Quantize Analysis by Adam Gabriel
It is that time of year where all 32 teams migrate to Indianapolis for the yearly NFL combine. Many NFL scouts and player personnel executives start to struggle with mentally weighting various facets that go into ranking players for the NFL draft. For example, how much value should players with fast 40 times receive versus slower players? You would expect that if the 40 really is a good measure of future success, players with fast 40 times at the Combine would go on to have good careers. But it is difficult to know just how much weight to place on a test through human observation.
Graph#1: 3 cone, Bench, 20, and 10 not significant predictors
This study uses forward stepwise regression and analyzing six years of NFL drafts and Combine results for over 180 cornerbacks. The results for the 10-, 20-, and 40-yard dash, standing broad jump, vertical jump, short shuttle, bench press and Wonderlic test were categorized for each player as was the player’s selection in the subsequent draft.
The graph above suggests that NFL teams as a whole view the 40-yard dash as the most important test followed by the short shuttle, the broad jump, the Wonderlic and then vertical jump. This does not mean that every team views the tests in the same way, but the graph does show how the NFL collectively used the data to rank players on draft day. The fact that a faster 40 is important for cornerbacks to be drafted earlier is not a shock to anyone who follows football. However, the regression tells you in a systematic way to what degree this happens.
Without going into a heavy discussion of statistical details, the resulting probabilities suggest it is highly unlikely that the order of importance of these tests occurred by chance. For example, we would expect that if the bench press really was more significant to NFL draft order, it would show up in the results.
*For the six-year period that was examined, when a player skipped a Combine test, his pro-day result for that test, if available, was included in order to complete the study.
Graph# 2: 3-cone, Bench, Wonderlic, 20, and 10 not significant predictors
Comparing the Combine test results with the players’ subsequent NFL careers to identify the most important predictors of average number of NFL games played. If indeed the 40 is an important indicator of future success, we would expect that players who recorded fast 40 times would be more likely to have success in the NFL. And, in fact, that is the case: NFL teams do an excellent job in correctly weighting the 40 as the most important predictor of cornerback success.
The results then deviate a bit. The data points to vertical jump being the second most important test for predicting average NFL games played, but NFL teams on average are actually weighting it too lightly when making draft selections. In addition, the Wonderlic has little predictive ability for corners, even though teams are weighting it as though it is the fourth most important factor.
The blue bar of the graph below shows current NFL Combine test weighting results in teams drafting cornerbacks who will play fewer average games/year than if the tests had been weighted optimally. The purple and red bars demonstrate that a team will draft 1.2 percent more talent at corner by increasing its weight on vertical jump and will gain another 0.4 percent advantage by decreasing its weight on the Wonderlic, for a total improvement in talent drafted of 1.6 percent. It is unlikely that this data occurred randomly and thus it is possible to use this information to adjust draft order.
The previous graph indicates that current NFL weighting of Combine tests is not completely efficient. Simply put, this study demonstrates that you will draft more talented cornerbacks by paying more attention to the vertical jump and ignoring the Wonderlic. In this study, more talent equates to more games played per year on average.
If a franchise makes adjustments based on these results, instead of drafting about 3 percent of the cornerback talent pool as one of 32 typical NFL teams in a typical NFL draft (100 percent divided by 32 teams=3.1 percent), over time you will capture nearly 5 percent of the available cornerback talent. That seems small but it is in fact an improvement in talent evaluation of more than 50 percent (1.6 percent divided by 3 percent.) Given the dynamic nature of the draft, that is a significant competitive advantage.
Keep in mind when teams struggle with positional need and draft board placement, if they are smart they will go back to a player’s DNA (film)to assign the proper grade, but with that said they must not ignore the data in front of them including the player’s body of work throughout their career and the quantize analysis staring them right in the face.