On Oct. 18, Tim Lincecum made his lone start of this postseason in Game Four of the National League Championship Series versus the Cardinals. He required 91 pitches to make it through 4.2 innings, and he allowed four runs on six hits and three walks against only three strikeouts.
Since then, Lincecum has returned to the bullpen, and in two appearances versus the Tigers, he has combined to throw 4.2 innings. That is the full extent of their similarity. Coming out of the ‘pen, Lincecum has not allowed a hit and has surrendered just one walk. He has struck out eight of the 16 batters he’s faced but has thrown only 64 pitches.
The difference between his results as a starter and as a reliever in his last three appearances is as stark as the difference between Lincecum as a starter before 2012 and in 2012, but is that the end of it? Should Lincecum become a fixture in the bullpen, and should we expect him to continue to be effective there?
The spike in fastball velocity that you would expect from a transition from starter to reliever has been there for Lincecum. In September, Lincecum made six starts. His average velocity of his four-seam fastball was down to 89.8 mph. In October, Lincecum has made six appearances, only one of which was a start. His average velocity of his four-seam fastball was up to 90.7 mph.
However, that is not the full story. In his Oct. 18 start, Lincecum averaged 90.5 mph on his four-seamer. His velocity was similar to that in his excellent relief appearances, but he was ineffective. Also, it seems a bit of a stretch to call his low-90s fastball an improvement. He sat in the low-90s in June and allowed 18 runs in 22.2 innings. As a reliever this postseason, Lincecum is still well short of his typical fastball velocity from even last season.
The lack of a full recovery of his once-dominant fastball makes Lincecum’s successful transition to the bullpen fairly surprising. Intuition tells us that it is easier for pitchers to succeed as relievers than as starters because, with shorter appearances, they do not have to conserve their energy, and they do not have to face batters multiple times. There is little evidence in Lincecum’s regular-season statistics that those were the major factors in his performance decline.
In 2011, Lincecum showed a typical drop in his effectiveness when batters faced him multiple times. The first time through the order, he allowed a triple slash of .190/.285/295. The second time through, he allowed a .220/.310/.346 line. The third time through, those numbers rose to .238/.293/.364. That trend was supported by his allowed BABIP, which increased from .246 to .288 to .313 in step with his opponents’ production increase each time through the order.
In contrast, Lincecum showed an unusual decrease in his allowed BABIP through the order as a starter in 2012. His first time through the lineup, he gave up a .337 BABIP and a triple slash of .253/.343/.405. His second time through, he allowed a .306 BABIP and a triple slash of .259/.328/.456. His third time through, he allowed a .278 BABIP and a triple slash of .247/.347/.416.
If that trend continued, there was little reason to expect Lincecum to become effective because he only had to face batters once in shorter relief appearances, and if was due to improve because of BABIP regression, it seemed as likely to happen for him as a starter as it would as a reliever.
Lincecum’s effectiveness out of the bullpen appears to be less about the role and more about a change in his approach. As a starter in the regular season, Lincecum followed the same patterns he did in previous seasons. His first pitch selections and strike percentages were very similar.
He threw a first-pitch fastball 41.4 percent of the time in 2011 and 40.1 percent of the time in 2012. Lincecum threw that fastball for a strike 52.2 percent of the time in 2011 and 56.1 percent of the time in 2012. However, his fastball was diminished, and whereas batters swung just 20.0 percent of the time and missed 4.9 percent of the time in 2011, they swung 23.2 percent of the time and missed only 3.7 percent of the time in 2012.
In the postseason, Lincecum has made a dramatic shift in his first-pitch selection. He has thrown his fastball more often—46.3 percent of the time—but has thrown it for a strike only 45.2 percent of the time. Batters have swung at those first-pitch fastballs only 9.7 percent of the time, less than half as often as in the regular season.
Meanwhile, Lincecum has been more willing to start batters with a strike with one of his secondary pitches. In fact, he has thrown 65.4 percent of his first-pitch changeups, sliders, and curves for strikes, up from 54.0 percent in the regular season.
The sample size is small, but Lincecum looks like he has made an adjustment in his new role as a reliever. The easy narrative is that Lincecum is more willing to show all of his pitches because he does not need to worry about the second and third times through the lineup.
However, that might be a similarly effective approach to take as a starter. Batters have been more aggressive against his diminished fastball this season. If Lincecum continues to throw strikes with his off-speed pitches early in counts, he could use their aggressiveness to his advantage and produce more swings-and-misses and weakly hit balls in play.
How Lincecum approaches batters going forward could be a significant factor in how the Giants decide to use him in 2013, and possibly beyond.
References & Resources
Pitch selection, velocity, and strike percentage data from PitchFX data at Texas Leaguers.
Lincecum’s splits by inning from Baseball-Reference.
All other statistics from FanGraphs.