What if a guy took his turn at bat every time through the lineup and nobody noticed?
In these stat-happy, hyperanalyzed times, when at a glance we can see how this guy did in night games in 2003 or whether that guy has hit any home runs on 2-2 counts this year, what if there was a guy who took his cuts and was for the most part ignored.
There is such a guy. He plays in the National League. He's called the pitcher.
Pitchers' hitting has to be the single most disregarded statistical category in baseball. Fielding is hard to measure accurately, but at least people try, and fielding gets discussed. You can even find fielding stats fairly easily on the Web, for what they're worth.
But try to find pitchers' hitting stats. You can do it, but it's not easy on most of the big mainstream sites where current-year stats are found.
ESPN.com lets you see the stats for players at any position -- except pitcher. At both Yahoo Sports and FoxSports.com, you're invited to click on any position to see stats for those player, but if you click on "P" you get pitching stats.
CBS.SportsLine.com won't let you see just pitchers' hitting stats, but you can look at all players' stats and sort by position, and the pitchers will all be together, divided into starters and relievers. But you can't sort by any other category and compare only pitchers. SI.com and MLB.com both show you pitchers' hitting as its own category.
And here's the sum total of all you'll ever hear about pitchers' hitting in the broadcast realm: "He looks like he handles the bat pretty well." This is said by a TV or radio announcer, usually the analyst, after the pitcher has swung and missed or fouled off a pitch without looking like Aunt Sally swatting mosquitos after one too many high-balls.
We aim to change that around here. We're going to come up with a new stat. We're going to call it PRV, for Pitcher Run Value, though the only reason we're calling it that is because we're hoping people will pronounce it out, as in "He's the National League perv leader."
And when I say "we" I'm not just following a fusty journalism custom. I mean we. You and me. You're going to help me. Especially if you're good at math and computers and you like baseball. And that third one isn't even required.
Pitcher run value will be the only mainstream statistic -- once we've sold it -- that measures a pitcher's entire contribution to the team effort. Essentially, it will take his offensive production -- meager in most cases -- and subtract it from his earned run average. The result will look like an ERA, but lower.
Let me be frank, and I don't mean Francisco, who has never batted. PRV is not going to rock your world. It won't turn some clown who can't pitch into the equal of Sandy Koufax if he can really turn on an inside fastball. It's just a way of focusing the picture a little.
I also want to say I'm not the genius discoverer of the fact that pitchers bat. The so-called Davenport Translations, developed by Clay Davenport and available at Baseball Prospectus, measure a pitcher's batting and fielding contributions using the stat EQR, or equivalent runs.
But, with all due respect, Joe Bleacher doesn't get much out of hearing that Tom Glavine has 4.4 EQR so far this year. Since PRV is an adjusted ERA, it's easy to digest.
Why do this? Because N.L. pitchers really do bat. And some of them help their team when they do. American League ERAs are higher than National League ERAs because A.L. pitchers have to face nine hitters while N.L. hurlers face eight hitters and an almost-automatic out. If a pitcher can hit, even if he just hits like a weak-hitting middle-infielder, he turns his team's N.L. lineup into an A.L. lineup.
And that means you should look at his ERA the way you'd look at an A.L. starter's ERA. That is, it can be higher than another N.L. pitcher's ERA, and still be just as good or better.
For lousy hitters, the guys who eke out a hit every few months, pitcher run value will lower an ERA by a meaningless point or two, say from 4.43 to 4.42. But for those few guys who can really, ahem, handle the bat, it might knock a half a run off. That 4.43 ERA might become a 3.93 PRV. That's quite a difference.
What I'm saying is that not all pitchers with 4.43 ERAs are created equal, if some of them can hit. I should also say that fielding should be a part of this equation too, but I am technologically, conceptually, mathematically and even physically unable to take fielding into account at the moment.
When you volunteer to help me, maybe we can figure out how to work fielding in to get the most complete picture of a pitcher's contribution to his team on the day he starts.
Here's what we'll do. We'll take the stat runs created. It's an attempt to measure how many runs a hitter is responsible for. There are a lot of formulas, some of them extremely complicated, but for now I'll use the simplest, which is a hitter's total bases divided by his plate appearances, times the sum of his hits and walks.
This formula might undervalue pitchers' offensive contributions because it ignores something they do a lot, sacrifice bunt. That's OK. We can use a more complex formula when you've fired up your spreadsheet and you're running the numbers for me. And anyway I don't think a pitcher's bunts account for many runs in a year.
We'll take the pitcher's RC and divide it again by his plate appearances to figure out how many runs he produces each time he bats. Then -- this is the tricky part -- we'll multiply that to show how many runs he would create per nine innings pitched. That's so it'll match his ERA, which measures not how many runs he allows per appearance or per inning, but per nine innings. We'll call this stat RC/9.
Now, figuring out the multiplier took some figuring out. Most starting pitchers -- they're the only ones we're talking about here, since relievers rarely bat -- hit about 2.5 times per game, tops. But then, most starting pitchers only throw about six innings a game, give or take a few outs. So multiplying by 2.5 would tell us what they contribute per game, but not per 9 innings.
I consulted with several smart people who have written or contributed to books and magazine or Web articles about esoteric baseball stats, and on the advice of one of them, Tom Tango, I settled on 3.5.
Tango, a well-known sabermetric analyst and a co-author of "The Book" -- and a person who does not endorse the creation of this stat -- told me he figured that in recent years, starting pitchers averaged about that many plate appearances per nine innings pitched. Some spot checking confirmed at least that it's not an unreasonable number.
So, once we figure out how many runs a pitcher creates per plate appearance and multiply it by 3.5 to get how many runs he creates per nine innings pitched, we can subtract that number from his ERA to get his PRV, which remember you pronounce like a word.
It's not a perfect measure. It's a blunt instrument. Most of the smart people I consulted as I started thinking about PRV pointed out that it would be a pretty useless stat, because it would deal in such small sample sizes, a few dozen plate appearances per year, and therefore wouldn't be very predictive.
As an example, Carlos Zambrano of the Chicago Cubs is a career .230 hitter who last year hit .300 with six doubles, a triple, two homers and a slugging percentage of .462. In short, he was a better hitter than most National League catchers. This year, he's 1-for-22. Did he suddenly go from good hitter to lousy one, or is he just in a little slump? It's a third of his season, but for a position player it would be the equivalent of one bad week.
But if PRV is a blunt instrument that isn't predictive, so is ERA. It won't necessarily tell you who the best pitcher is, but it'll give you a decent idea. There's some luck involved in posting a low ERA. But even though ERA isn't very good at saying whether a pitcher is likely to do as well next season, it does help tell the story of what actually happened this season.
PRV is the same. It makes the picture a little clearer. National League pitchers do bat, and if one can contribute in that area, it makes him the equivalent of a better pitcher.
I'll show you what I mean. Here are the top 10 starting pitchers in the National League by ERA, through Sunday:
Pitcher | ERA |
1. Webb, Ari. | 2.01 |
2. Arroyo, Cin. | 2.40 |
3. Martinez, N.Y. | 2.50 |
4. Glavine, N.Y. | 2.59 |
5. Penny, L.A. | 2.62 |
6. Carpenter, St.L. | 2.63 |
7. Lowe, L.A. | 2.68 |
8. Schmidt, S.F. | 2.70 |
9. Myers, Phi. | 2.90 |
10. Oswalt, Hou. | 3.11 |
And here are the top 10 starting pitchers by PRV:
ERA | RC/9 | PRV | |
1. Webb, Ari. | 2.01 | .073 | 1.94 |
2. Arroyo, Cin. | 2.40 | .240 | 2.16 |
3. Glavine, N.Y. | 2.59 | .320 | 2.27 |
4. Penny, L.A. | 2.62 | .235 | 2.38 |
5. Martinez, N.Y. | 2.50 | .029 | 2.47 |
6. Carpenter, St.L. | 2.63 | .007 | 2.62 |
7. Lowe, L.A. | 2.68 | .033 | 2.65 |
8. Schmidt, S.F. | 2.70 | .031 | 2.67 |
9. Myers, Phi. | 2.90 | .007 | 2.89 |
10. Oswalt, Hou. | 3.11 | .023 | 3.09 |
Not much change there. In fact, it's the exact same cast of characters. Look, I told you PRV wouldn't rock your world. Somebody already invented, like, slugging percentage, OK?
But look how Tom Glavine and Brad Penny leapfrogged Pedro Martinez. By ERA, Pedro is the Mets' best starter by a tenth of a run. But counting their hitting, Glavine's better by a fifth of a run. And look at the distance Bronson Arroyo puts between himself and Martinez with his hitting.
Here's a good time for a reminder about sample-size issues. I'm not ready to rate Bronson Arroyo ahead of Pedro Martinez because of Arroyo's hitting. Arroyo increased his RC/9 from .109 to .240 with two singles and a double in three trips Saturday.
But listen, you'll take Arroyo or Glavine the way they're pitching this year even if they make Aunt Sally's fly-swatter look like Wonderboy. Or if they make Pedro "Aunt Sally" Martinez look like Babe Ruth with the stick. The way these 10 guys are throwing, any hitting they do is gilding the lily.
Where PRV gets interesting, I think, is when you look at pitchers who aren't mowing down the league. Look down the ERA list and you might find some guys who are more valuable than you think because they can use their bats to mitigate a little of the damage their arms do.
For example, here are two guys with roughly equal, below-league-average ERAs, but notice how one of them makes himself a little better -- by about seven runs, or three quarters of a win, over the course of a 200-inning season -- by helping himself with the stick.
ERA | RC/9 | PRV | |
1. Hernandez, Wash. | 5.10 | .331 | 4.77 |
2. Suppan, St.L. | 5.06 | .016 | 5.04 |
If you're a National League general manager looking at picking up one of these two guys, wouldn't you consider Livan Hernandez's offense -- not to mention fielding, which we're ignoring for the moment -- when trying to decide between them? I would. Why not? Every little bit might make a difference. And you just never hear about this.
That's going to change. Drop me a line -- I'm serious here, my e-mail address is at the bottom of the column -- if you can help me refine this stat, and especially if you can help me run the numbers every once in a while.
In the meantime, I'll give you the National League top 10 in RC/9, minimum 20 plate appearances. To provide some context, if we multiply Albert Pujols' runs created per plate appearance by 3.5, we would get .905. If we did the same for Randy Winn, whose batting, on-base and slugging averages are all almost exactly league average, we'd get .432.
RC/9 | ERA | PRV | |
1. Park, S.D. | .463 | 4.26 | 3.80 |
2. Mulder, St.L. | .437 | 4.71 | 4.27 |
3. Thomson, Atl. | .383 | 4.65 | 4.27 |
4. Hernandez, Wash. | .331 | 5.10 | 4.77 |
5. Glavine, N.Y. | .320 | 2.59 | 2.27 |
6. Sosa, Atl. | .280 | 5.07 | 4.79 |
7. Arroyo, Cin. | .240 | 2.40 | 2.16 |
8. Penny, L.A. | .235 | 2.62 | 2.38 |
9. Marquis, St.L. | .170 | 5.03 | 4.86 |
10. Wright, S.F. | .153 | 4.20 | 4.05 |
- - - - - - - - - - - -
Shares