Drafting Decisions and the Salary Cap 2015

Normally this time of year I’ll do some different financial looks at the draft and ways to price draft picks, so here we go with this year’s entry.  The idea for this actually came from a post on Chase Stuart’s Football Perspective where there was a consideration as to how much, in terms of cap room, would a team creatively give up to acquire a draft pick.  While there are a lot of different ways to look at that question I thought it might be worthwhile to try to actually compare the expected rookie performance over a 4 year period with a comparable veteran player performance over the same time period. Using the veteran player’s contract as a “market price” we can re-value a draft pick in veteran dollars.

The Metrics

I went back and forth with a few ways to compare players, which is by no means easy in the NFL. It came down to either using my games/games started/honors matrix or Approximate Value as defined by Pro Football Reference. I decided to use AV for this since I think games/starts won’t be a fair barometer for veteran NFL players who are already proven to be gameday active most weeks whereas many rookies, specifically later round rookies, are not given  that opportunity.

Secondly I wanted to only look at four year windows since the four year period is our rookie contract length. While first rounder’s have fifth year options, those prices are based on a veteran marketplace so I didn’t think that would be used in the evaluation. So when we look at AV we are looking at a draft picks AV from year 1-4 in his NFL career and a veteran’s AV from year 5 onward.

I looked at every draft from 1994 through 2011 and compiled the combined AV for every draft pick made in that time frame. Each slot was then broken into quartiles to determine performance ranges that we can expect for the players. We set our upper limits using the IQR for each slot (the lower limit in almost all cases was going to be 0). Originally I was going to adjust for outliers but once I started looking at the data in many cases that didn’t seem to make sense since there is no logical reason for say a player drafted at 13 to be an outlier for that individual selection but be fine for pick 14-20.

Similarly I looked at every veteran player that has been in the NFL since 2011 and determined their combined AV for that four year period. Each player’s APY was then used as the salary basis for the player. The one limitation of that is that for some players that had multiple contracts over that time period I am using their most current negotiated deal. While that may not be perfect I think for the sake of these posts it’s fair to use. For players with a 0 AV quartile I used the cap charge vet minimum salary plus full bonus ($665K).

The Salary Calculation

With that data we can now create a matrix that allows us to determine the average veteran salary that corresponds to the expected AV range for each draft pick. There are different ways to use the financials, which I’ll touch on in some later posts, but the most basic method is to simply average the dollars across each quartile.

So let’s look at the number 1 overall pick for a quick overview of how the process works:

What those numbers tell us is that we have a 25% chance of drafting a player whose AV will be above 47.8 over the first four seasons of his career. Veterans in the NFL who accounted for totals in that range had an average salary of $14.37 million per year.  The team has a 25% chance of drafting a player worth $7.44 million, a 25% chance of drafting a player worth $4.325 million, and finally a 25% chance of drafting a player worth $1.975 million.   That averages to a bit over $7 million a year, which is the expected return for a player deemed worthy of a number 1 pick.

Because we get some fluky draft results, such as draft pick number 10 performing far worse than a pick after, I wanted to smooth the expectations by taking a forward average approach. This should take into account the fact that every team selecting in a given slot is likely considering drafting a player valued a few spots ahead. This also makes pretty much all our individual outliers that don’t make much intuitive sense to classify that way fit in with the expected results with few exceptions.

This was a purely subjective decision on my part just based on how we often hear draft grades give or mock draft debates on decisions.  I’d be more than happy to adjust results based on alternate draft possibilities if people had any ideas on the subject. Here is the way I broke things down:

Here is what our results look like using that formula

We can further (really we could have just done this initially) work with the data by determining a best fit line to use to calculate the draft pick value in terms of veteran dollars, which I will do in the following sections.

The Draft Pick Benefit

In the coming days we will look at a few different ways to interpret the numbers, but for this post let’s look at where the draft is most underpriced.

The biggest benefit to drafting in the NFL is the top of the second round.  This is somewhat expected as the NFL put a relatively large devaluation between the end of the 1^st and start of 2^nd round despite there being little difference in performance.  The top four picks in the second round are expected to give a team about $1.65 million per year in benefits. I was a bit surprised to see the top picks of the third round providing similar discounted prices. A team is essentially getting $2.5 million in value for just  $750,000 per year. The 16^th pick in the draft is our initial first rounder to make the list, adding about $1.6M in value. As I’ve discussed before the best values in that first round really come into play in the mid first round.

Our least value received comes with the 4^th pick in the draft, which gives us just $320,000 in additional value. The fifth pick in the draft is also poor with just $400,000 in benefits. These rank with all the 7^th round picks that provide limited value.

As for the first pick, it is a mid value pick. A team gains about $1.2 million per year in benefits. It’s the best value among the top 7 picks in the draft, despite having the highest price tag.  When we go back and look at the old CBA it is pretty clear how overpriced those picks were. Sam Bradford, the final number one pick of the non-slotted system, would have been overpriced by about $5.2 million per year.  To justify that pick you would have had to be one of the top picks ever at the position.

We can also look at the charts to better identify the gain that teams realize by being awarded compensatory draft selections. A third round pick is worth close to $2 million per year in true value and the team gains over $1.2 million per year in value. I think a team needs to strongly consider that when giving up comp picks for lower valued players. In general if you are in line for a 3^rd-5^th round comp you probably should protect that if possible.

The following chart shows the values per draft slot

Updating the Trade Value Chart

By using our expected salary values we can now create a new trade value chart that can be used to evaluate a trade.  With each player we can determine salary above replacement (again the 665K veteran) and re-assign the old trade value chart which overvalues the top picks by too much. However this is only part of the puzzle, which we’ll focus on a in part 2 later this week.