Joe Maddon is Using J.P. Howell Correctly

by R.J. Anderson on Jun 16, 2009 10:00 AM EDT

I've beaten the tar out of J.P. Howell being used in the highest leveraged situations and Joe Maddon is really doing a brilliant job of doing just that since Troy Percival finally broke. Howell has appeared in 15 games the last 30 days, which dates back to when Percival was still pitching, and his average leverage index is 2.14; Percival's in those two games was 2.52. Howell's average pLI with Percival around was 1.21, and without is 2.18. Let's take a look at when Maddon is using Howell:

@ FLA entered bottom 7^th, ---, 0 out, ahead two, 2.91 aLI

@ FLA entered bottom 9^th, ---, 0 out, tied, 1.89 aLI

@ CLE entered bottom 7^th, ---, 0 out, down one, 0.53 aLI

MIN  entered top 8^th, --3, 0 out, ahead three, 2.35 aLI

MIN entered top 9^th, 1-3, 1 out, down one, 1.71 aLI

KCR entered top 9, ---, 0 out, ahead four, 0.21 aLI

KCR entered top 9^th, ---, 0 out, ahead one, 2.22 aLI

@ NYY entered bottom 8^th, 1--, 1 out, ahead one, 4.62 aLI

@ NYY entered bottom 8^th, 123, 1 out, ahead two, 3.88 aLI

LAA top 9^th, ---, 0 out, down one, 0.66 aLI

WSN entered top 8^th, --3, 2 outs, tied, 2.74 aLI

WSN entered top 9^th, ---, 0 out, ahead one,2.38 aLI

That's 15 appearances, two save chances, a 1.73 FIP, and the highest leverage on the team during that time. Maddon's not using Howell as a closer at all, he's using him as a relief ace, a fireman, the iceman, whatever you want to label him, Maddon is using him perfectly. Seriously. He could've easily said he was giving the set closers job to Jason Isringhausen or Dan Wheeler and most people would've bought into it, instead Maddon has done something far, far savvier and the Rays are a better team because of it.

Fantastic job, Joe.

Tweet Comment 43 comments | 0 recs |

Do you like this story?

Comments

Better bullpen management = better bullpen results

Who woulda thought?

by matthan on Jun 16, 2009 10:09 AM EDT reply actions

Better pitching also equals better results

by tallyray on Jun 16, 2009 10:56 AM EDT up reply actions

You mean better pitchers means better results

Don’t forget the fat tub of lard is gone….

by matthan on Jun 16, 2009 1:27 PM EDT up reply actions

Bottom of 7 no on no out levergae

Down 1 i is a .53
while up two in the bottom of the 7th is 2.91

The cost of giving up an additional run is greater down 1 than up 2 no?

Follow Me on Twitter @FreeZorilla

by FreeZorilla on Jun 16, 2009 10:16 AM EDT reply actions

at least much closer

Follow Me on Twitter @FreeZorilla

by FreeZorilla on Jun 16, 2009 10:16 AM EDT up reply actions

It's averaged from his entire inning I believe.

by R.J. Anderson on Jun 16, 2009 11:27 AM EDT up reply actions

Still, there are some unusual findings on the actual leverage table

For example:
Bottom of 7 runner on 3rd no outs

The leverage index for a 1 run deficit is 2.0, while the leverage for a 2 run deficit is 2.2. I’m in the middle of reading the Book now so I’m sure I’ll understand better in a week or so.

Follow Me on Twitter @FreeZorilla

by FreeZorilla on Jun 16, 2009 11:52 AM EDT up reply actions

Howell needs help.

by walkoffwalk on Jun 16, 2009 10:23 AM EDT reply actions

He has help. The bullpen is en fuego.

Follow Me on Twitter @FreeZorilla

by FreeZorilla on Jun 16, 2009 10:31 AM EDT up reply actions

Average Leverage Index

I’m new to all of this and was hoping someone might provide some insight into how the aLI is calculated. I find the NY set interesting because they are both similar situations; tying run on first or tying run on second with the go-ahead run on first.

Just trying to learn—thanks for any help.

by timmyj on Jun 16, 2009 10:38 AM EDT reply actions

I second this request for informational assistance.

B Rad the Ray Fan

by BWoodrum on Jun 16, 2009 10:41 AM EDT up reply actions

Here you go.

http://www.fangraphs.com/blogs/index.php/get-to-know-leverage-index/

by R.J. Anderson on Jun 16, 2009 11:26 AM EDT up reply actions

Between the Numbers is pretty good.

by R.J. Anderson on Jun 16, 2009 11:38 AM EDT up reply actions

When comparing figures like tRA or FIP

What is the minimum amount of innings in one season to make that figure reliable? I’m trying to some analysis and obviously I want to toss out pitchers that barely threw any innings, but I’m not sure what the proper benchmark should be.

by matthan on Jun 16, 2009 10:38 AM EDT reply actions

Better yet maybe a # of pitches?

Although IP would do

by matthan on Jun 16, 2009 11:04 AM EDT up reply actions

At this point, I'd say more than 20 innings.

by R.J. Anderson on Jun 16, 2009 11:31 AM EDT up reply actions

Thanks. I'm using 30 IP.

I wish I could use more but that would kill my sample in a couple different ways. I’m trying to determine Hickeys influence.

by matthan on Jun 16, 2009 1:29 PM EDT up reply actions

Are you just doing starters or relievers?

If it’s the former, 30-40 makes sense, the latter, I’d still try and stick with 20-30.

by R.J. Anderson on Jun 16, 2009 1:45 PM EDT up reply actions

Both. I’m going to look at it separates but probably end up combining all pitchers. I’m not focused on the mean, but rather year over year changes for each pitcher. I want to find if the change is significant. So it doesn’t really matter if the pitcher was a starter or a reliever. Although switching from a starter to a reliever (or vice versa) would be a bad thing for this analysis since relieving is easier than starting and that variable would explain the increase or the decrease. The first way I am looking at it is whether there is a difference between a year with Hickey and a year without Hickey immediately before or after it. Then I’m going to also take a look at whether pitchers improve or get worse while under at least two years of Hickey. I’m hoping the sample is large enough to eliminate the impact of say age (e.g. Garza’s increase v Wheelers decline). Of course with this there is also survivorship bias which is partially why a high IP is needed.

by matthan on Jun 16, 2009 2:12 PM EDT up reply actions

I'm also using his time with the Astros

by matthan on Jun 16, 2009 2:15 PM EDT up reply actions

In other words

I’m looking at say Roger Clemens w/o Hickey in 2004 and comparing it to w/ Hickey in 2005 and also Roger Clemens w/o Hickey in 2007 and comparing it to w/ Hickey in 2006. And say Brian Stokes in 2008 w/o Hickey and 2007 w/ Hickey. Even though one is a starter and one is a reliever it shouldn’t matter since the expectation is that pitching coaches do no matter and that their metrics should stay the same.

by matthan on Jun 16, 2009 2:19 PM EDT up reply actions

I'm not sure you'll find much.

But if you do go through, I don’t know if Roger Clemens is the best measuring stick for a pitching coach.

by R.J. Anderson on Jun 16, 2009 2:21 PM EDT up reply actions

Jim Hickey broke Roger Clemens!

by rglass44 on Jun 16, 2009 2:26 PM EDT up reply actions

Actually...

Clemens had better years in 2005 and 2006 with Hickey than he did w/o him in 2004 and 2007.

by matthan on Jun 16, 2009 2:28 PM EDT up reply actions

Of course that could be attributed to a different variable rather than Hickey, but that is why the same is far larger than 2.

And yes I know you were being sarcastic. It is obvious that his roid raging wife is the one that broke him.

by matthan on Jun 16, 2009 2:32 PM EDT up reply actions

sample

God I’m the worst typer when I’m at work

by matthan on Jun 16, 2009 2:33 PM EDT up reply actions

Wayne Graham was the only person to sucessfully break Roger Clemens!

Tools Whore

by Tyler on Jun 16, 2009 4:58 PM EDT up reply actions

You can’t pick and choose the data though. Something like this requires very diverse data. The larger the set the more able I’ll be able to dismiss things like age, talent, injuries, roids, etc.

If we just wanted to look at their impact on young pitchers it would require a sample size of young guns that would be nearly impossible to get.

by matthan on Jun 16, 2009 2:27 PM EDT up reply actions

But I do agree with you that I don't expect to find much

I really don’t expect there to be a significant difference between the performance of a year with Hickey and a year without it for the same group of pitchers.But there is really only one way to find out.

I’ll finish it up between Neimann pitches.

by matthan on Jun 16, 2009 2:39 PM EDT up reply actions

You're talking to yourself.

by Suttree on Jun 16, 2009 2:47 PM EDT up reply actions

I do that all the time

It is nice to talk to someone with equal intelligence

by matthan on Jun 16, 2009 3:40 PM EDT up reply actions

You sir

are a formidable opponent

Rays Win!

by Sandy Kazmir on Jun 16, 2009 3:55 PM EDT up reply actions 1 recs

I just finished up

There is a statistically significant difference between a year (30IP or more) under Hickey and a year (30IP or more) under a different pitching coach in back to back years (Hickey/NonHick, NonHick/Hick). The FIPs of the pitchers go down when they are under another pitching coach. There is no statistically significant relationship for K/9, BB/9, K/9, or ERA (what a horrible metric). I’m going to throw in tRA when I have time. I also want to see how pitchers do over multiple years with Hickey. I’m not sure if I’ll make a fanpost now or wait until I get everything done.

by matthan on Jun 16, 2009 4:06 PM EDT up reply actions