Math Rushers: Waldo
“Box score scouting” is as or more dangerous in the sport of football as any other on Earth. With 22 moving parts on any given play, it’s hard to come up with a metric to compare and/or contrast players that aren’t influenced by the system the player plays in, the strength of his opponent, or the strength of the teammates next to him.
The most important day for football evaluations comes in the spring: the NFL Draft. Teams hedge their futures on players. One single player can elevate the entire franchise to another level, ending up in raises or championships for those around them, as well as for himself.
On the contrast, some set a team back for years as they’re forced into the lineups due to their investment value. Imagine how different Buffalo would have been like had they taken Brian Orakpo or Clay Matthews, two future Pro Bowl rushers, over Aaron Maybin, who started one game in his NFL career. Their 11th overall selection was spent on someone who is now out of the league. He was thought to be a “value pick” at the time. Clay Matthews, who the Packers traded up in the back end of the first round for, was considered a reach. He signed a deal worth nearly 70 million dollars last spring.
If there were a time when a football metric measuring an independent player’s skill would be important, it’d be draft day.
In 2011, a poster on Football’s Future by the name of Waldo had come up with something impressive. That something was a metric for 3-4 OLBs, in essence, pass-rushers. He predicted the risk/reward of a pass-rusher based solely on Combine numbers, and when absolutely needed, Pro Day numbers. The numbers give us the isolation that is so hard to find on the football field. In his post, he explained why, by his measurement, Cameron Wake was destined to be a great rusher coming out of Penn State, and why Justin Houston was a top tier rusher, not a third round pick like he would eventually end up as.
In an attempt to make Waldo’s formula even better, I added some criteria of my own, to see if I could make it more efficient. Moving on with this project, I will not directly spill the beans on how I run the formula/filters/whatever you want to call it to get the results for this test. For those with lack of faith in me, turn around now before investing your time. For those who believe me, just know that it’s based loosely on those Waldo numbers.