[Cowboys scouting director Tom Ciskowski] told me that in the initial [training camp] practices each scout is assigned a single position to evaluate. They then produce a unit evaluation, where each unit is ranked from its best player on down. The scouts do not write full reports but offer "one liners" of each player’s game.
When the Cowboys begin playing opponents, the scouts are sent DVDs of the games, so that they have an understanding of each player’s performance in live action. The reports are used as baselines; when the scouts evaluate college players, they do so in relation to the talent on the roster. "For example, if you’re grading a center and he’s similar to Andre Gurode," he said, "you would compare him to Gurode." "You don’t want to bring in players who are not as good or who can’t compete with the talent you already have on your squad."
-- "The Teaching Continues," BSR, August 2, 2008
The Cowboys are deep in their draft preparations, a process they begin each year well before the college season starts. As this quote demonstrates, the team tries to create a consensus among its scouts of its own talent. Each scout who goes on the road has a solid idea of the talent on his team's roster and can make sound evaluations of a prospects ability to make the team, in comparison to the talent already in house.
That's why my first mock stressed players on the low end of the roster, the 3rd year guys who just don't have starting talent, or the talented, but aged. Flozell Adams may be old, but how many of the current crop of '10 tackle prospects could push him?
This represents only a starting point in constructing a draft board, one which will have little relation to the ones you see on mock lists you find online or in magazines. Here's a primer on key board-building terms.
Vertical ratings -- these are the top to bottom ratings within a position group. Try it yourself. Let's take one position. Let's say safety. Rate the five guys on the team, from one to five. Mine would look like this:
- Gerald Sensabaugh
- Ken Hamlin
- Alan Ball
- Mike Hamlin
- Pat Watkins
The final numbers may seem a bit unfair to Watkins. He's been a special teams ace for a while and he started a game last year, while the younger Hamlin missed much of the season with a broken wrist. Still, Watkins is a four-year vet and has only started when injuries pushed him into the lineup. It's not apparent that he's suddenly going to challenge Sensabaugh or Ken Hamlin for downs from scrimmage.
Now, let's look at a vertical rating of college prospects. Let's look only at free safeties, and take one of the better known media graders, the National Football Post's Wes Bunting:
- Eric Berry, Tennessee
- Morgan Burnett, Georgia Tech
- Nate Allen, South Florida
- Major Wright, Florida
- Taylor Mays, USC
- Earl Thomas, Texas
Let's go to the first obvious vertical question: how would each of these guys compare to Ken Hamlin and Alan Ball, the 1st and 2nd team Cowboys' free safeties?
That leads us to another vital scouting question: where does each of these guys rate by round? In other words, how many of these guys would get a first round grade? How many a second? Teams not only grade players within their groups, but they rate them in relation to other players at their position in other years. This is a very important distinction, because you may be the top player at your position but if you don't merit a top grade, a team is wasting a pick if it picks you solely on need.
Let's look at wide receiver, for an example. Two years ago, not a single wide receiver was selected in the 1st round, even though WR is the most picked offensive position in the first. This means that none of them rated as high as a Jeremy Maclin or a Percy Harvin, guys who went in the top 20 last year. Which brings us to the skill that really earns scouting departments their money:
Horizontal ratings: this is the ability to distinguish between players with similar grades. Let's say you have 24 players with a first round grade in a given year. (More on this in a minute.) You have two DEs with first round grades. You can tell which one should be higher than the other, but how does that top DE compare to the quarterback with the first round grade? Or to the cornerbacks up there?
As Michael Lombardi pointed out in a piece last year, this is where most mistakes are made. A team may need a tight end and have one with a first round grade on the board when it's pick comes up, but if it has three other players with higher first round grades also available when it picks, and it still goes for strict need, it has left value on the board and cheated itself. Which brings us to another critical horizontal point:
Teams don't build boards according to 32 player brackets. Look at the top rating services and you'll see the top 30 or top 32. This means nothing to many real war rooms. The Cowboys give a player a round rating and then slot him within that round. That round number fluctuates; the first is usually between 18 and 25 players deep. The subsequent rounds are bigger. In 2008, for instance, Dallas assigned 1st round grades to just 21 players. Last year, 23 players got that grade.
The next player on the board will begin the 2nd round. Let's say there are 20 first rounders this year, for arguments sake. The board would rate the turn like this:
- 1-18, Joe Bob, QB
- 1-19, Billy Buck, DE
- 1-20, Harry Doe, OT
- 2-21, Smith Jones, RB
and so on. When you look at the Cowboys' pick at 27, weight the odds of getting a first round value. This is just as important, if not more so, than filling that crying need you see there. Let's say there are 22 1st round grades this year. And let's say the last first rounder, a cornerback, slides to 27. What do you do? If you're Dallas, you either take him, or look for a trade partner who wants him.
This scenario isn't as unlikely as it may seem. Something like this happened two years ago. Dallas had assigned 21 1st-round grades and had the 22nd and 28th picks. They had a decent chance of landing one of the last 1st rounders with their first pick but looked likely to pick one of their top 2nd round prospects with the next selection.
Then,teams went OT crazy. Starting with Ryan Clady, six of the ten players picked between 12 and 21 were OTs. Eight OTs were picked in the first round. Some represented reaches. The NFL Networks' Mike Mayock said recently that he had assigned the Texans' OT pick Duane Brown a 3rd round grade. I had a source tell me something similar about the Lions' OT pick Gosder Cherilus.
Drafting is a zero-sum game. One player's rise means another falls. The OT flurry pushed a handful of top-level candidates into the 22 range; when Dallas came on the clock at 22, three of its 1sts were still available. This is the type of horizontal choice which tests organizations. Dallas did what draftniks and front office types profess but frequently fail to do -- it picked off its board.
Felix Jones and Mike Jenkins were available and Dallas liked both. It has Jones rated higher and took him first. A lot of amateur draftniks ripped Dallas for not taking Jenkins at 22, because a lot of semi-pro boards had him higher. Dallas didn't however, and they took their higher rated guy first. When Pittsburgh and Tennessee took RBs next, their strategy was validated. Dallas wanted Jones and would not have gotten a shot at him at 28. (Jeff Fisher admitted later that day that he wanted to draft Jones.)
Those two RB picks also kept Jenkins on the board and gave Dallas a shot at a trade up. When the Cowboys consummated a deal with Seattle, they left the round with two players from their 1st round tier.
Which brings us back to your data. Are you reading sites or services who grade talent this way? Some do build boards in this manner. The War Room has created a 7-round board for the Sporting News for several years. I find them to be required draft reading for this reason. Bunting, who I linked earlier, using a grading system which avoids round grades, but if he has 24 players this year who he feels can be "featured players."
Is your favorite draft drawing a line on rounds? When a prospect is rated, is he compared to other players at his position in this year and in other years? If it isn't, your supplier isn't building his board the way the pros build theirs.
Tuesday: Combine? We don't need no stinkin' Combine!