## NFL Draft Efficiency Before and After the Rookie Wage Scale

By Kevin Meers

With the new collective bargaining agreement came the rookie wage scale and a lengthy discussion of how the rookie wage scale would change draft strategy. Using salary data from www.spotrac.com, this analysis presents two models for rookie contract value based on overall pick drafted before and after the new CBA. Combining these models with my previous analysis of Career Approximate Value (a metric from www.profootballreference.com), I present a model for CAVOA/dollar both before and after the 2011 CBA to see how the rookie wage scale affected draft efficiency. (Note that throughout this post, salaries are in tens of thousands of dollars.)

Let’s start by modeling rookie salaries before the 2011 CBA. Salaries peak at the first overall pick, and rapidly decline until pick 34, after which salaries stay fairly constant. This sharp change makes a corner in the salary curve, so we need two equations to accurately model rookie salaries:

If pick ≤ 34, salary = -2197 * ln(pick)+8211.7 (R2: 0.92)

If pick > 34,  salary = -127.1 * ln(pick) + 843.74 (R2: 0.68)

Combining these equations, we get the model below (“eDollars” is expected salary):

Having accurately modeled rookie salaries, we can get CAVOA/dollar at each pick simply by dividing the expected CAVOA by the expected salary of each pick, which results in this graph:

CAVOA/dollar starts at 0.060 at the first overall pick and rises to 0.499 at pick 56 where it begins to fall until the end of the draft. The first 21 overall picks are the most inefficient of any in the draft, but the expected value of those picks is so large that teams accepted the relative inefficiency of those picks. However, it is clear that early first round picks are extremely inefficient compared to the rest of the draft. This inefficiency was one of the reasons that the league instituted the rookie wage scale. But how successful was this policy in making early draft picks more efficient?

The 2011 draft provides some answers. Like in 2010, there is a kink in the salary curve, but significantly later in the draft at pick 68. After this point, salaries become almost constant. Because of this corner, there are two equations to model salaries before and after pick 60:

If pick ≤ 62, salary = -529.2 * ln(pick) + 2451 (R2: 0.97)

If pick > 62, salary = -68.67 * ln(pick) + 574.29 (R2: 0.80)

Combined, these equations form the model below.

Rookie salaries under the rookie wage scale have a very similar shape to salaries before the 2011 CBA. However, the first overall pick’s expected salary is over \$55,000,000 less under the 2011 CBA. Because of that shift, rookie salaries are significantly lower at every pick than in previous years.

Having created this model, all that is left to do to get Career Approximate Value Over Average/dollar is divide expected CAVOA by the expected salary of each pick to get CAVOA/dollar. The result is below:

CAVOA/dollar begins at 0.203, and steadily rises until the 61st pick, where it peaks at 0.497. From pick 62 on, CAVOA/dollar falls to 0.112 at pick 252 (the final pick of last year’s draft). So picks become more and more efficient until the 61stoverall pick, where efficiency begins falling. The average CAVOA/dollar is 0.294 and the median is 0.288, which makes most of the first round relatively inefficient. Picks 25-147, on the other hand, are all relatively efficient. Given their inefficiency and low expected value, picks after 147 have severely limited value. It is not the case that no individual picked after 147 can be valuable. Rather, those picks have a low CAVOA and are financially inefficient. In terms of trading draft picks, trading out of the first round and into the second definitely increases the efficiency of a team’s draft. However, trading back after the 61st pick sacrifices both overall value and efficiency, which is the opposite of what teams, in general, want to do.

There are some significant differences between these two CAVOA/dollar models. Earlier draft picks are much more efficient under the 2011 CBA than in previous years. Efficiency starts higher and peaks later in the draft. However, the new CBA makes late round draft picks much less efficient. This change is great for teams, as they are better able to invest significant money in top draft picks without sacrificing as much efficiency as they had to before. While teams should love this change, future early round draft picks lost a lot of money because of the rookie wage scale. All in all, the league accomplished its goal of making rookie contracts more efficient.