Posts Tagged ‘analytics’

Jul
25
2009

ArticlesFun with Data – NFL Combine Results

2861856437_7037474353[1]Strength, Speed, Agility – what more could you ask for in an athlete? (besides skill and talent, I suppose). The NFL combine results do a great job of serving as a benchmark for athletics, and it gets especially interesting when you start to segment results by position, and examine correlations between the various tests.

Due to my interest in sports, and my nerdy tendencies, I’ve gone through and charted out the results.

The reason for doing this followed from an article I read where the following was stated:

It’s also rare to see someone with a 30 inch or better vertical jump who can’t run a sub 5.0 forty.

The results seem to support this statement- Examine where the best fit line crosses the 30 inch plane below (at 5.00 sec). (click on the chart to see a larger version)

40 time vs vertical leap

40 vs. vertical.small

The following chart has a very high R^2 value – showing that 72% of the of the performance in broad jump can be explained by the performance in the vertical leap. This makes sense.

vertical leap vs broad jump

vertical vs broad jump.small

The final chart shows who is really scary – Look for the outliers. Those in the 4.5 second range with over 25 reps on the bench are people that are different from you and I.

40 time vs bench reps – determines who is really scary

40 vs. bench.small

From each of the graphs, there are formulae that can be used to predict what the results in one test will be based another. Below are the formulae, accompanied with the R^2 values, or the explanatory power of the formula. The higher the R^2 value the greater the explanatory power (Max of 1.0).

Chart Formulae

formulas2

As well, a table of the raw combine data (which the NFL does a bad job of doing) – The names also link to the player profiles, though I suspect these links will break once the NFL draft page is removed. You can sort by column by clicking on the headers. (click the Read More link to see the tables)

(more…)