clock menu more-arrow no yes mobile

Filed under:

The Birth Of A Stat - Part 1: The Think Tank

As Dane alluded to earlier this week, some of us like to get our stat on every once in a while.  Those of you who know me know that this might be an understatement in my case. I spend hours sifting through simple statistics and calculating complex stats a couple times a week.  But through my Top Draft Philosophies series, I have come across a problem.  There are no statistics that are even remotely relevant to draft value.  I think it's about time to fill that void.  Follow me to the clearing beyond the end of the jump to find out how.

Weighing team drafts and draft classes is such a subjective chore.  A perfect system is nearly impossible.  But then, that is true of virtually all measures made through complex statistics.  The goal here today is to devise a system that accurately reflects each draftees value to date.  It'll be a tough task, granted.  But I believe we have one hell of a think tank here.  And I have a general idea of how to accomplish this.  

The actual mathematics of this system can and will be figured out.  Please don't feel that your input is not valuable because you aren't sure how to mathematically apply it.  That's my job, though suggestions are certainly welcome.  In fact, they are encouraged.  The more heads the better.  And no matter how trivial, simple, complex or even dense your idea may seem to you, it may help more than you believe.  This isn't going to be a hasty project, and this is just the first part of this series.  As I sift through your ideas and input, I'll adjust fire, test, tweak, retest, pull out some hair, re-retest and then come back with the problems with the first draft of the statistic.  This will be an ongoing process and your input is vital.

So here's what I have, so far.  I'll leave the complexities of the math out, for now.  All of this is going to be built off of the foundation provided by Pro-Football-Reference.com's Doug, using his AV system.  Now, I don't see this system as fool-proof, nor does Doug claim it is.  In fact, it's the honestly about his system that endeared me to it when I first saw it.  And it's the constant tweaking to get it right that legitimizes it.  Doug's system isn't about his own pride, but about an honest to goodness attempt to quantify something that seemed overwhelming to most.

There are several reasons I chose this specific statistic for the baseline by which to help measure drafted players.  First, while not perfect, it is the best (IMO) metric around for raw measurement of a players impact on a team, which is, of course part of what we are trying to determine.  Second, the fact that a team's overall performance is calculated into the individual measure is critical to measuring a players impact as a draft pick.

Not that, while I think the Weighted Career AV is a fine metric for measuring careers after they are finished (and only of similar lengths, really) it is the raw Career AV that will be used here.  We aren't trying to measure players, but measure players impact as it relates to the value of their drafting.

That being said, AV alone cannot measure accurately the effects of a draft.  This is where I need ideas.  The WCG thinktank, if you will.  I have several thoughts on other stats to use in calculating this new metric.  Here's what I've come up with....

Ratio of games started vs possible games started.

Pro Bowl/All-Pro nods

Games played vs possible games played

And some more vague ideas that I need to flesh out like.....

The idea that it also help to rate the player vs others at his position, with some sort of minimum control.  I was thinking of using some sort of ratio based on fantasy ranking vs fantasy average.

indiviual position metrics.  Covered, for the most part by using the AV metric as the baseline.  But it may require some tweaking to avoid specific positions affecting the outcome one way or the other by virtue of position drafted alone.

How to account for players washing out in a single season or even before final cuts.  There will have to be some sort of modifier for the amount of time a player remains with the team.  I mean, what's a better pick?  A one year wonder who is stellar for a single season but washes out in his sophomore season, or a Hunter Hillenmeyer type guy?

Accounting for catastrophic injury (career ending) that wasn't foreseeable and accounting for the injury prone that were.  This might be one of the most subjective ideas.  But it has to be dealt with.  Billy Sims, for example, was not a bad draft pick because he was devastated by a knee injury.  Chris Williams was, considering he was injured when he was drafted, and the Bears knew.  But how do we control that?

And one final thought.  We will not be rating players here, but a players value to the team upon being drafted.  That brings me to two thoughts.  One, because all picks are not equal, some sort of multiplier should be invoked to cover the difference between drafting Peyton Manning in the first round and Tom Brady in the sixth round.  Regardless of the debate over who is better, I don't believe it can be argued that Brady wasn't a better value.  And Two, since the era of free agency removes some of the control from the team over where the player plays, I'm thinking a players entire career should be considered, regardless of his location.  Additionally, because we are rating the pick, and not the ensuing decisions by management, all players should be considered over their entire career, as a team that drafts a future Hall of Famer but then releases him out of stupidity or impatience after one season still drafted a future Hall of Famer.  Again, we are measuring the draft picks, not the personnel decisions.

That's my starting point.  Let's hear what you think.  How much do you think I should weigh each part?  Am I missing something?  Am I putting to much weight into something?  What, how and why?  Gimme your ideas and your critiques, Ladies and Gents, and let the birthing of a new statistic begin.