Manipulating Luck: Brewers Starter Support | Disciples of Uecker

Disciples of Uecker

We'd like to go to the Playoffs, that would be cool.

Before Franchise Pitcher Yovani Gallardo struggled toward the end of the season, there was a good case to be made that the righty was the Brewers’ Most Valuable Player. Entering August, the Brewers were 10-12 in his starts, despite the fact that Gallardo was consistently hovering around eight runs prevented for the season (for context, that’s quite good: in the 2014 NL, an eight runs prevented performance ranked between #18-#20 among regular NL SP). To that point, Brewers bats only scored 72 runs in Gallardo’s 22 starts, far below their season average, and the bullpen allowed 25 runs during Gallardo’s outings. To reach a .500 record, Gallardo only had 47 runs to work with; remarkably, Gallardo’s excellent performance nearly matched that ridiculous standard.

Of course, Gallardo’s last ten starts were rough: the bullpen allowed 19 runs, the bats dried up to 26 runs scored, and Gallardo himself allowed 32 runs. This August and September performance absolutely reflects the total nature of the collapse, and it also reflects the generally frustrating nature of 2014: even when Gallardo was pitching poorly down the stretch, the Brewers still lost three quality starts, including two starts in which Gallardo did not allow a single run. It is no wonder that the Brewers went 12-20 during Gallardo’s 2014 starts, and those quality start losses are painful reminders of exactly how the Brewers lost their contending status (it’s one thing to simply get whipped day in and day out, but the Brewers left a lot of winnable games on the field).

One of the toughest elements of statistical analysis is manipulating and understanding completely contextual data, like run support, bullpen support, and fielding support. Statistical analysis provides several excellent tools for abstracting players’ performances and judging expected outcomes for specific types of performances. This is not a dismissal of those tools, for it is important to have a vague understanding of, “if the Brewers employ player X, how many more games will they win?” Yet, the trouble with the pennant race is that it is nothing more than sheer luck: it is the intersection of hot streaks, injuries, inexplicable support, etc. There are plenty of examples of poorly supported players that should have performed much better; my favorite is 2011 Zack Greinke, who could not receive solid support from one of the strongest teams in franchise history, let alone 2004 Ben Sheets (who received little support from an awful team). The 2014 Brewers were squarely between great and awful, and they were not able to complete their contending run despite a handful of solidly average starting pitchers: between Gallardo, Kyle Lohse, Wily Peralta, Matt Garza, and Fastballer Mike Fiers, the Brewers boasted 824.3 IP and 21 runs prevented (thanks mostly to Fiers); the Brewers only managed 69 wins against 63 losses during these pitchers’ starts.

Thankfully, terribly poor support was basically limited to Gallardo, as Peralta and Lohse enjoyed good support, and Fiers and Garza still had solid support. Marco Estrada and Jimmy Nelson received support that was arguably below average, due mostly to bullpen runs allowed during their outings. Here are how the Brewers bats and relievers supported the Brewers starters (ranked by quality start losses):

Brewers Support GS RS Bullpen RA RA to .500 Actual RA Run Differential (W-L) QS L
Gallardo 32 98 44 54 86 98 RS / 130 RA (12-20) 8
Peralta 32 129 34 95 88 129 RS / 122 RA (19-13) 7
Garza 27 111 34 77 77 111 RS / 111 RA (13-14) 7
Lohse 31 149 34 115 87 149 RS / 121 RA (19-12) 4
Estrada 18 91 30 61 60 91 RS / 90 RA (9-9) 3
Nelson 12 37 18 19 40 37 RS / 59 RA (4-8) 3
Fiers 10 34 7 27 17 34 RS / 24 RA (6-4) 2

What is striking about the bullpen support is that the Brewers, in general, had a solid bullpen in 2014. In fact, their relievers were approximately 10 runs better than the NL / Miller Park average last season. So, this support chart simply goes to show that even when a team is better than average, its performance is not distributed evenly:

Brewers Support Bullpen RA / GS
Estrada 1.67
Nelson 1.50
Gallardo 1.38

Garza 1.26
Lohse 1.10
Peralta 1.06
Fiers 0.70

Lohse and Peralta were the true leaders of the rotation, as both pitchers worked the most innings and prevented the most runs among regular starters. Interestingly enough, they also received the best support, which is somewhat unfortunate: both pitchers could have excelled with slightly worse support, leaving one to wonder whether teams simply “feel” good when some pitchers pitch. Everything rolled when Lohse pitched: the bats knocked the ball around the park, the relievers were great, and Lohse was pretty good himself. In fact, even Lohse’s defensive support was great, while Gallardo’s (and Peralta’s, and Nelson’s) was not particularly good; once again, even though the Brewers had an above average defense, its efforts were not evenly distributed:

Brewers Support BABIP LD%
Nelson .345 17%
Peralta .300 16%
Gallardo .298 16%
TEAM .293 17%
Garza .273 17%
Lohse .272 20%
Estrada .259 20%
Fiers .227 15%

What else could have Gallardo done in 2014? His fastball improved, which helped influence the type of contact he allowed, and it just did not matter: the fielders, batters, and relievers simply did not support him. Just in case you’re skeptical about the idea of quality start losses, noting that even a pitcher that allows three runs in 6 IP should not necessarily be expected to win, Gallardo was exceptional in his quality start losses (as were his teammates):

QS Losses GS IP R Runs Average
Gallardo 8 52.7 8 1.37
Garza 7 50.3 13 2.33
Peralta 7 45.3 14 2.78
Lohse 4 26.7 8 2.70
Nelson 3 19.0 7 3.32
Estrada 3 18.0 8 4.00
Fiers 2 14.0 5 3.21
Total 34 226 63 2.51

This might be the most depressing chart I’ve seen yet, pertaining to the 2014 Brewers: in these 34 losses, the Brewers starters were basically as good as Clayton Kershaw, as they prevented more than 40 runs. So, the Brewers’ quality start losses were not simply “ho hum / poor” quality starts of the 6 IP / 3 R variety: they were the 6.64 IP / 1.85 R variety!!! Given that the bullpen allowed 1.25 R/G on average in 2014, Brewers bats arguably should have been able to win at least a portion of these games. Even if you’re inclined to argue that a team should be expected to lose a certain portion of quality starts, the Brewers arguably ruined at least 12 great outings in 2014, or outings in which the starter allowed 0 or 1 run:

“Great” Losses (0-1 R) # of Losses
Gallardo 5
Peralta 3
Garza 3
Lohse 1

These charts show why it is important for the Brewers to improve their depth for 2015, especially by bolstering 1B production (and, arguably, 2B and SS production, although that is a different article). While the club has veteran performers that also need to improve (hopefully simply by returning to full health), the Brewers truly need to score more runs in 2015. By looking at 2014 support charts for the starting pitchers, it should be clear that there are very real consequences for having a below average offense. Not only did the Brewers lose a lot of very winnable games in 2014, their below average offense was unevenly distributed, which completely wasted a solid performance by Gallardo. One of the keys for improving in 2015 will be correcting performances in a way that maximizes both pitching and batting performances: therefore, it is also worth questioning whether the Brewers will have a more efficient club in 2015 simply by improving the offensive contributors. For, if team support is simply unevenly distributed, one must ask, “which pitcher will receive poor support in 2015?”

Baseball-Reference. Sports Referece, LLC., 2000-2014.
MLB Advanced Media, LP., 2014.
2011-2014 Rotations with Runs Prevented
2014 NL SP 100+ IP rankings

Share Our Posts

Share this post through social bookmarks.

  • Delicious
  • Digg
  • Newsvine
  • RSS
  • StumbleUpon
  • Technorati