BPA absent Positional Value adjustments

Last post I attempted to explain, in no uncertain terms, why taking a Guard in the first round makes little sense (especially in the top half of the first round).  However, I get the sense that many people remain unconvinced.  Today, I’ll take a slightly different look at it to see if I can sway the remaining holdouts.  I apologize for the repetition in subject-matter, but this is an EXTREMELY important topic, as it goes to the heart of Optimal Draft Strategy.

First, I want you to ask yourselves if you agree with the following statement:

GMs should, in general, take the Best Player Available (“BPA”) with every draft pick.

I’m guessing most readers here would back that strategy, and I certainly do.  HOWEVER, it is not enough to just endorse that statement and use it as the basis for a draft strategy.  First, you must definite exactly what BPA means.

Here is where I am seeing some confusion and where I believe there is a big disagreement.  The people calling for taking Warmack high in the draft (forget about the Eagles for a moment, I’m speaking in broader terms), are using BPA as support.  The problem is that, in their BPA definition, they do not seem to be adjusting for positional value, which is a MASSIVE mistake.

Rather than attempting to explain why with a rational argument (for that see my last post), I’ll try to illustrate it.  Here are the top ten OTs and Gs from this past season, according to Pro Football Focus.  They are listed in order, with their original draft pick included.

Screen Shot 2013-03-04 at 11.34.40 AM

There are two big takeaways from this chart.  First, look at the draft picks.

I know I’ve said this repeatedly, but if you want an “elite” OT, you have to take him in the first round.  The same clearly does not hold for the Guard position.  This provides a clearer illustration of the opportunity cost argument I made last week.  There are really only two ways to get an elite OT in the NFL, draft him or sign him in FA. (I know we traded for Peters, but that’s a RARE exception).  Additionally, signing an elite OT in FA requires a huge contract.  Therefore, the BEST way to add a top OT to your team is to draft him in the first round.  Choosing a G means you are forgoing that opportunity.

The second point I want to highlight from the chart above gets at positional value.  Again, I want you to answer a question:

All other things being equal, would you rather trade Anthony Davis for Evan Mathis (or any other top Guard) or vice-versa?

I’m a big Mathis fan, but that’s a no-brainer.  Clearly most of the league agrees as well, hence the salary differences between OTs and Gs.

Assuming you agree, that means the #10 OT is worth more than the #1 G.

How about Brandon Albert versus Evan Mathis?

This one probably creates a bit more disagreement, but my guess is, on the whole, people would take Albert, who ranked as the 25th best OT according to Pro Football Focus.

Regardless, clearly there is a discrepancy in value between OTs and Gs.  Also, while I’ve focused on Gs and OTs, the same analysis can (and must be) done with all positions, and then incorporated into each team’s prospect rankings.  To not account for this is a huge mistake and one that appears is made by a lot of the writers who support taking a G high in the draft under a flawed concept of BPA.

To make sure I’m being clear: This should be incorporated into the “tiered” rankings I advocated previously.  So it is theoretically possible for a G to be the best choice with a top 15 pick, but it would require an amazing G prospect that can overcome the positional value difference to make it to the top 1 or 2 tiers, as well as every other player (from positions of greater value) within the same prospect tier being taken before said top 15 pick.

Hopefully this has convinced a few more of you, or at least provided a clearer explanation of what I was getting at last week.  This isn’t exactly Eagles-specific, because I don’t think there’s any way they take Warmack at #4, but it applies to any team with a top 15 pick.

Lastly, I’ll leave you with an analogy.  A frequent defense of G picks (and OL picks in general) in the first round has been that they are “safer”.  In general, that is true, as our strategy chart showed.  However, this is only one half of the equation.  As anyone in finance knows, evaluation is always a question of risk vs. REWARD.  To look at one side and ignore the other is a recipe for sub-optimal decision-making.

For example, given the choice between to raffles, and ignoring an external factors, which would you rather enter:

A) 60% chance of winning $100.

B) 75% chance of winning $75.

While the odds of success for raffle B are significantly higher, the correct answer (in a vacuum) is A, since the expected payout is greater.  It’s impossible to apply such specific values to players, but it’s a useful exercise nonetheless.  Here, imagine a DE as raffle A and an OL as raffle B.  To take the OL just because he offers less risk (25% chance of miss versus 40%) would clearly be a poor decision.

Advertisements

7 thoughts on “BPA absent Positional Value adjustments

  1. “First, you must definite exactly what BPA means.” – When I read that, I quickly thought about how great you are going to do in law school… hahaha. Complements aside, I have a great idea for a quantitative analysis to demonstrate exactly what positional-adjusted BPA is. Here is the gist of it:

    Basically, we already have a natural experiment that determines exactly how much each team comparatively values each position. That is, if we aggregate each team’s contracts per position while leaving out the rookie deals (rent control), we would see exactly what proportion of the aggregate salary cap total (cap*32) the league values each position. Then we can take a player’s absolute combine rating ( on a 1-100 scale) and multiply it by (1+[% of overall cap allocated to the position in the league]). What is your opinion on doing an article to this effect? If it seems too labor intensive, I would be more than willing to collaborate on it.

    • I like it. What do you mean by “absolute combine rating”? I think putting together a weighted measure of positional importance is valuable by itself, but if we can figure out a way to apply it to the draft board it obviously becomes much more intriguing.

      • The average amount a team is willing to spend (non rookie) on a position is the perceived value that the position adds to a team. We take that figure as being representative of the comparative difference in team-value-added (TVA) between hypothetical players at different positions. For example, assume teams spend 10% of their cap space on a LT while only 5% on a LG. This demonstrates subjective comparative value in how the NFL views the two positions.

        Next let’s say we have two draft prospects, a LT who is a 90 and an LG who is a 93. We preform the respective weighted adjustments for the LT (90*1.1 = 99) and LG (93*1.05= 97.65), which tells us that the positional value-adjusted BPA rating between the two has the LT as the better draft value (and thus worthy of the higher pick). If you were being literal as to what I mean by “absolute combine rating,” it would simply be the average of the player rating from three reputable sources (NFL, ESPN & Scouts?).

      • I was being literal. I figured you were working from Scouts Inc. ratings, but wanted to make sure. Regardless, its worth exploring.

        I’ll start putting the salaries together, depending on how fast I can compile them. If you want to help, pull together consensus prospect ratings (let’s just work from first round prospects for now) and email me. If not, no worries, I can handle it, itll just take a few more days.

        Thanks for the idea. Im intrigued to see where we can take it.

        On Mon, Mar 4, 2013 at 6:16 PM, Eagles Rewind

  2. I started compiling the first round prospect grades and realized that ESPN and Scouts Inc. use the same ratings. Anyway, I took their top 32 prospects (all you can get without being an insider member) and paired their respective NFL.com ratings in a spreadsheet. I am going to look into whether there are any other reputable sources that numerically grade NFL prospects. Do you know of any? I was unable to find your email on the site; if you’re looking to keep it private just email me at the email I post under.

      • NFP is on a 1.0-10.0 scale, so I will just multiply it by 10. I am going to send you the incomplete spread sheet now; just add in prospects 32-60 (first two rounds) from ESPN/Scouts (paste as HTML) and ill add and average in their respective NFL and NFP ratings tonight.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s