Studying the draft record of NFL Teams
In the recent DRAFTMETRICS article “Late Round Draft Picks: The Key to Success?” the 2012 season was reviewed to determine whether success in the late rounds was an important success factor for ”good” teams. In this article, DRAFTMETRICS digs back into history to see if success at the draft has more to do with drafting skills or accumulating extra draft choices.
DRAFTMETRICS focused its review on five-year starters produced in the 1993 to 2006 drafts. A player must have started at least eight games in each of at least five seasons to be counted as a five-year starter by DRAFTMETRICS. This time period was selected because it allowed adequate time (seven seasons) for players to become five-year starters. The Browns and Texans were excluded from the analysis because of the small number of data point as they entered the league in 1999 and 2002, respectively.
I have received several requests to do an analysis such as this for General Managers as their “draft record” may be at least as relevant as an individual team. DRAFTMETRICS cannot do that at this time but will add the General Managers to its data base over the summer months and will be in a position to do such an analysis in the future.
For purposes of this article, teams are given credit for a player they drafted regardless of whether he started for that team for all five seasons. For example, Antonio Cromartie was drafted by the Chargers, where he started for three seasons before moving onto the Jets where he attained the five-year starter milestone. The Chargers receive credit for all of his starts because they drafted him, and the purpose of this exercise is to measure drafting ability.
There is a wide variation in the number of five-year starters resulting from the draft choices of NFL teams during the study period, with the Packers and Steelers each drafting 35 and the Lions at the low end with 17. The average number of five-year starters for each team is 26. Here is how each team stacks up.
This leads to the issue of determining why a team ended with more or fewer five-year starters than the average. Were they better judges of talent? Or was it simply a matter of accumulating draft choices?
DRAFTMETRICS tried to answer these questions by first calculating how the actual number of five-year starters a team produced compared with the number they should have given the number and location
of their draft choices. This was done by categorizing each team’s draft choices into the seven Value Groups and applying the average league results (from the DRAFTMETRICS “Digging Deeper into Draft Probabilities” article). As a reminder, the Value Groups and the probabilities of drafting a five-year starter in each is shown below
After making that calculation,DRAFTMETRICS then determined the variation from the average that resulted from a team’s draft position and number of draft selections. The following table summarizes the results of the two calculations.
The “Efficiency” column shows how many more or fewer (indicated with a minus sign) five-year starters produced compared to what they would have been expected to produce, The “Choices” column shows the effect of their draft position and number of choices on the actual number of five-year starters. For
example, the 49ers draft choices produced 4.52 more starters than would have been expected. Their draft position and number of choices cost them 0.52 five-year starters, leaving them with a net total of four five-year starters more than the average.
The best and worst from the above table are as follows:
<p> Three teams stand out in these numbers, two of them good and one bad. The Packers and the Steelers represent the good. It is interesting to compare how they achieved their efficiency ratings. The Packers were very consistent. They had only two selections in the first 13 choices, but after that they had
positive efficiency in every Value Group exceptthe 67-86 picks. Green Bay did very well in the late rounds with at least of a margin of two five-year starters above average in each of the Value Groups after the 86th pick. The Steelers, on the other hand, achieved two-thirds of their positive efficiency from the 87-149 picks. The Lions were pretty bad across the board, but especially so with the 14-40 picks and 87-149 picks. Overall, though, they produced fewer than the expected number of five-year starters in five of the seven Value Groups.
Finally, DRAFTMETRICS cannot leave this subject without a brief discussion about the “L” word, or Luck in this case. If a team truly had a superiorscouting and front office staffin comparison to its competition, one would expect a fair amount of consistency in draft results. Recognizing that injury and non-football related matters can cause some bumps in the road, this consistency seemed to be lacking in our review (the Packers looking like an exception).
One example illustrates the point. With selections 14-40 the Eagles had one of the worst records of any team, with 3.85 fewer five-year starters than expected. With selections 41-66 the Eagles had one of the
best records, with 2.03 more five-year starters than expected. There may be explanations other than luck, but it was the same group of guys making the selections in both cases and in one case they stunk and in the other they were geniuses. It does cause you to wonder, though, if the draft is more like blackjack than bridge.