Inside the Numbers: Power Ratings
by Will Stewart, TechSideline.com
TSL Extra, Issue #17
Last year, about this same time, we ran a nifty article on power ratings for VT football players, and we've returned for a second annual Power Ratings article.
The premise is simple. Message board poster "Technocrat" devised a way to calculate power ratings for Virginia Tech football players based on their height, weight, and 40-yard dash times. You can build a power rating formula that is as simple or as complex as you want it to be. Technocrat's power rating is pretty simple, from the standpoint that it only requires three points of data as input: height, weight, and 40 time.
You could devise a more complex formula that incorporates weightlifting numbers, vertical leap, shuttle times, etc., but we like the simplicity of Technocrat's formula, so it's what we go with. We're not sure exactly what it means, but you can boil it down to this: shorter, heavier, faster guys will get a high power rating; tall, skinny, slow guys won't. Think of a bowling ball hitting you at a high rate of speed versus, say, a beach ball rolling slowly into you.
What follows is an explanation of Technocrat's formulas, and then I'll give the results so you can find out which players scored highest in his power rating system. Since this is our second year doing this, will provide some comparisons to last year's figures and let you know who improved their power rating the most.
As always, those of you who want the executive summary can skip ahead to the results (see the paragraph titled "The Numbers") to see how the players rank. The engineering and scientific types out there can muddle through the next paragraph (titled "The Formulas") on your way to "The Numbers."
As is always the case with "Inside the Numbers," the results are interesting and illuminating, and you can learn a lot about the VT football players just from examining the numbers.
Remember (one last warning), if the technical stuff puts you to sleep, you can jump right ahead to the next section ("The Numbers").
Technocrat's power rating formula is simple. He calculates a mass component and divides it by a speed component to give his final power rating. So if a player's mass component is 180.0 and their speed component is 60.0, their final power rating is 180/60 = 3.000.
The speed component is created by squaring a player's 40-yard dash time and then multiplying by a "speed weighting" factor, which I'll describe later:
Speed Component = [(40-time)^2] * Speed Weighting Factor
To derive the mass component of the power ratings equation, you must first calculate a player's Body Mass Index (BMI). BMI is calculated by converting the player's weight to kilograms (divide pounds by 2.2 kilos/pound) and their height to meters (divide total inches in height by 39.37 inches/meter). You then divide weight (in kilos) by height (in meters) squared, and this gives you the player's BMI:
BMI = Body Mass Index = Weight in kilos / (Height in meters)^2
By the way, Technocrat didn't just make BMI up. Go to yahoo.com or your favorite search engine and do a search on "body mass index" and you'll get at least one link to the formula shown above.
To get the mass component of the power rating, you multiply the BMI by a "mass weighting factor."
Mass Component = BMI * Mass Weighting Factor
Lastly, to get the final power rating for the player, you divide the mass component by the speed component:
Power Rating = Mass Component / Speed Component
So what are the "speed weighting factor" and the "mass weighting factor"? They are numbers that you use to assign more "weight" to either the mass part of the equation or the speed part of the equation.
So if your mass weighting factor is 2 and your speed weighting factor is 1, that doubles a player's power rating over a 1:1 ratio. Likewise, a mass weighting factor of 1 and a speed weighting factor of 2 cuts the player's power rating into half that of a 1:1 ratio.
Note that changing the weighting factors does not alter how the players rank relative to one another, because they're just multipliers. If you change them, they affect all players' ratings proportionately. So if player A has a higher power rating than player B, his rating will remain higher no matter what you do with the weighting numbers.
The original spreadsheet that Technocrat sent to me had a mass-to-speed ratio of 5 to 3, so he chose to give more weight to the mass portion of the equation. That's fine with me, and as I mentioned, it doesn't affect how they rank with respect to one another.
So let's start crunching some numbers!
Again, this formula only has three input variables: height, weight, and 40 time. All of the data for the spreadsheet that Technocrat sent to me were taken from "Gentry's Iron Palace" on BeamerBall.com, for Winter Max testing (done in the month of February).
Of course, some players were out with injuries and were not able to test, so the figures entered for those players were the most recent ones available (in most cases, fall of 2001). This information is also available on BeamerBall.com.
Technocrat did all the research and entered all of the data into the spreadsheet before sending it to me, so any errors in data entry can be attributed to him. Feel free to berate him on the message board if you find any mistakes.
Having said all that (drum roll, please), here are the power ratings for the top 15 players on Virginia Tech's spring football roster. Their ranking last year, if applicable, is also included:
Seven players from last year's top 15 graduated, and seven of the other eight -- Colas, Burnell, Robinson, Wilson, Lewis, Adibi, and Suggs -- are present in this year's top 15.
The only player to drop out of the top 15 is fullback Marvin Urquhart, who fell to 20th this year. Ironically, Urquhart lost 10 pounds and shaved his 40 time from 5.00 to 4.93, but the result was that his power rating dropped, from 2.504 to 2.439.
Defensive end Cols Colas is the reigning power rating champion for the second year in a row, turning in a nearly identical power rating to last year. Colas' height and 40 time are the same as last year, but his power rating dropped slightly because he lost a pound from last year, going from 240 to 239.
Last year, Wayne Briggs and Jarrett Ferguson were right on Colas' heels, but this year, the gap is much wider between Colas and the #2 finisher, Keith Burnell. Burnell, who is moving from tailback to rover this spring, increased his weight from 206 to 207.5 and dropped his forty time from 4.28 to 4.21 … and, believe it or not, lost a full inch of height, from 72 inches (6-0) to 71 inches (5-11).
Vegas Robinson holds steady from last year, and Joe Wilson makes a move from #10 to #4, based mainly on dropping his forty time from 4.78 to 4.65.
Posting an incredible improvement is fullback Josh Spence, jumping from 34th last year to 5th this year. Spence gained 10.5 pounds, dropped his forty time from 4.56 to 4.51 … and lost one and a quarter inches, from 73.00 to 71.75.
So that's obviously one way to increase your power rating: shrink.
Jason Lallis, moving up from #30 to #6, did not shrink, but he did gain 6 pounds and drop his forty time, from 4.82 to 4.66. Kevin Lewis comes in at roughly the same place as last year.
Newcomers from the 2001 recruiting class into the top 15 are DeAngelo Hall and Tim Sandidge. The coaches have raved about Hall since he signed, and Sandidge's high power rating supports the positive comments they have made about him. They are trailed at #10 by Nathaniel Adibi, who makes a slight move from #11 to #10.
Out of the remaining five players in the top 15, Anthony Nelson makes a move into the top 15, and Sam Fatherly posts a huge jump from #49 to #13, based on dropping his forty time from 4.60 to 4.39.
Which players made the biggest increase in their power rating from 2001 to 2002? Here are the top 15 in terms of percentage improvements.
Note the high number of young players in this list. Many of them were from the 2001 recruiting class and thus made the expected big improvement from the fall to the spring. In addition to those young players are some older players the Hokies are hoping will step up in the 2002 season: Jason Murphy and Tim Sandidge at DL, Mike Daniels at Whip linebacker, Chris Clifton at QB, and Josh Spence at FB, and Ernest Wilford at WR.
The data that went into this article are available as a web page or a Microsoft Excel 97 spreadsheet. The data include not just current scholarship players, but walk-ons and members of the 2002 recruiting class (using publicly available data, since the new recruits have not been tested by VT yet).
To see the full list of players ranked by power rating, check out this web page:
To download the data in Microsoft Excel 97 spreadsheet format, head to this link:
(Right-click the link and do a "Save Link As" or "Save Target As" to save the Excel file to disk.)