Tuesday, April 26, 2011

What Can You Expect From a Draft Pick?

The NFL Draft is fast approaching; so, there is no better time of the year to analyze the overall draft performance of the league and the relationship between draft position in NFL success. The draft can change team's fortunes if they draft the right player (Tom Brady in the 6th around) or the wrong one (Ryan Leaf), often for seasons at a time. Coaches, scouts, and general managers spend months and months before the draft analyzing every facet of a player--from the way he handles himself during interviews to the speed he can run a 40-yard dash. Not surprisingly, the 1st-round draft picks are the most scrutinized, as they are given millions upon millions of dollars and expected to contribute at a professional level almost immediately. However, what should the expectations be for those picks, as well as the players taken in the following two rounds--who often end up being major contributors as well? What can teams expect out of a player picked in the first 100 picks? We might just have an answer...

In order to quantify the league’s draft performance, I analyzed the 20 drafts from 1987-2006. The study was limited to 20 years because it was big enough to be of a sufficient sample size, yet small enough to be completed in a short period of time. Drafts after 2006 were not used because it is often difficult to analyze a draft without five years of player performance to evaluate. The metric used to measure the draftee’s value and subsequently, the league’s drafting ability was “number of seasons in which the player was the primary starter at his position”; it’s a stat that I found at Pro-Football-Reference.com. This stat would not work well in comparing one player to another, but when comparing one set of 100 players to another set of another 100 players, it is probably the best measure to use--even when taking into account that kickers and punters have longer careers. Draft position is a surprisingly good predictor of player quality. I thought it was a good predictor through the first 30 picks or so, but it is actually a great predictor through 80-90 drafts picks. The graph below shows the average number of years as a starter for the first 100 picks of the draft. The important measure here is the Ten-Pick Moving Average: anything can happen from one pick to another, but the top 10 picks are almost always better than the next 10 picks, the second ten picks are almost always better than the next ten and so on. Any fan would expect the further down in the draft you go, it’s less likely that you’re going to find starter material.

Although there is a downward trend as the draft progresses, this downward trend decreases at a slower rate, suggesting that difference between successive draft picks becomes smaller as the draft goes on. The 98th pick in the draft, on average, has had 4.9 years as a starter, which is more than draft picks 37-97. It's obvious this is an outlier when one looks at who was drafted 98th overall. Between 1987-2006, some 98th picks included Rich Gannon (28,000 career passing yards), Jon Kasay (who kicked at least 20 FGs in nearly every season from 1991-2010), and Derrick Mason, the still-active 37-year-old receiver with nearly 12,000 career yards. In addition, plays like Houston TE Owen Daniels, Green Bay linebacker Na'il Diggs, and former Chiefs Pro Bowl linebacker Donnie Edwards.  The next graph uses the same attribute, seasons as a starter, to measure the NFL’s overall drafting effectiveness. It takes the top 100 draft picks as measured by total seasons as a starter, and the top 100 draft picks in the order they were picked and measures the relationship between the two in terms of percentage. The logic behind it is simple, even if the prior explanation seems convoluted. It basically measures how many of the top 100 picks ended up being one of the top 100 players in the draft class. Here’s an example to illustrate the measure: In 1987, the top 100 picks totaled 358 seasons as a starter while the 100 players with the most seasons a starter had a total of 607; this translated to the top 100 picks having 59% of the total seasons as a starter. This graph shows that in the last season of this study, 2006, NFL front offices did their worst job scouting and drafting since 1990.

There's A Stat For That

No comments:

Post a Comment