What better time to analyze the Brewers' bullpen than now, at the AS Break?
Like last year, the common sense fan tells you that our bullpen is a problem, and not getting the job done. However, the bullpen is not as bad as most people think. Now, the pen is not as good as last year's pen, but on the whole there are some very positive spots in the pen, to overshadow the negative ones.
First, the overall stats (I used everyone that pitched more than 1 game out of the bullpen):
Overall: 116 IR/32 IS (27.6%); 76/91 leads converted (83.5%)
Guillermo Mota (34 IP, 33 G): 8 IR/1 IS; 24 R, 37 H, 17 BB; 10/12 leads converted (83.3%)
Salomon Torres (49.3 IP, 43 G): 21 IR/5 IS; 18 R; 43 H, 20 BB; 20/23 leads converted (86.9%)
Brian Shouse (33 IP, 44 G): 42 IR/16 IS; 9 R; 25 H, 8 BB; 13/16 leads converted (81.3%)
Eric Gagne (23.3 IP, 24 G): 3 IR/1 IS; 19 R; 29 H, 16 BB; 10/15 leads converted (66.7%)
Carlos Villanueva (27 IP, 16 G): 9 IR/0 IS; 8 R, 23 H, 5 BB; 7/7 leads converted (100%)
Mitch Stetter (22 IP, 22 G): 6 IR/0 IS; 6 R; 11 H, 15 BB; 2/2 leads converted ( 100%)
David Riske (29 IP, 27 G): 15 IR/4 IS; 15 R; 28 H, 14 BB; 10/12 leads converted (83.3%)
Seth McClung (20.3 IP, 11 G): 2 IR/1 IS; 8 R; 15 H, 13 BB; 1/1 leads converted (100%)
Tim Dillard (8.3 IP, 8 G): 3 IR/2 IS; 10 R; 10 H, 5 BB; 1/1 leads converted (100%)
Mark DiFelice (13.7 IP, 10 G): 5 IR/2 IS; 7 R; 13 H, 0 BB; 1/1 leads converted (100%)
Derrick Turnbow (6.3 IP, 8 G): 0 IR/0 IS; 11 R; 12 H, 13 BB; 1/1 leads converted (100%)
Zach Jackson (3.7 IP, 2 G); 2 IR/0 IS; 2 R; 5 H, 2 BB; n/a leads converted
On the whole, allowing fewer than 30% of inherited runners to score is pretty solid, and 83.5% leads converted is slightly below average. Gagne's problems account for 5 of the 15 blown leads alone, so without the right-hander, the pen has converted 66/76 leads (86.8%). That clip is pretty much average, perhaps above average (I do not yet have the full NL averages)
In advanced terms, beyond leads converted and IR/IS, we can also analyze how the relievers are being used, and whether or not their usage patterns are warranted given their success rate.
For this aspect of the pen, I utilized an adjusted WHIP, which accounts for IR/IS along with the reliever's own baserunners (H, BB), and weighs those baserunners against the number of baserunners scored. Basically it's a full picture of the frequency with which a reliever allows their total baserunners to score (inherited and their own).
The basic idea is that a reliever will generally be more successful if they allow a lower percentage of runners to score.
Adjusted WHIP: (137 R + 32 IS)/(251 H + 128 BB + 116 IR) = .3414
Mitch Stetter (6 R + 0 IS)/(11 H + 15 BB + 6 IR) = .1875
Carlos Villanueva (8 R + 0 IS)/(23 H + 5 BB + 9 IR) = .2162
Zach Jackson (2 R + 0 IS)/(5 H + 2 BB + 2 IR) = .2222
Salomon Torres (18 R + 5 IS)/(43 H + 20 BB + 21 IR)= .2738
Seth McClung (8 R + 1 IS)/(15 H + 13 BB + 2 IR) = .3000
Brian Shouse (9 R + 16 IS)/(25 H + 8 BB + 42 IR) = .3333
David Riske (15 R + 4 IS)/(28 H + 14 BB + 15 IR) = .3333
Guillermo Mota (24 R + 1 IS)/(37 H + 17 BB + 8 IR) = .4032
Eric Gagne (19 R + 1 IS)/(29 H + 16 BB + 3 IR) = .4167
Derrick Turnbow (11 R + 0 IS)/(12 H + 13 BB + 0 IR) = .4400
Mark DiFelice (7 R + 2 IS)/(13 H + 0 BB + 5 IR) = .5000
Tim Dillard (10 R + 2 IS)/(10 H + 5 BB + 3 IR) = .6667
Of the regulars, Mitch Stetter, Carlos Villanueva, Salomon Torres, and Seth McClung are the best relievers in our bullpen, with Brian Shouse and David Riske not far behind. Each of those pitchers is above the league average for basic WHIP, although basic WHIP does not include IR/IS numbers, so the league average is estimated:
Actual WHIP (R/(H+BB)): .3615
NL League Average WHIP: .3714
The pen's actual WHIP % scored is higher than the overall % because the IR/IS % is fairly low for the entire bullpen. However, the pen as a whole is still allowing runners to score at a lower percentage than average pitchers in the NL.
The real question is, do Yost's usage patterns reflect the success of the relievers, or rather, does Yost use relievers in roles they are best suited for?
In order to get an idea of Yost's usage patterns, I looked at two areas:
(1) Appearance distribution. If we look at the innings in which relievers are entering games, are they entering games in one inning only, or are they used flexibly? My intuition is that a good bullpen will have both flexible relievers and some steadfast relievers, although my own personal opinion leans towards a more flexible pen, so I valued pitcher flexibility in this study. This means that if a reliever has a lower percentage of their overall appearances invested in one single inning in the ballgame, they will be more valuable (because they can enter the ballgame in multiple IP).
(2) "Clutch" distribution. This is simple: how many tie games does a reliever appear in? How many games within one run does a reliever appear in? (This includes one run defecits, one run leads, and ties, overall).
This point is simple: a reliever that can successfully pitch in many close games, that reliever will be more valuable than a reliever that does not pitch in close games. But, this can also signal managerial misuse in other cases -- does a bullpen feature too many relievers that pitch poorly, but appear in lots of close games?
Appearance Distribution:
McClung: 3 ea/11 appearances in the 6th and 7th innings (27.5% each, 55% total); 2/11 tie games, 2/11 games within 1 R
Riske: 8 ea/27 appearances in the 6th, 7th, and 8th innings (30% each, 89% total); 1/27 tie games, 12/27 games within 1 R
DiFelice: 3/10 appearances in 5th inning (30%); 0/10 tie games, 1/10 games within 1 R
Stetter: 7/22 appearances in the 7th inning (32%); 3/22 tie games, 8/22 games withint 1 R
Villanueva: 6/16 appearances in the 6th inning (37.5%); 1/16 tie games; 6/16 games within 1 R
Dillard: 3/8 appearances in 8th inning (37.5%); 2/8 tie games, 3/8 games within 1 R
Turnbow: 3 ea/8 appearances in 8th and 9th innings (37.5% ea, 75%); 1/8 tie games, 3/8 games within 1 R
Shouse: 18/44 appearances in 8th inning (41%); 7/44 tie games, 17/44 games within 1 R
Torres: 18/43 appearances in 9th inning (42%); 8/43 tie games, 22/43 games within 1 R
Gagne: 17/24 appearances in 9th inning (71%); 4/24 tie games, 9/24 games within 1 R
Mota: 26/34 appearances in 8th inning (76.5%); 10/34 tie games, 15/34 games within 1 R
Jackson: 2/2 appearances in 7th inning (100%); 0/2 tie games, 0/2 games within 1 R
Clutch Distribution:
Torres: 8/43 tie games (19%), 22/43 games within 1 R (51%)
Riske: 1/27 tie games (4%), 12/27 games within 1 R (44%)
Mota: 10/34 tie games (29%), 15/34 games within 1 R (44%)
Shouse: 7/44 tie games (16%), 17/44 games within 1 R (39%)
Gagne: 4/24 tie games (17%), 9/24 games within 1 R (37.5%)
Villanueva: 1/16 tie games (6%); 6/16 games within 1 R (37.5%)
Dillard: 2/8 tie games (25%), 3/8 games within 1 R (37.5%)
Turnbow: 1/8 tie games (12.5%), 3/8 games within 1 R (37.5%)
Stetter: 3/22 tie games (14%), 8/22 games withint 1 R (36%)
McClung: 2/11 tie games (18%), 2/11 games within 1 R (18%)
DiFelice: 0/10 tie games (0%), 1/10 games within 1 R (10%)
Jackson: 0%
Some observations:
(1) Yost's usage of Mota and Gagne has hurt the pen to some extent -- both of these pitchers are struggling, but they have been used in fairly inflexible ways. Although both allow a high percentage of runners to score, they pitch in a disproportionate number of close games, and they also are not used flexibly, distributing their IP over different points in the ballgame.
SOLUTION: Bump Mota and Gagne to junktime usage, or leads that are 3 runs or larger. Keep them out of as many tie games as possible, and pitch them in fewer one-run games.
(2) Yost's usage of Riske is perfect. He is matching Linebrink's role last year, not inheriting many 3-run leads (or fewer), but pitching in a good number of tie games and 1-run defecits, as well as some 1-run leads. Yost also enters him into the game at any point, as he has equal appearances entering the game in the 6th, 7th, and 8th.
Note: Keep it up, Yost! And also, hopefully Riske can find his form that he had early in the year, so that he can help this pen out with his flexible arm.
(3) Yost's usage of Villanueva is lacking. He is one of the best relievers on the team, in terms of allowing runners to score, and he stranded all his inherited runners to boot. However, Yost enters him into the game on a fairly rigid basis (in the middle of the game, the 6th inning specifically), and pitches him in entirely too few close games.
SOLUTION: When moving Mota and Gagne out of their roles, use Villanueva to pitch the close games they would normally inherit. Utilizie Villanueva's multiple IP ability in a more flexible manner.
(4) Torres' usage since becoming a closer is lacking. With Mota and Gagne in need of a demotion, Torres is not the best candidate for a shoirt reliever that does nothing but pitch in a single inning.
SOLUTION: Use Stetter to cover either Mota or Gagne's job, and use Torres in a more flexible way.
(5) Pitch Shouse with bases empty more. He has been a great fireman for two years, but his skills have shifted thus far -- he is allowing his own runners to score infrequently, but is much worse with inherited runners.
He is the only pitcher to have more inherited runners than games, by a long shot.
(6) McClung cannot get back to the pen soon enough. With Mota and Gagne demoted, he could be a solid middle innings option.
****
My pen:
Closer flexible: Torres
Closer two: Stetter
Close game flexible: Villanueva
Close game flexible: Riske
Close game short: Shouse
Middle game: McClung
Mop up / garbage time: Mota or Gagne