Posted in MLB, Philadelphia Phillies, Sabermetrics | Print | 18 Comments »
Over the years here, Charlie Manuel‘s decision-making has been a frequent topic of discussion, ranging from his overuse of starting pitchers, improper use of his relievers, reliance on the sacrifice bunt, and more, there is usually at least one or two controversial tactical decisions in any given week. There were a few in this afternoon’s series finale, a 4-3 extra-innings loss in San Francisco as Jonathan Pettibone opposed Barry Zito.
The first questionable decision Manuel made was to allow Pettibone to hit for himself to lead off the sixth inning. The right-hander had thrown 90 pitches over five innings but labored hard in the fifth inning allowing one run while throwing 24 pitches. In his prior three starts at the Major League level, Pettibone maxed out at 96 pitches. It’s not that Pettibone can’t go over 100 pitches, it’s that he has typically run out of gas by the time such a decision needs to be made. In that fifth inning, his fastball only once crossed 92 MPH (in 24 pitches) after doing so three times in 13 pitches in the first inning. The 22-year-old still needs to develop the stamina to pitch deep into games.
Additionally, in a close game — the score was 2-1 — who is more likely to keep the game close? A fatigued starting pitcher his third time through the opposing lineup, or a reliever fresh out of the bullpen? According to Baseball Reference, opposing hitters posted a .695 OPS against starting pitchers the first time, .734 the second, and .775 the third. Against relievers, they had a .704 OPS the first time.
This is without addressing the offensive aspect. We don’t have anywhere near a large enough sample to really know how good Pettibone is with the bat, so let us assume he is league-average. The average NL pitcher has a .149 weighted on-base average, and we can safely assume Pettibone doesn’t hit for much power, so it’s almost all singles. The right-handed pinch-hitting options include Erik Kratz (career .287 wOBA vs. LHP) and Freddy Galvis (.296). To put it in a broader scope, the run expectancy with the bases empty and no outs (start of an inning) is 0.48 runs; with a runner on first base and no outs is 0.82; with the bases empty and one out, it is 0.27. It is not a huge difference, but paired with multiple tactical mistakes over the course of a game, it adds up quickly.
Pettibone got burned in the sixth with a lead-off walk to Brandon Belt, but rebounded to get the next two outs. By that point, he had thrown an additional 17 pitches, putting him at 107 pitches. The right-handed Guillermo Quiroz, with a .255 wOBA, strode to the plate. The two options here are: 1) take Pettibone out and bring in a right-handed reliever; or, 2) intentionally walk Quiroz to bring up the pitcher and potentially force the Giants to pinch-hit, in which case you take Pettibone out for a reliever. Manuel opted partially for #2, but left Pettibone in as the Giants let their starting pitcher hit. With the bases loaded, Pettibone threw a 90 MPH fastball and Zito put in play, just out of first baseman Kevin Frandsen‘s reach to plate the Giants’ third run.
That play brings up an interesting question: why was Frandsen holding on Quiroz? Quiroz, a catcher, is not a good runner and holding him on only leaves more room for a ground ball to sneak by. Observe:
There are two outs. The only reason you’d ever want to play with the first baseman up like that is if there are fewer than two outs and you want to go for a force out or tag play at home plate. If the first baseman corrals a ground ball with two outs in this situation, his play is 100% to first base, so he should have been playing back. Quiroz is only scoring on balls deep to right-center, and holding him on or not has zero effect on that.
After that, Manuel finally took Pettibone out for reliever Jeremy Horst to face the left-handed Gregor Blanco. Horst couldn’t find the plate and walked Blanco on four pitches, bringing up the right-handed Marco Scutaro. I didn’t have a problem with leaving Horst in for three reasons: 1) control isn’t something he usually struggles with, so there is no need to worry about the four pitches he threw to Blanco; 2) Marco Scutaro doesn’t show a platoon split over his career: .322 wOBA against right-handers and .325 against lefties; and, 3) Horst hasn’t shown a platoon split in limited MLB action: 4.16 xFIP against left-handers and 4.31 against right-handed pitchers. I’m not sure opting for a right-hander like Phillippe Aumont or Chad Durbin would have been an improvement.
Rewinding a bit, in the top of the sixth, Rollins had reached second base with a one-out double, bringing up Kevin Frandsen. On a 1-2 count, Rollins attempted to steal third base but was thrown out. As with leaving in Horst, I also didn’t have an issue with Rollins’ stolen base attempt despite the poor outcome. As Josh Goldman showed at FanGraphs in November 2011, the break-even rate for stealing third base with one out is 69 percent. Rollins, entering his 13th full season, is a career 83 percent base-stealer overall and 88 percent stealing third base. Opposing base-stealers had a 62 percent success rate against Zito, but a majority of runners were thrown out by Bengie Molina and Eli Whiteside. Opposing base-stealers had a 79 percent success rate against Quiroz, who had never caught Zito prior to this afternoon’s game. In 2012-13 combined, Rollins was a perfect 14-for-14 stealing third base before his attempt today.
As you can see in the above .gif, Rollins ran on a high-and-inside fastball all of 84 MPH with a right-hander batting, all in all not a terrible choice. He takes a weird route as he ends up stepping briefly on the infield grass, which I can only imagine was a mistake. Additionally, his slide was terrible, something you rarely hear about with Rollins.
Finally, getting to third base allows him to score the tying run on some infield grounders and most outfield fly balls. On second base, he only scores on some singles and all extra-base hits. Frandsen is not an extra-base hit machine, as 75 percent of his 210 career hits entering the game were singles.
(To clarify, Rollins likely attempted the steal of his own volition, so this shouldn’t be blamed on Manuel.)
The last strategical point I would like to touch upon is Manuel’s continued refusal to use “closer” Jonathan Papelbon in high-leverage situations in a tie game on the road. Between the time Pettibone exited the game in the sixth and when the game ended in the tenth, ten Giants batters came to bat with a leverage index of 1.00 (average) or greater, including eight in excess of 2.00 (high-leverage). Papelbon, ostensibly the team’s best reliever and certainly the most expensive, really hasn’t been used in the most important situations. Here’s a breakdown of the batters he has faced by leverage index bucket:
The Phillies are more or less assured victory in a large majority of the games in which they use Papelbon, which largely invalidates the need for a $50 million reliever at that point. Between April 23 and today, Papelbon has been used five times in a span of 16 days, only one time involved a save situation, none of them high-leverage. This afternoon was a perfect opportunity to use Papelbon. The Phillies were in trouble in both the ninth and tenth innings and could use Papelbon’s high strikeout rate (30 percent career) and high infield fly ball rate (15 percent career) to limit base advancement. This is not to say the use of Mike Adams and Antonio Bastardo was in error, but that Papelbon could have been brought in at the first sign of trouble since he is the team’s best reliever.
Some of these tactical blunders are nothing new — Manuel has been making them since he took over the team in 2005. However, the 2007-11 era teams were good enough in all facets of the game to make up for Manuel’s poor decision-making. Now, with a depleted starting rotation and one of the league’s weaker offenses, Manuel’s impact on the game is magnified as the Phillies can afford few mistakes.