|Draft Metrics: Successful Drafting - Who Does It and Why|
|By: Tony Villiotti|
|February 1, 2011|
In this article, Draft Metrics will examine whether the NFL teams that are most successful in the draft do it by accumulating extra selections, by being more efficient/have more good fortune with their draft choices, or both. Although most of the draft attention is on the early rounds of the draft, a
significant amount of talent comes in the later rounds as well. Are the best drafting teams more effective in those rounds than their competition? Draft Metrics will address all these issues in this article.
In order to have credible information on which to base conclusions, only the 1991 through 2004 drafts (or 14 years) will be included. The more recent drafts are excluded because the careers of the drafted players are still evolving. It was further decided to limit the analysis to the 28 non-expansion teams. Doing that allows us to ignore any aberrations caused by team start-up issues and provides data for the full 14 year study period for each of the teams analyzed.
Draft Metrics considered several possible measures for this study including:
• Number of 3-year starters, meaning a player starting 8 or more games for 3 or more seasons; |
• Number of 5-year starters, meaning a player starting 8 or more games for 5 or more seasons;
• Number of players selected for one or more Pro Bowls
• Number of players selected as All Pros one or more times
Total number of starts was ignored in this analysis because it tends to give greater weighting to older draft selections. A draft selection in 1995 for example would tend to have more starts than a draft selection in 2004 even though their careers may ultimately end up the same.
The chart to the right shows the percentage of draft choices and each of the four considered measures by Value Group.
See the previous article The Relative Value of Draft Choices on the website for an explanation of Value Groups. Essentially, though, it is Draft Metrics position that each draft choice within a Value Group is of equal value. Value Group 1 includes selections 1-13; Value Group 2 includes selection 14-28; Value Group 3 includes selections 29-48; Value Group 4 includes selections 49-74; Value Group 5 includes selections 75-114; Value Group 6 includes selections 115-200; and Value Group 7 includes selections 201 and higher.
Draft Metrics also reviewed the four measures from the perspective of a "batting average". That is, if you have X number of choices in a Value Group, how many 3-year starters, 5-year starters, Pro Bowl selections and All Pro selections should you expect to end up with? The batting averages are as follows:
So if a team had 10 VG1 selections they would expect to get 8+ 3-year starters, 7+ 5-year starters, about 5 players who make the Pro Bowl at least once and 2+ All-Pro selections.
After due consideration, Draft Metrics concluded that the number of 5-year starters would be the best single measuring stick. Teams ranked by number of 5-year starters drafted during the study period are as follows:
Being a 5-year starter does not mean that a player started all 5 years for the team that drafted him. Rather, it means that he started 5 years or more during his NFL career. For purpose of this analysis, though, he is counted for the team that drafted him. This is consistent with the purpose of this analysis, which is to measure the success of each team's drafting efforts. In a future article Draft Metrics will address the issue of draft choices who leave the team that drafted them and why.
The next step in the analysis was to determine whether draft performance was a function of the number of draft choices the team had or how efficiently the teams used their draft choices. For the number of draft choices, Draft Metrics compared the number of draft choices (or the volume) each team had in comparison to the average (120.5 per team for the 14-year period).
In order to assess efficiency Draft Metrics compared the number of 5-year starters a team drafted to the number they should have drafted given the number of choices it had in each Value Group. For example, the Tampa Bay Bucs had six selections in VG1. Draft Metrics multiplied 6 by .736 (the VG1 "batting average for 5 year starters), calculating that the Bucs should have ended up with 4.42 starters from VG1. This process was then repeated for each Value Group and the results were totaled for arrive at the total number of expected 5-year starters. This was then compared to the actual results, as shown in the chart above, to determine whether a team did better or worse than expected. This determined how efficient each team was in their use of draft choices.
The following chart summarized both the volume and efficiency of each team. If a team was within 5% of the average number of draft choices (or between 118 and 126) they were classified as having an average amount of draft choices. Teams with fewer than 118 draft choices were classified as having a low volume and those with more than 126 draft choices were classified as having a high volume.
The same 5% rule of thumb was used to classify the efficiency of teams. If a team was within 5% of the calculated expected number of 5-year starters it was classified as average. Teams that had less than 95% of the calculated number of starters were classified as having low efficiency and those with at least 105% of the calculated number of starters were classified as having high efficiency. The chart provides a two dimensional look at each team's draft results.
Teams with high volume and high efficiency (plenty of picks and making good use of them) obviously get the most value out of the draft and those with low volume and low efficiency (fewer picks and don't make good use of them) get the least.
The final step in the Draft Metrics analysis is to determine whether there are any lessons to be learned about the root of an efficient team's success. In other words are those teams more successful than the others in early rounds, late rounds or all the way through. The following chart shows aggregated actual versus expected performance for the teams in each of the three efficiency categories.
Another way of looking at this issue is to see is to track the cumulative number of 5-year starters by Value Group. The following chart reflects this tracking.
Just to be clear on what this represents, 46% of the 5-year starters drafted by the "High Efficiency" teams were in the first three Value Groups; 61% were in the first four Value Groups and so on.
Looking at both analyses indicates that the best drafting teams tend to draft better than their competition throughout the draft but it is slightly more pronounced in the later rounds. Differences are most pronounced in VGs 1 through 3 (midway through the second round) and then again in VG6 (mid 4th round through early 7th round). As an aside, it is somewhat difficult to relate the Value Groups to rounds due to the awarding of compensation picks. In theory, the draft has 224 picks (7 rounds times 32 teams) but the drafts typically have about 250 picks due to the compensatory picks.
A few more parting notes, before wrapping this thing up:
• The Patriots were the clear leader in accumulating draft picks with 143, or about 22 more than the 28-team average
• The Steelers were the leader in the number of 5-year starters and did it by consistently exceeding the expected results through all but 1 of the Value Groups
- Only in VG 4 were they slightly below the expected number of 5-Year starters
• There were several "whiffs" in the draft results where a team was "0 for Value Group"
- Most notable was in VG7 where the Bears, Dolphins, Ravens, Saints and Seahawks were a combined 0 for 131.
- Earliest whiffs were Chargers and Falcons in VG2, but with only a total of 5 picks between them
And really finally, Draft Metrics divided the Value Groups into 3 categories: Early (VGs 1-3), Mid (VGs 4-5) and Late (VGs 6-7) and identified the best and worst in each group and is shown below. The appendix to this article reflects results by Value Group for each team.
• VGs 1-3:
- Most Picks: Rams (28), Bengals (27), Cardinals (27), Patriots (27)
- Fewest Picks: Chiefs (16), Dolphins (16), Broncos (18)
- Best Batting Average: Rams (.750), Steelers (.714), Seahawks (.708)
- Worst Batting Average: Bengals (.370), Chiefs (.375), Eagles (.400)
• VGs 4-5:
- Most Picks: Cowboys (42), Bears (39), Titans (38)
- Fewest Picks: Ravens (22), Raiders (23), Saints (23)
- Best Batting Average: Eagles (.353), Steelers (.344), Cardinals (.323)
- Worst Batting Average: Bengals (.111), Chiefs (.161), Raiders (.174), Saints (.174)
• VGs 6-7:
- Most Picks: Packers (81), Patriots (80), Steelers (78)
- Fewest Picks: Redskins (56), Lions (58), Bengals (59)
- Best Batting Average: Bengals (.203), Packers (.160), Steelers (.141)
- Worst Batting Average: Chargers (.028), Cowboys (.029), Eagles (.029)
Note: This article was originally published on Tony Villiotti's website, DraftMetrics.com and is being reprinted here with the permission of the author.
Draft Metrics was established in 2010 but its roots were planted long before. Villiotti's obsession with the NFL Draft began in 1969.
Over the years, his interest shifted from predicting draft choices to trying to better understand the draft's importance by examining its eventual outcomes.