Friday, June 22, 2012

Buck Showalter: Tactical Genius?

Jim Johnson's Dominance

Let's start this conversation by saying that this is going to be one of those stat-heavy posts.  In this case we'll be utilizing the win probability data from fangraphs, which you can find here.  The first point I'd like to note is that according to Win Probability Added, Jim Johnson is the best pitcher in baseball this season with a fairly sizable lead over Justin Verlander.  Here's an explanation of WPA if you're not familiar with the stat.  Not necessarily that important, just something interesting to think about as we look at the bullpen usage more in depth.

Buck Showalter's Bullpen Usage

The stat I really want to focus on is gmLI which basically gives an indication of how intense the situation is when the pitcher enters the game.  Using the same page from before we can sort the pitchers who have thrown at least 10 innings in MLB by gmLI.  If you want more info on Leverage Indices and how they work, click on the link in the first sentence of this section.

As you can see, Jim Johnson is tied for 9th in MLB with an average leverage index of 1.79 for the games he enters.  As stated in the fangraphs explanation, anything above 1 is an average leverage appearance, so that figure shows that Buck is using his best pitcher in the most critical situations.  Johnson's 1.79 leverage index leads the Orioles, but looking at the rest of the staff gives more insight into how Buck has used the bullpen.

If we sort out the Orioles' pitchers we can look at the average leverage index when each guy enters the game, and get an idea of who pitches in the highest leverage and lowest leverage appearances.  Johnson paces the O's with a gmLI of 1.79, followed by Strop at 1.71.  These two guys have pitched in the most crucial situations, and their performances have backed up Buck's aggressive usage.

The O's also have a bunch of pitchers between average leverage at 1.00 and 1.16, slightly above average.  From most to least these guys are: Ayala, Lindstrom, O'Day and Patton.  All of these guys have provided quality innings out of the bullpen, though these mainly come earlier in games which might play a role in their lower gmLIs.  At the opposite end of the spectrum Kevin Gregg and Dana Eveland bring up the rear with 0.81 and 0.60 gmLIs.  These are almost certainly a function of their usage, as Gregg rarely pitches in close games and Eveland is the de facto long-man coming in when a starter exits early.

What Does It Mean?

Well to assess the seasons these guys have had, we'll use a stat called ERA-.  Basically ERA- adjusts a pitchers ERA for their league, home park, etc.  For this particular stat a score of 100 would be league average, and the lower your number is the better.  Here are the ERA- stats (among other stats) for the O's pitchers:

Johnson, Lindstrom and Strop have absurd ERA- stats of 29, 31 and 34 respectively, showing that they've been absurdly better than league average to this point.  O'Day and Ayala both also have scores below 50.  This presents an opportunity for Buck to perhaps spread out some of the high leverage situations among these guys as well as they've shown the capability to handle it.  Will they be this good for the whole season?  Probably not.  What we have, and can hope to continue to see is that Buck using his pitchers well. I'm thinking of Buck putting Jim Johnson out there in tie games against the Phils and Yankees earlier this season rather than saving him for the "traditional save chances".

Buck has used the bullpen really well so far this season despite them having to throw a lot of innings.  That said, he needs to continue to manage the bullpen well to put guys in positions to succeed.  Using Johnson as he has allows him to utilize his best pitchers in the most critical situations - giving the O's the best chance to win.  Buck's decision on the lineup and who starts where might be questionable, but his usage of the pitching staff has been really solid so far.


  1. It's interesting that the starters all have a gmLI between .87 and .88. I'd be interested to see you do some research on what the average gmLI is for all starters compared to the O's rotation.

    I'd do it myself but I don't want to rip your work off and be a jerk like that.

    1. Good question Lance. Here is the gmLI for starters who have qualified by throwing the minimum number of innings this season:

      As you can see, the gmLI ranges from 1.04 on the high end to .82 on the low end. Of the 110 pitchers, all but 12 of them have gmLI scores between .86 and .89. I would imagine this is a function of starting games as gmLI is a score for when a pitcher enters the game.

    2. Interesting. I wonder if that's a product of our starters rarely making it beyond the 5th and 6th innings (outside of Hammel and Chen), or Buck inserting the bullpen when certain SP's get themselves into dangerous situations?

    3. It's possible. Remember that gmLI tells you the leverage index for when a pitcher enters the game. The stat: exLI would be more telling for our starters as that tells you the average leverage index when they leave the game (e.g. reliever takes over).

    4. O's rotation has a .87 gmLI, as does 20 other teams, but O's rotation has a 1.09 exLI which is 12th highest in MLB.

      O's bullpen has a gmLI of 1.21, 8th highest in MLB, and has a 1.27 exLI which is also 8th highest.

      Interesting numbers there. Bullpen seems to be fairly "clutch"...which go figure, they have the highest clutch rating on fangraphs at 3.62. Pirates are 2nd at 3.42.

      Next two teams after them are the Rangers and Twins at 2.17. I'd say that's a large reason why the O's have won so many close games.