I find myself increasingly frustrated with OPS as a measure of offensive performance. I grossly overvalues singles(they are not twice as valuable as walks) and undervalue HR's in relation to singles(2.5X) but have it about right or slightly overvalued in relation to walks(5X). That is compounded by adding two ratios with different denominators(PA for OBP, AB for SLG%). It also ignores the contribution or lack thereof of SB's in offensive performance. What a mess!!
I got to thinking, how could you combine the best parts of OBP and SLG% into one % number that would be easy to understand and give a more accurate value of the players offensive performance?
In my profession, they have something called Relative Value Units(RVU) which are supposed to quantify the relative value of different medical activities for the purposes of calculating reimbursement. For instance, a typical office visit to a doctor for a common cold may have an RVU of 0.8 while a complicated new patient with multiple chronic health problems might be weighted with an RVU of 2.0 while a complicated surgical procedure that takes several yours may have an RVU value of more than 10.
The problem with the RVU system is that it may not be weighted properly. Right now, it appears that outpatient office care is undervalued in relation to procedures which is exacerbating a growing shortage of primary care physicians. Still, if the weighting problem can be corrected, it's theoretically a good system.
I decided to assign a RVU value to several offensive parameters and then use TPA as a denominator to arrive at a single number% that is easy to understand and possibly reflects more accurately the true offensive value of the player.
I am sure someone has calculated the relative value of these parameters in terms of run production somewhere, but I don't know where to find it. So, I made some educated estimates that ore open to revision. Here they are:
BB/HBP= 1. This seemed like a good starting point because it equals one base gained.
Single= 1.2. A single is more valuable than a BB, because it is more likely to advance runners who are already on base. How much more? Should it be 1.3? 1.4? I chose 1.2 because it seemed right.
Double= 1.8. A double is more valuable than a walk and a single, but almost certainly not twice as valuable. But it does immediately get a runner into "scoring" position and is more likely to drive in runs than a single. I decided on an RVU of 1.8, but this is open to adjustment.
Triple= 2.5. Similar arguments as for doubles. Is this too low? I thought a triple is more like a double than a HR.
HR= 4!!. The Big Kahuna! HR is the only offensive event that produces a run without the help of another batter. About 40% of all runs scored in baseball are the result of HR's. I gave it a full 4. Should it be even higher? I thought about giving it as much as a 5!
SB= 0.5. I think of a SB as turning a walk or single into a double, except it is not quite the same since a double can advance other runner and drive in runs while a SB does not. O.5 seemed about right.
CS= -1.0. A CS erases a walk or single, except the single might have had prior benefit of advancing runners. I settled on a -1.0.
SF= 0.5. A SF drives in a run, but also produces an out. Runs are very precious but so are outs. I split the difference.
SH= 0.2. A SH advances a runner but results in an out. A small net positive. Some may argue this one, but the effect is generally negligable due to relatively few events.
GIDP= -1.9. A GIDP result in an out plus it causes an additional out negating a teammante's single or walk. A DP can occassionaly advance a runner or even drive in a run, so I didn't go all the way to -2.0 but close. Is it an even more catastrohic event?
Anyway, I thought that was enough to do some calculations. I looked up a variety of offensive players to see what effect it might have on how their offensive contribution is stratified. I simply looked up the offensive events for each player, mulitiplied each event by it's RVU and then divided by TPA. Here's what I found:
ARod: OPS= 1.067. RVU%= .691. No surprise here. ARod had the highest OPS and the highest RVU%.
Prince Fielder: OPS= 1.013. RVU%= .667. Again, no surprise. OPS and RVU% assign about the same relative offensive value.
Hanley Ramirez: OPS= .948. RVU%= .604. Sill not a huge surprise. Hanley had a lot fewer HR's than ARod and Fielder, but had a lot more singles, doubles, triples and SB's. Note that the relative difference between RVU% is greater than for OPS due to Fielder's superior power. Did I overvalue HR's in RVU%?
Adam Dunn: OPS= .940. RVU%= .607. Dunn had a slightly lower OPS, but a higher RVU% mainly due to more HR's. which more than made up for lack of speed and huge difference in BA.
Here's an interesting group:
Ichiro: OPS- .827. RVU%= .510.
Kevin Youkilis: OPS= .818. RVU%= .525. Youkilis has a much lower BA(.357 vs .288) and Ichiro even had a higher OBP(.396 vs .390) but Kevin had more HR's(16 vs 6) and even more doubles(35 vs 22) for a better RVU% in spite of far fewer SB's and more GIDP's. Which is the better number here, OPS or RVU%?
Dan Uggla: OPS- .805. RVU%= .517. Uggla had both a low BA(.245) and a low OBP(,326) but had a lot of value due to all the XBH's(49 doubles, 31 HR's). More than one way to skin a cat?
Chris Young: OPS= .763. RVU%= .517! Here's a guy who's OPS may grossly undervalue him. He had a terrible BA and OBP(.237/.295) but just raked the stuff he did connect with(29 doubles, 32 HR's) He helped himself with a good SB%(27/33) and did not hit into DP's. Which is a better reflection of his value, OPS or RVU%?
Here's an interesting guy:
Curtis Granderson led the league in triples with 23! OPS= .903, RVU%= .598! Despite mundane HR's(23) and OBP(.361), he helped himself a lot with the triples and also with a 26/27 SB's. Note that his RVU% is right there with Dunn and Hanley and his OPS my undervalue him even though it is very good.
What do you all think?