In two of the three games this series, the Marlins were afforded a tight lead over the Atlanta Braves late in the game. In all three games, the two teams were very close. In the two Marlins losses, however, the bullpen disastrously blew the games wide open with horrendous pitching. Games 2 and 3 were some of the worst bullpen performances that I have witnessed in recent memory.
Series Hero: Nate Robertson (0.252 WPA)
Series Goat: Brian Sanches (-0.374 WPA)
Impressed By: Robertson (6 IP, 5 K, 2 BB, 0 HR, 0 R, 2 H)
Depressed By: The entire bullpen (see below)
Bullpen implosions lead to two losses
In both games, there were strong drops in the Marlins’ win expectancy. In the second game of the series, the Marlins led 2-1 entering the 7th inning. However, the bullpen at that point took over, surrendering six runs and 0.568 WPA to the Braves in just that inning alone. The Marlins sent Burke Badenhop, Dan Meyer, and Brian Sanches in an attempt to finish the inning, but none of those players accomplished much. As a group, those three relievers combined for three walks, five hits, no strikeouts, adn the aforementioned six runs. Sanches himself melted down for a WPA of -0.460.
The next night saw a similarly poor performance. The Marlins had climbed back into a 3-2 deficit by the end of the fourth. After Ricky Nolasco was pulled due to a rain delay, the Marlins sent out Jay Buente, Tim Wood, Taylor Tankersley, Dan Meyer, and Clay Hensley in that order. The collective unit gave up the remaining five runs, although this time it was spread out over the five remaining innings. This included a disastrous eighth inning by Meyer in which he allowed two singles and five walks, including one with the bases loaded. Meyer was able to record a double play with the bases initially juiced, but then an error from Jorge Cantu scored another run. Meyer’s final out mercifully came against relief pitcher Takashi Saito,
Fredi kept Nolasco despite rain
Manager Fredi Gonzalez ran into an interesting situation in the fourth inning. After Cameron Maybin walked to load the bases, Fredi was given the choice to either stay with starter Ricky Nolasco at the plate with one out or pull him in favor of a pinch hitter like Wes Helms. At the time, I asked whether or not the right call was to pull Nolasco from the game. Thanks to run expectancy, we can actually come up with a crude estimate of which option would have been the better call.
By leaving Nolasco in the game, the Marlins would presumably plan to pitch him at least two more innings. Leaving him in, however, would likely squander the team’s bases loaded opportunity. According to these RE charts provided by Amazin’ Avenue’s James Kannengeiser and DRaysBay’s R.J. Anderson (thanks guys), the run expectancy with bases loaded and one out was about 1.55-1.65 runs (scored by the end of the inning). The RE with bases juiced and two outs is about 0.85 runs. Giving Nolasco some credit for his shot at getting a hit (which he of course did), we can expect that keeping Nolasco at the plate cost the team an average of 0.5 to 0.6 runs.
What about the alternative? Well, assume we send Mike Lamb to the plate. Given his projected values by ZiPS, Lamb would project to add 0.04 runs to the situation given his odds of hits or outs. However, Lamb is a below average hitter, and starter Tim Hudson is a slightly above average hitter. As a result, I too about 0.1 runs away from that with some back-of-the-envelope calculations, giving him an expected value of -0.06 runs. Knock another 0.02 runs off of that for the pinch-hitting penalty (players hit worse off the bench than as starters), and you’re down to -0.08 runs. Finally, consider that Clay Hensley, the likely long reliever for the game, is projected at about 0.2 runs worse in two innings than Nolasco, dropping the expectancy to -0.28 runs, still a good 0.2 runs better than the best Nolasco expectancy.
Then, of course, figure in the fact that the rain would have likely prevented Nolasco from going back out anyway, and you have to think that Fredi should have pinch hit for Nolasco, regardless of the actual result.