nmculbreth

05-14-2009, 12:11 PM

I came across this article this morning and thought I'd post it here for others to see.

http://www.hardballtimes.com/main/article/thou-shalt-sow-but-thou-shalt-not-reap/

Thou shalt sow, but thou shalt not reap

by Colin Wyers

May 14, 2009

On May 10, pitcher Micah Owings bought the Reds a shot at extra innings with a pinch hit home run. This brings up the perpetual question: Is Micah Owings best utilized as a hitter or a pitcher?

We should begin by talking a common language; it's possible to compare hitters to hitters and pitchers to pitchers with metrics like OPS and ERA respectively (albeit with the recognition that there are better alternatives to those as well). But to compare apples to oranges we need a pan-fruit metric. As longtime readers might expect, my metric of choice for this purpose is Wins Above Replacement. For those who aren't longtime readers, WAR is essentially an expression of how many wins a player is worth compared to the generally available replacements available, such as minor-league journeymen or free agents willing to work for the league minimum.

[For a longer discussion of WAR, see my previous series "How to measure a player's value," parts one, two and three.]

So let's look at Micah Owings in two parts: the pitcher who is and the position player who could be.

Owings the pitcher

There are other good hitting pitchers in the league, like Carlos Zambrano, Jake Peavy and Yovani Gallardo. This doesn't ignite controversy the same way that Micah Owings does when he swings the stick. This is because they have all established themselves as pitchers. Owings hasn't.

Owings owns a career 4.89 ERA and a career 4.80 FIP. These are not spectacular numbers, but they aren't terrible. Going into this season, Owings had accumulated 2.6 WAR in 250 innings, according to Fangraphs. The average baseball player (hitter or pitcher) will accumulate 2.25 WAR in a full season (about 120 innings for a starting pitcher).

Again courtesy of Fangraphs, his projected FIP-ERA going forward is 4.59; that works out to almost five runs per game. In a league where the average is 4.65, and in a hitter's park like the Great American Ballpark, that works out to a projected 1.33 WAR in 120 innings for Owings. (Probably a little better than that, since he would still accumulate some value from his hitting even as a pitcher.) So a below average pitcher, to be sure, but it's possible that the Reds don't have a better starting pitcher than Owings available.

Owings seems to be comfortably nested in the grey area where he's neither a great asset or a great liability as a pitcher. But what about as a hitter?

Owings the hitter

The short answer here is that we don't know how well he'd perform as a hitter. For the sake of intellectual curiosity, we can go ahead and step through some calculations that seem reasonable and should at least put us in the ballpark, but I want to start off right here with the caveat that the error bars on what we're doing here are huge.

Or to be more plain, this is basically a wild guess. An educated wild guess, maybe—is there such a thing? Anyway, if there is an educated wild guess, this is one.

And let's start off with the first of the two big, unsubstantiated assumptions that are required to make this guesstimate work out. Projection systems are largely based upon the premise of regression to the mean, which is that over time, extreme observations become less extreme. Given a small number of observations, we would assume that a guy who has hit poorly will hit better than he has so far, and a guy who has hit well will hit worse than he has so far. The more observations we have, the more we can have confidence in a player being above or below average as a matter of true talent, rather than it simply being due to random chance.

The issue with regression to the mean here is determining the mean. For the majority of pitchers, regressing their hitting performance to the mean results in simply laughable numbers, because pitcher hitting is really a separate thing from position player hitting, and that almost every pitcher is a well below average hitter. So simply feeding a pitcher's batting line into the typical projection system will not provide usable results.

What we are going to assume here is that, if Owings were to devote himself to being a hitter full-time, that he would regress to the league mean of position players hitting, not the mean of pitchers hitting. This has the benefit of making sense, at least. But it's a really unvalidated assumption, with few data points to go off of (outside of Babe Ruth and Rick Ankiel there's not a lot of evidence to go around, and those two are both special cases indeed). If I had a giant robot following me around helping me perform sabermetrics, he would be flashing warning lights and screaming "DANGER COLIN WYERS, DANGER!" But it's the best guess I have for the time being.

Using Sal Baxamusa's Marcels spreadsheet and Owings' career hitting stats thusfar, we get a forecasted batting line of .285/.343/.478, or about a .350 wOBA. That works out to roughly six runs (or a half-win) above average with the bat in 650 plate appearances.

Now, as it just so happens, the positional adjustment for a corner outfield spot in 650 plate appearances is right around seven runs. So if we're correct, Micah Owings could be about a league average player in the outfield, assuming he plays defense about as well as a typical corner outfielder. Can he do that?

Honestly, this is even more of a guess than his hitting projection. He's a young guy and apparently athletic (and he obviously has the arm to play somewhere like right field). So he could be. Or he could not be. I know I'm being vague here, but that's all I can be with the data at hand.

So, using a set of favorable assumptions, it looks like we could make a case for Micah Owings being a league-average player in the outfield. Again, very large error bars surround this entire endeavor and it's possible that he's an utter trainwreck as an outfielder and hits like a utility infielder. We really don't know for sure. But let's assume for a second that this WAG is essentially correct. Should the Reds try to convert him to an outfielder right now?

Transactional costs

There are real costs to trying to convert Owings to a full-time position player, such as figuring out who to displace from the outfield to make room for him and figuring out who to take his innings in the rotation. There's a lot of chaining going on here, and a lot of things to figure for the Reds.

There's also a lot of risk here. It's real easy to run some numbers and come up with "about 2-2.5 WAR," and another thing entirely to teach a pitcher to play the outfield, to have him abandon his craft and devote himself to something else entirely. There's the chance that in taking Micah the pitcher and making him Micah the hitter you're left with neither Micah at all.

And the Reds just moved into a three-way tie for first place in the NL Central this evening, which says they're probably not in a mood to take those kinds of risks right now. So Owings will probably stay in the rotation, and probably should stay in the rotation.

But boy, it's fun to dream about, isn't it.

http://www.hardballtimes.com/main/article/thou-shalt-sow-but-thou-shalt-not-reap/

Thou shalt sow, but thou shalt not reap

by Colin Wyers

May 14, 2009

On May 10, pitcher Micah Owings bought the Reds a shot at extra innings with a pinch hit home run. This brings up the perpetual question: Is Micah Owings best utilized as a hitter or a pitcher?

We should begin by talking a common language; it's possible to compare hitters to hitters and pitchers to pitchers with metrics like OPS and ERA respectively (albeit with the recognition that there are better alternatives to those as well). But to compare apples to oranges we need a pan-fruit metric. As longtime readers might expect, my metric of choice for this purpose is Wins Above Replacement. For those who aren't longtime readers, WAR is essentially an expression of how many wins a player is worth compared to the generally available replacements available, such as minor-league journeymen or free agents willing to work for the league minimum.

[For a longer discussion of WAR, see my previous series "How to measure a player's value," parts one, two and three.]

So let's look at Micah Owings in two parts: the pitcher who is and the position player who could be.

Owings the pitcher

There are other good hitting pitchers in the league, like Carlos Zambrano, Jake Peavy and Yovani Gallardo. This doesn't ignite controversy the same way that Micah Owings does when he swings the stick. This is because they have all established themselves as pitchers. Owings hasn't.

Owings owns a career 4.89 ERA and a career 4.80 FIP. These are not spectacular numbers, but they aren't terrible. Going into this season, Owings had accumulated 2.6 WAR in 250 innings, according to Fangraphs. The average baseball player (hitter or pitcher) will accumulate 2.25 WAR in a full season (about 120 innings for a starting pitcher).

Again courtesy of Fangraphs, his projected FIP-ERA going forward is 4.59; that works out to almost five runs per game. In a league where the average is 4.65, and in a hitter's park like the Great American Ballpark, that works out to a projected 1.33 WAR in 120 innings for Owings. (Probably a little better than that, since he would still accumulate some value from his hitting even as a pitcher.) So a below average pitcher, to be sure, but it's possible that the Reds don't have a better starting pitcher than Owings available.

Owings seems to be comfortably nested in the grey area where he's neither a great asset or a great liability as a pitcher. But what about as a hitter?

Owings the hitter

The short answer here is that we don't know how well he'd perform as a hitter. For the sake of intellectual curiosity, we can go ahead and step through some calculations that seem reasonable and should at least put us in the ballpark, but I want to start off right here with the caveat that the error bars on what we're doing here are huge.

Or to be more plain, this is basically a wild guess. An educated wild guess, maybe—is there such a thing? Anyway, if there is an educated wild guess, this is one.

And let's start off with the first of the two big, unsubstantiated assumptions that are required to make this guesstimate work out. Projection systems are largely based upon the premise of regression to the mean, which is that over time, extreme observations become less extreme. Given a small number of observations, we would assume that a guy who has hit poorly will hit better than he has so far, and a guy who has hit well will hit worse than he has so far. The more observations we have, the more we can have confidence in a player being above or below average as a matter of true talent, rather than it simply being due to random chance.

The issue with regression to the mean here is determining the mean. For the majority of pitchers, regressing their hitting performance to the mean results in simply laughable numbers, because pitcher hitting is really a separate thing from position player hitting, and that almost every pitcher is a well below average hitter. So simply feeding a pitcher's batting line into the typical projection system will not provide usable results.

What we are going to assume here is that, if Owings were to devote himself to being a hitter full-time, that he would regress to the league mean of position players hitting, not the mean of pitchers hitting. This has the benefit of making sense, at least. But it's a really unvalidated assumption, with few data points to go off of (outside of Babe Ruth and Rick Ankiel there's not a lot of evidence to go around, and those two are both special cases indeed). If I had a giant robot following me around helping me perform sabermetrics, he would be flashing warning lights and screaming "DANGER COLIN WYERS, DANGER!" But it's the best guess I have for the time being.

Using Sal Baxamusa's Marcels spreadsheet and Owings' career hitting stats thusfar, we get a forecasted batting line of .285/.343/.478, or about a .350 wOBA. That works out to roughly six runs (or a half-win) above average with the bat in 650 plate appearances.

Now, as it just so happens, the positional adjustment for a corner outfield spot in 650 plate appearances is right around seven runs. So if we're correct, Micah Owings could be about a league average player in the outfield, assuming he plays defense about as well as a typical corner outfielder. Can he do that?

Honestly, this is even more of a guess than his hitting projection. He's a young guy and apparently athletic (and he obviously has the arm to play somewhere like right field). So he could be. Or he could not be. I know I'm being vague here, but that's all I can be with the data at hand.

So, using a set of favorable assumptions, it looks like we could make a case for Micah Owings being a league-average player in the outfield. Again, very large error bars surround this entire endeavor and it's possible that he's an utter trainwreck as an outfielder and hits like a utility infielder. We really don't know for sure. But let's assume for a second that this WAG is essentially correct. Should the Reds try to convert him to an outfielder right now?

Transactional costs

There are real costs to trying to convert Owings to a full-time position player, such as figuring out who to displace from the outfield to make room for him and figuring out who to take his innings in the rotation. There's a lot of chaining going on here, and a lot of things to figure for the Reds.

There's also a lot of risk here. It's real easy to run some numbers and come up with "about 2-2.5 WAR," and another thing entirely to teach a pitcher to play the outfield, to have him abandon his craft and devote himself to something else entirely. There's the chance that in taking Micah the pitcher and making him Micah the hitter you're left with neither Micah at all.

And the Reds just moved into a three-way tie for first place in the NL Central this evening, which says they're probably not in a mood to take those kinds of risks right now. So Owings will probably stay in the rotation, and probably should stay in the rotation.

But boy, it's fun to dream about, isn't it.