### What Exactly is OPS?

Dan Fox has written yet another excellant article over at the Hardballtimes.com:

Run Estimation for the Masses

While many know that OPS correlated suprisingly well with runs scored, most don't know exactly why it works as well as it does. I certainly didn't. While I'm familiar with maybe half a dozen run estimation formulas that are intuitive, OPS really doesn't fall into that category. While OXS (and by extension, Bill James' Basic Run Created) just feels "right", adding OBP and SLG seems like a messy shortcut. They don't even use the same scale, afterall(OBP goes from 0 to 1, SLG from 0 to 4).

In actuality, OPS simply mirrors the many run estimation equations that use linear weights. Dan Fox shows that with a little algebraic manipulation, OPS can be written in the form:

(.5 x single + .75 x double + 1 x triple + 1.25 x HR + .25 x W) x (4/PA)

The relative values given to each baseball event in OPS almost EXACTLY matches the weights usual used in linear weights based run estimation equations. Tangotiger found those relative weights (using data from 1999-2002) to be:

single:.47

double: .75

triple: 1.03

HR: 1.4

non-int BB: .33

Other than slightly undervaluing HR's and walks, OPS is almost dead on. Apparently OPS is alot "prittier" than I ever gave it credit for.

Run Estimation for the Masses

While many know that OPS correlated suprisingly well with runs scored, most don't know exactly why it works as well as it does. I certainly didn't. While I'm familiar with maybe half a dozen run estimation formulas that are intuitive, OPS really doesn't fall into that category. While OXS (and by extension, Bill James' Basic Run Created) just feels "right", adding OBP and SLG seems like a messy shortcut. They don't even use the same scale, afterall(OBP goes from 0 to 1, SLG from 0 to 4).

In actuality, OPS simply mirrors the many run estimation equations that use linear weights. Dan Fox shows that with a little algebraic manipulation, OPS can be written in the form:

(.5 x single + .75 x double + 1 x triple + 1.25 x HR + .25 x W) x (4/PA)

The relative values given to each baseball event in OPS almost EXACTLY matches the weights usual used in linear weights based run estimation equations. Tangotiger found those relative weights (using data from 1999-2002) to be:

single:.47

double: .75

triple: 1.03

HR: 1.4

non-int BB: .33

Other than slightly undervaluing HR's and walks, OPS is almost dead on. Apparently OPS is alot "prittier" than I ever gave it credit for.

## 0 Comments:

Post a Comment

## Links to this post:

Create a Link

<< Home