Baseball Notes

In the SABR community, a lot has been made in the recent past about the shortcomings of stats like batting average and RBIs; how they don’t tell the full offensive story and that other, more complex rate stats like OPS or VORP or EqA do a better job of encompassing a player’s offensive worth.

As you know, my favorite of the “obscure” statistics is VORP (and you can search posts on here to learn more about it if you are unfamiliar). To me, it fully explains how a player performs offensively above what a team could do or would have to do in his place.

Batting average is misleading because you have many hits that get by, let’s say, a poor defensive shortstop where an above average SS would have made the play. Or what about a ball hit to centerfield that is routine to Andruw Jones, but might cause Chris Burke to run the wrong route, he fails to reach the ball, and the batter still gets credit for a hit since there was no play on it. Slugging percentage is good for determining power, but you are comparing all player’s SLG equally, where it is universally known you expect your first basemen and corner outfielders to hit for more power and your middle infielders to be more contact/speed guys. On base percentage works the best out of those three, but whether you get a walk, HBP or homerun, it all counts the same towards your OBP total. OPS is a good indicator of overall value, i.e. how much you get on base plus how much power you have. But the problem here is that they are both weighted equally since you just add the two totals. But in a bottom of the ninth, one out situation, isn’t a triple more valuable than a walk? Isolated power (ISO, SLG minus AVG) is newer and interesting, but you then get obscure cases like Adam Dunn or Jason Giambi with low AVG and high SLG or Freddy Sanchez with high AVG and low SLG.

So while I am sure this has been done before somewhere, I though I might take four commonly used stats (AVG, OBP, SLG, and OPS), and correlate them to VORP to see what happens, and which of these comes out to be a truer indicator of offensive value. This is by no means a proper test nor is it valuable anywhere else other than this site, but I thought I would give it a try.

First we need to define our sample group. Just using the top 25 or so players by VORP in 2006 would skew results because you are just using the best players, who will usually have high combinations of average, OBP, and SLG. So I thought I would take the top 200 players in 2006 by VORP and, starting with number one, take every eighth player, giving me a 26-person sample size. Doing this, we get this group of players:

Albert Pujols 85.4
Grady Sizemore 69.1
Vladimir Guerrero 63.9
David Wright 54.3
Robinson Cano 49.1
Ichiro 46.4
Carl Crawford 41.1
Michael Cuddyer 36.3
Adrian Gonzalez 32.8
Orlando Hudson 30.8
Omar Vizquel 28.0
Paul Lo Duca 27.2
Tadahito Iguchi 25.5
Adam Dunn 23.5
Dave Roberts 22.2
Ty Wiggington 20.5
Nick Markakis 19.4
Jose Lopez 18.3
Brian Giles 17.7
Ronny Paulino 16.4
Ronnie Belliard 15.9
Gregg Zaun 15.3
Chone Figgins 14.2
Mike Napoli 13.0
Mike Jacobs 12.2
Jeff Cirillo 11.1

These 26 players give us a good cross section of outfielders, infielders, and catchers; superstars, above average, and regular players. You have NL and AL, lefties and righties, power and speed.

Correlating each of these players’ 2006 VORP to AVG, OBP, SLG and OPS, you get the following results:

This chart provides a small glimpse of clarity that shows batting average, as so often thought, is probably not the best predictor of a hitter’s true value, at least for this small sample size.

Admittedly, this is a small sample, so let’s expand it some to see if we get different outcomes. But, instead of adding more players from the same year, I will use data from 2004 and 2005, and stick to the same parameters we used for 2006.

The results for 2004:

And for 2005:

As you can see, the theory holds true for each of the past three years. I suspect if we were to go back another three or 13 or 30 years, we would find the same trend as well.

The lesson from this? Unless you are playing rotisserie fantasy baseball, ignore batting average when looking for the value of a hitter and look a little deeper. You find examples of this all over baseball. Take these stats from player X in 2006:

446 AB, 113 Hits, 106 Ks, .253 Avg

Not too impressive at first glance, is it? Well player X is actually Jason Giambi, cleanup hitter for the Yankees in the ‘06 season. Here are his other numbers:

37 HR, 113 RBI, 110 BB, .413 OBP, .558 SLG, VORP of 47.4.

And there are multiple examples like this one from every season. One last classic example is Mark McGwire from 1990. That year his average was a miserable .235. Nonethless, he had a VORP of 34.0 (top 40 in the majors). Compare that to Brad Ausmus and his .230 average from 2006. Given that his average was five points lower, Ausmus’ VORP for the year? Negative 17.5.

And thanks to Jared Benge, my resident brainiac and statistician, for checking my work.