MLB: Converting a starter to a reliever

[Justin7]

Assume you have good projections on how a player will do in a starting role. How will he do as a reliever? Is "10% better" a good descriptor? Are there any good articles on this problem?

As always, points to very good answers.

[MonkeyF0cker]

Is "10% better" a good descriptor?

Wow.

[mathdotcom]

Monkey is a perfectionist and does not like these blanket numbers since in theory this is a difficult problem. One reason is that it's endogenous why the pitcher has moved from a starting to relieving role.

In practice, is using such a rule of thumb really that horrendous if you don't expect the reliever to face many batters?

I am curious why you suggest 10% better rather than... 10% worse... is that an empirical fact?

[Insoluble]

Using ZiPS projections to estimate starter/reliever conversion:

http://www.baseballthinkfactory.org/oracle/discussion/start_relief_projection_toy_for_microsoft_excel/

Also, an article regarding the "break even point in converting a starter to reliever:

https://www.baseballprospectus.com/article.php?articleid=16435

If you don't have a subscription you can get a sneak peek at the numbers here:

http://www.insidethebook.com/ee/index.php/site/comments/breakeven_point_of_starter_relief_conversion/

[Justin7]

Monkey is a perfectionist and does not like these blanket numbers since in theory this is a difficult problem. One reason is that it's endogenous why the pitcher has moved from a starting to relieving role.

In practice, is using such a rule of thumb really that horrendous if you don't expect the reliever to face many batters?

I am curious why you suggest 10% better rather than... 10% worse... is that an empirical fact?

It is easier to pitch 1-2 innings than 5+. If a reliever starts, his whip gets worse. If a starter moves to relief, his whip improves.

Late inning scoring/inning is roughly 10% less than early inning scoring. Ergo, relief pitchers allow 10% less scoring than starters (this obviously ignores a gazillion situational factors, but is a starting point.) But are relievers worse, or comparable to starters? Or even better? I had hoped there would be a nice academic article reviewing this.

[mathdotcom]

It is easier to pitch 1-2 innings than 5+. If a reliever starts, his whip gets worse. If a starter moves to relief, his whip improves.

Yes but usually it's the starters who have recently performed poorly that get moved to the pen so if you add 10% to their year-long record (which is relatively decent compared to their last few appearances) you might be overestimating their ability.

[MonkeyF0cker]

Yes but usually it's the starters who have recently performed poorly that get moved to the pen so if you add 10% to their year-long record (which is relatively decent compared to their last few appearances) you might be overestimating their ability.

Or underestimating. Would he even know?

He might as well throw darts.

How much could a reduced WHIP be attributed to situational appearances where a double play ends the inning (adding 2/3 of an inning to his WHIP stats without the runners counting toward his WHIP)?

Meh. Why am I trying to help him?

[evo34]

http://www.bloguinpro.com/statspeak/2010-articles/april/to-start-or-relieve.html

[flocko76]

some starters did well, Brett myers, John smoltz, kerry wood, although you probably only hear about the successes. Even ryan madson was a starter converted to relief. I would factor in which team is doing the converting. Good teams like the Phillies seem to have more success than bad teams.

[antifoil]

maybe you could look to see how these players perform the first time through the line up. Some of the starters that can't make it have problems turning over a line up 2 or 3 times for various reasons.

there could a better correlation between a relieve outing and the first 3 innings of an outing. Of course, you still run into the fact that relievers tend to throw harder because they don't need to pace themselves and cut back on their repertoire of pitches.

i also disagree with monkey's assertion that converted starter turn into one out situational guys or even used highly in those situations. LOOGY's and the like are platoon guys.

[MonkeyF0cker]

maybe you could look to see how these players perform the first time through the line up. Some of the starters that can't make it have problems turning over a line up 2 or 3 times for various reasons.

there could a better correlation between a relieve outing and the first 3 innings of an outing. Of course, you still run into the fact that relievers tend to throw harder because they don't need to pace themselves and cut back on their repertoire of pitches.

i also disagree with monkey's assertion that converted starter turn into one out situational guys or even used highly in those situations. LOOGY's and the like are platoon guys.

When did I say that? My example was purely ONE hypothetical which could lead to reduced WHIP. I'm not sure how you get a blanket statement out of it.

However, relievers (outside of closers) are largely used situationally. How many managers typically throw a left-handed reliever following a left-handed starter? Not many. Why? Because the lineup of the opposing team is generally platooned against the starting pitcher.

[donkson]

There's a chapter in 'the Book' that talks about changing between starter and reliever, it also mentions equivalent changes in era for each.