More draft theory

[dryheat44]

The Jumping-Off point:

That pretty much summarizes my exact thoughts on Willis. And it's made me ponder this impromptu equation:

impact = talent X fit X need

Guys like Carriker and Staley come up as legitimate Pats 1st-rounders despite marginal need because of their outstanding talent and fit scores. Patrick Willis, meanwhile, comes across as a terrific talent at a key position of need yet not a perfect fit. I'd never pass him up at #24 because I think he could succeed at the Pats' ILB position, but it wouldn't be the best use of his talents.

Meanwhile the more I watch Stewart Bradley, the higher he climbs. Clearly less talented but looks like an ideal fit, yielding a solid impact...and likely an excellent draft value, given that the talent part of the equation holds fairly equal across teams while the fit varies wildly. More gut reactions on highlight clips say that Brandon Siler looks pretty average on both talent and fit; Lawrence Timmons looks enormously talented and a terrible fit; Jon Beason simply looks all wrong. (I wish I could speak more intelligently on Beason, but he just doesn't look anything like a Pats ILB to me.)

So right now my ILB-or-bust board is pretty scant:

1. Patrick Willis (outstanding talent, average fit)
2a. David Harris (excellent talent, very good fit)
2b. Stewart Bradley (very good talent, excellent fit)
gap
3. Paul Posluszny (very good talent, very good fit)
gap
4. Anthony Waters (potentially excellent talent and fit, but hard to evaluate)
gap
5. Desmond Bishop (good+ talent, very good fit)

Let's start with the formula: impact = talent X fit X need

I don't think Patchick that as a literal mathematical formula, as need would be the overriding concern. I also don't think all three are considered in equal measure. What I've come up with is:

I = 30%T+50%F+20%N

I don't know if that's entirely accurate, but it seams reasonable. To test it out, I considered 12 prospects that would be a good cross-section between those three attributes. They were graded on a simple 5 point scale, so the gaps between them aren't representative. But according to this theory, the Patriots priority list would look like this (among the 12 I ran numbers for):

1. Darrelle Revis 4.8
2. David Harris 4.7
3. (tie) Adam Carriker, Robert Meachem 4.4
5. Joe Staley 4.1
6. Anthony Spencer 4.0
7. Josh Wilson 3.8
8. Brady Quinn 3.7
9. Patrick Willis 3.5
10. Chris Houston 3.2
11. Zach Diles 3.1
12. Jamaal Anderson 2.8

Is this the order of preference Belichick might have? Probably not, since 1) Carriker would probably go ahead of Harris and Meachem, and 2) Diles would almost certainly be on the bottom. But with another attribute figured in, like "Positional Rarity", we could probably solve the first one, and with a wider scale the second flag would probably be resolved.

I'll be doing some more thinking and refining. Any thoughts in the meantime?

 

[Boston Tim]

Let's start with the formula: impact = talent X fit X need

I don't think Patchick that as a literal mathematical formula, as need would be the overriding concern. I also don't think all three are considered in equal measure. What I've come up with is:

I = 30%T+50%F+20%N

Given that players develop and improve at different rates and we're talking about impact over at least an entire season (or even several seasons), I think you should breakup talent into "floor" and "ceiling" and then determine priorities on the group of four. Actually, you could eliminate "Need" altogether and just increase/decrease floor based on how quickly the player must be ready. This also allows you to use different formulas for each position.

QB: I = (10% x Floor) + (40% x Fit) + (50% x Ceiling)
ILB: I = (40% x Floor) + (40% x Fit) + (20% x Ceiling)

Or is this just crazy?

 

[Ochmed Jones]

I would make a small change to the formula. I would make talent 50% and fit 30%.

 

[sebman2112]

You can't minimize talent as having only 30% importance.

 

[dryheat44]

You can't minimize talent as having only 30% importance.

That's what I'm struggling with exactly. But I think fit is more important, to a point. I might need a tiered system, i.e. this formula applies only to guys who rate a seven or above in talent.

 

[Boston Tim]

Step 1: setup a formula for each position (preferably using "floor/ceiling/fit")
Step 2: apply appropriate formula to draft candidates for each position.
Step 3: by position, create ranked player list.
Step 4: group players by similar score for each position (one to many players in group.)
Step 5: using the player's scores, create an average score for each value group.
Step 6: using the group's average score, assign the group to a place within a round (high round 1, mid round 1, low round 1, etc.)
Step 7: compare the patriots draft picks to the group/round mappings.

This sounds like fun. Maybe I'll take a pass this weekend.

 

[patchick]

After I posted that, it occurred to me that "impact" isn't the same as draft value. Part of any value calculation, football or otherwise, is rarity. So the rarity of the player's skills or body type and probably the dropoff to the next player at the position has to factor in. (You might even want to add a factor for financial considerations; e.g. a 1st-round CB is more of a bargain than a 1st-round OG.)

If you factor in bonus points for rarity value, Carriker suddenly soars.

I really like some of these calculations...if nothing else, they help focus my thinking in interesting ways.

 

[SamBamsFan]

Baseball has its sabremetrics. DH44 has definitely turned me on to to Wilson. I like Usama Young. We have to figure out a metric that gets us from JW @ say 45, vs UY @ 125 as relative cost/value. This I think is one for the offseason.

 

[Mainefan]

And yet, after all the calculation, study and formulas, it is possible--not even unlikely--to choose players who seem perfect on paper, yet fail on the field.

What qualities cannot be measured, no matter how hard you try? My leading candidate is the adjustment to NFL speed. Some people, I think, simply do not have the mental capacity to do this--and it's not intelligence or athleticism that they're missing. I don't think motivation plays a role either.

I wonder if there might be a way to test for this ability, perhaps using reaction time testing or something of the video game variety.

 

[dryheat44]

And yet, after all the calculation, study and formulas, it is possible--not even unlikely--to choose players who seem perfect on paper, yet fail on the field.

What qualities cannot be measured, no matter how hard you try? My leading candidate is the adjustment to NFL speed. Some people, I think, simply do not have the mental capacity to do this--and it's not intelligence or athleticism that they're missing. I don't think motivation plays a role either.

I wonder if there might be a way to test for this ability, perhaps using reaction time testing or something of the video game variety.

Well, I'm trying to determine who Belichick is likely to draft...not how good that player is going to be. If that were possible, I'd give the GM a lifetime contract at 10 million a year.

 

[Mainefan]

Yes, Dryheat, and you are doing a fine job at it. However, I suspect that BB is trying to determine how good his potential draftees are. I'm just wondering if adjusting to NFL speed adjustment might be the key. This doesn't quite rise to the level of theory--let's say it's a guess.

 

[dryheat44]

Yes, Dryheat, and you are doing a fine job at it. However, I suspect that BB is trying to determine how good his potential draftees are. I'm just wondering if adjusting to NFL speed adjustment might be the key. This doesn't quite rise to the level of theory--let's say it's a guess.

I don't doubt that at all. I'm guessing that if you're not quick to react in NCAA, you haven't got a prayer in the NFL. Let's wear BBs shoes. We're looking over game film, you're watching guys do private workouts.....what indications do you think there would be that would show a capacity for playing at NFL speed? I would guess that watching film, you would (let's use a linebacker as an example) note how quickly after the snap he makes his first move (assuming his move was that correct one). Ditto for a safety or defensive lineman. I would imagine the biggest key for an offensive lineman is moving through space and hitting a moving target in the run game, quickness in setting up your pass blocking stance in the pass game. How quickly and decisively a running back makes his first step after taking the ball, and upon making his first cut. How quickly a cornerback transitions from delivering a jam into his backpedal.

Definitely an interesting topic.

 

[shakadave]

Another factor for the formula is Probability Players At That Position Are Not Busts. For example:

Any WR: 50%
Any QB: 60%
Any LB: 75%
Any CB: 80%
Any DL: 60%

(Just my estimates of the numbers off the top of my head)

 

[Boston Tim]

Look at the 1:1 matchups that the prospect had with players that entered the NFL the previous year and who also adjusted well to the speed. That tells you that the prospect managed to play against an NFL calibre player while in college.

 

[Box_O_Rocks]

I'm not sure an equation is the best way to define a player. This site http://www.newerascouting.com/index.php?c=30&a=139 does an assessment based on general traits and position specific traits.

Though if you are so inclined, perhaps a similar approach is to take patchick's equation and rework it a bit:

impact = (talent + knowledge + training)/fit

talent: foot speed, vision, hand speed, explosion, burst, arm strength, reaction time

knowledge: would be hard for me to define, I don't even know what I don't know.

training: same as knowledge, I have little clue on what each position needs for training.

fit: would be measurables (height + weight + arm length + speed + quickness + strength) and so called intangibles (leadership, poise, determination, work ethic).