Roch: Hitting Coach Terry Crowley is coming back

[KAZ97]

Just to make sure my method wasn't systematically biased, I ran the same analysis for Eddie Murray's tenures as hitting coach.

Eddie was pitching coach for Cleveland from 2002 to June 4th, 2005 and then for the Dodgers from 2006 to Jun 14th, 2007. Using the same methods, this time using data from 2002-2007, here is Eddie compared to other hitting coaches:

Under Eddie Murray: .264/.335/.422 in 29,088 PAs

Under other coaches: .270/.340/.426 in 105,131 PAs

It seems this provides evidence that Eddie is about a league average hitting coach and that having better players can make your overall tenure numbers look pretty good.

[Ohfan67]

Just to make sure my method wasn't systematically biased, I ran the same analysis for Eddie Murray's tenures as hitting coach.

Eddie was pitching coach for Cleveland from 2002 to June 4th, 2005 and then for the Dodgers from 2006 to Jun 14th, 2007. Using the same methods, this time using data from 2002-2007, here is Eddie compared to other hitting coaches:

Under Eddie Murray: .264/.335/.422 in 29,088 PAs

Under other coaches: .270/.340/.426 in 105,131 PAs

It seems this provides evidence that Eddie is about a league average hitting coach and that having better players can make your overall tenure numbers look pretty good.

I think these analyses are an impressive amount of work and I am super impressed that you and others have done them. However, I think in this case having more data is not really helpful. When you average 105,000 at bats you have tremendous statistical power to compare means, but in this case I think it makes the comparison pretty meaningless. You are looking at hundreds of players across a random portion of their playing career. Those players include some great hitters, a lot of mediocre hitters, and a good number of really bad hitters. Just think about how weak the 7, 8, and 9 O's hitters have been over the last ten years. You don't know when players were hurt, where they were in their development (a rookie taking instruction may be very different than a successful veteran taking instruction), etc. When you pull 100,000 at bats from the career of a "long-lived" coach like Crowley as you did in this case I bet that those 100,000 at bats would be really similar to 100,000 randomly collected at bats from across all major leaguers of that era. To really test that idea you would have to do a bootstrap type simulation where you randomly pull a 100,000 at bats from the data set 10,000 times and then compare each of those to Crowley's numbers, but I think if you did that there wouldn't be much difference. There's just so much noise you basically just get average MLB performance. Any real differences that would occur are likely to be far more influenced by other factors than the hitting coach, even though you are doing a very admiral job of trying to isolate that effect. For example, when the hitting coach changed the ballpark changed.

[KAZ97]

I think these analyses are an impressive amount of work and I am super impressed that you and others have done them. However, I think in this case having more data is not really helpful. When you average 105,000 at bats you have tremendous statistical power to compare means, but in this case I think it makes the comparison pretty meaningless. You are looking at hundreds of players across a random portion of their playing career. Those players include some great hitters, a lot of mediocre hitters, and a good number of really bad hitters. Just think about how weak the 7, 8, and 9 O's hitters have been over the last ten years. You don't know when players were hurt, where they were in their development (a rookie taking instruction may be very different than a successful veteran taking instruction), etc. When you pull 100,000 at bats from the career of a "long-lived" coach like Crowley as you did in this case I bet that those 100,000 at bats would be really similar to 100,000 randomly collected at bats from across all major leaguers of that era. To really test that idea you would have to do a bootstrap type simulation where you randomly pull a 100,000 at bats from the data set 10,000 times and then compare each of those to Crowley's numbers, but I think if you did that there wouldn't be much difference. There's just so much noise you basically just get average MLB performance. Any real differences that would occur are likely to be far more influenced by other factors than the hitting coach, even though you are doing a very admiral job of trying to isolate that effect. For example, when the hitting coach changed the ballpark changed.

That was exactly my expectation. That's why I ran it for Eddie Murray, because the results for Crowley were so unexpected that I wanted to run it for someone else as a bit of a sanity check.

As you describe, I would have expected that any comparison would come out as about league average. The fact that the evidence is consistent with Terry Crowley being significantly better than his contemporaries was quite surprising.

[Ohfan67]

That was exactly my expectation. That's why I ran it for Eddie Murray, because the results for Crowley were so unexpected that I wanted to run it for someone else as a bit of a sanity check.

As you describe, I would have expected that any comparison would come out as about league average. The fact that the evidence is consistent with Terry Crowley being significantly better than his contemporaries was quite surprising.

But I'm not sure that one comparison really means that Crowley was better than Murray, or at least better than "his contemporaries". My main reason to question this assertion is statistical...one of these comparisons isn't very meaningful when you're talking about a factor (hitting coach) that has such little effect relative to other things. That's why I suggested the bootstrap method is really the only way to statistically test for the Crowley effect. Just comparing snapshots of Crowley versus others using your methods doesn't really address the issue very well because of all the other factors involved. Also, Eddie Murray was apparently a pretty bad hitting coach. He was fired after a relatively short time and has not been rehired. Crowley's long tenure alone would suggest that he is not really bad. So is that a fair comparison for Crowley? Know what I mean?

I am really impressed with what you did though. I'm just thinking out loud and not trying to criticize. I use a lot of statistics at work and enjoy thinking about statistical problems like this one.

[Ohfan67]

p.s. I think this analogy is correct: comparing Crowley and Murray would be like comparing only two data points in a data set and then making conclusions.

[Frobby]

But I'm not sure that one comparison really means that Crowley was better than Murray, or at least better than "his contemporaries". My main reason to question this assertion is statistical...one of these comparisons isn't very meaningful when you're talking about a factor (hitting coach) that has such little effect relative to other things. Also, Eddie Murray was apparently a pretty bad hitting coach. He was fired after a relatively short time and has not been rehired. Crowley's long tenure alone would suggest that he is not really bad. Know what I mean?

To me, I don't look at KAZ's results as "proving" that Crowley is an above average hitting coach, but I do find that they negate any supposed "proof" that Crowley is a bad coach because statistically the Orioles have been a below average offensive team.

I also think we kind of skipped over the part of his data that showed that the players he coached who also played for other teams during the test period (1)had slightly lower walk rates under their other hitting coaches, and (2) were, generally, not very good players. To me that data really belies the notion that Crowley is teaching them to be undisciplined.

[Frobby]

It's not using historical PECTOA comps, but that doesn't mean we don't have any data regarding Terry Crowley's performance. Here is some empirical evidence of the job Crowley has done. I posted this earlier in the thread, then realized I left the pitchers plate appearances in. I took them out, and here is a comparison of Terry Crowley vs other hitting coaches.

Again, this is using data from the 2006-2009 seasons. I took all the hitters who played under Terry Crowley who also played under another hitting coach during that same time frame. Some played on other teams before their time with Terry Crowley, some after and some both.

Under Terry Crowley: .266/.326/.415 in 14,770 PAs

Under other coaches: .257/.315/.389 in 14,825 PAs

Another thing I find interesting: if my math is correct, this means that the players who played only for the Orioles in the 2006-09 period hit approximately .280/.346/.423 in about 10,000 PA. They were considerably more disciplined hitters than the ones who came and went. Not too surprising since Markakis, Roberts, Mora, Wieters and Reimold all fit in that category and they probably represent the overwhelming majority of the 10,000 PA.

[BaltimoreTerp]

http://weblogs.baltimoresun.com/sports/orioles/blog/2010/10/sources_crowley_and_orioles_ag.html

Although no official announcement has been made, the Orioles have reached an agreement with Terry Crowley, which will make the big-league hitting coach an organizational-wide offensive evaluator in 2011, according to two sources.

...

During next season, Crowley is expected to have a wide range of responsibilities, which would include tutoring specific minor league hitters, evaluating potential amateur draft choices and scouting hitters from other organizations that could be available via trade or free agency.

[Capn Vivi]

This part is particularly interesting:

Crowley, 63, had the option to remain with the big-league club in 2011, according to a source. But instead of signing on for his 13th consecutive season and 17th overall as Orioles’ hitting coach, he has chosen this newly created role, which will have him working with the organization’s minor leaguers and major leaguers in spring training.

So he essentially stepped down as hitting coach? He chose not to remain on the coaching staff?

[Crazysilver03]

This part is particularly interesting:

So he essentially stepped down as hitting coach? He chose not to remain on the coaching staff?

I would think the only reason he stepped down was becaus he was offered this organizational role and that kind of put the writing on the wall. Why else create a new role and offer it to him?

On the topic of the role, it sounds great and hopefully will help the minor leaguers.

[Frobby]

This part is particularly interesting:

So he essentially stepped down as hitting coach? He chose not to remain on the coaching staff?

Sounds like it. If I were 63 years old, a job like the one described might well be more attractive than being a hitting coach.

It is pretty ironic that, with all the screaming around here about how Crowley must go, it appears Showalter would have been happy to stay but Crowley chose to step aside. I say "it appears" because you never know if some of this is spin designed to prevent Crowley from being embarassed.

[Capn Vivi]

Roch just added this new detail in a recent blog post:

This is an interesting reversal from Crowley, who had agreed to return as hitting coach next season. I'm told he was given the option by manager Buck Showalter and accepted it, but three weeks later, he's apparently changed his mind and his title.

Definitely sounds like Crowley liked the sound of the new title and responsibilities. I think it can only be good for the organization, honestly. If there's one thing that Crowley is respected for around baseball it's his ability to evaluate and adjust the mechanics for hitters. Markakis and Pie absolutely sing his praises and his popularity amongst the players is really beyond disagreement.

[TonySoprano]

Sounds like it. If I were 63 years old, a job like the one described might well be more attractive than being a hitting coach.

It is pretty ironic that, with all the screaming around here about how Crowley must go, it appears Showalter would have been happy to stay but Crowley chose to step aside. I say "it appears" because you never know if some of this is spin designed to prevent Crowley from being embarassed.

I wouldn't count out the spin angle. Call it the Peter[Angelos?] principle in action; kick the guy upstairs to save face. It is possible that, at 63, Crow has grown tired of the traveling that is involved.

[crowmst3k!]

this is great news. Crow is a terrific swing mechanics guy. He will be an asset to the players in the minors.

[Moose Milligan]

Maybe Crow reads the OH and found out he's not as good of a hitting coach as he thought he was.