. Rd Ryan Shazier Sean Lee
Eric Kendricks 1-2 69.7%
Benardrick McKinney 2 70.5%
Kwon Alexander 3 72.2%
Davis Tull 3-4 80.8%
Jordan Hicks 4 73.8%
Bryce Hagar 5 78.4%
Jake Ryan 5 78.2%
Unfortunately, Shazier didn't run a 40 at the Combine, so the similarities we're seeing here may be a little skewed, but if the Cowboys are looking for guys fitting their two templates, they have enough guys to chose from, in almost any round. Interesting sidenote: Sean Lee and Bruce Carter have a 80.4% similarity score while Rolando McClain and Kyle Wilber have a similarity score of 84.6%.
Running Back
If we assume that DeMarco Murray is the prototype for the Cowboys, then we're out of luck again, because the similarity scores for Murray don't show a 2015 prospect. But what they do show is interesting anyway: the player with the highest similarity score to Murray is Darren McFadden (91.9%), which could indicate that the Cowboys are thinking of McFadden as more than just a third-down back.
But just because the top 10 similarity scores on Murray's card don't show a 2015 match doesn't mean we're out of options. Because we can also work this in reverse, and look at the 2015 prospects and their similarity scores. The third-highest score on Melvin Gordon's card is none other than DeMarco Murray (72.4%).
The similarity scores don't work as well here as they did at the previous two positions, in part because we don't know what the Cowboys prototype at running back is. If we take Murray and add Marion Barber, Felix Jones and recent draft pick Joseph Randle, we don't get a common prototype player. But perhaps that's the way it is for the Cowboys.
Anyway, if we use those four players as prototypes, here at the most similar players in this year's class.
Player Similarity Score
Name Proj. Rd Demarco Murray
6'1, 213 Joseph Randle
6'0, 204 Felix Jones
5'10, 207 Marion Barber
6'1, 221
Melvin Gordon 6'1, 213 Joseph Randle
6'0, 204 Felix Jones
5'10, 207 Marion Barber
6'1, 221
1 72.4% - - - - - -
Tevin Coleman 2 - - 89.5% - - - -
Duke Johnson 2 - - - - 91.0% - -
David Johnson 2-3 - - - - - - 75.4%
Zach Zenner 6-7 - - - - - - 78.6%
John Crockett 7-FA - - - - - - 77.9%
With these disparate prototypes, we are rapidly reaching the point where this exercise loses its usefulness.
And we run into similar at DE and CB. Absent a better understanding of what prototype player the Cowboys are looking for, these similarities aren't of much use. But for the positions above, I hope the similarity score was at least enough to provide some additional food for thought.
Continue reading...