January 21, 2009
Aaron Harang's Extra-Inning Adventure
Quantifying the contributions of a baseball manager is a task even the geekiest of us sometimes struggle to break down. Are there certain decisive standards to which managerial judgments can be compared? How much influence does the front office exert over his strategic choices? In such a role-defined era, which situations would merit the usage of a different reliever to set up the closer? Does the manager define these roles? What types of decisions can a manager actually be credited with, given that most alternatives would act the same? More questions along these lines could be added to the evaluation stew, but one axiom remains the same: the true job of a manager involves putting his players in a position to succeed.
A competent manager will never slot Willie Bloomquist in the cleanup spot. Nor will he allow Ivan Rodriguez to patrol center field. And he definitely would not ask his ace pitcher to throw four innings of relief, on very short rest, in a meaningless early season game when at least one alternative could have logged the same frames.
Wait, scratch that last one, because that hypothetical actually took place, when Dusty Baker called on Aaron Harang to pitch four innings out of the pen in an 18-inning game against the Padres on May 25. Now, Baker's reputation for handling a starting rotation is already about as sterling as Lindsay Lohan's for showing up on time. Justifiably or not, his reputation is directly related to his handling of Mark Prior and Kerry Wood, and their subsequent struggles. It should come as no surprise that four of his starters last season-Bronson Arroyo, Edinson Volquez, Aaron Harang, and Johnny Cueto-all finished in the senior circuit's top twenty in pitcher abuse points.
Due to the commonly expressed opinion that Harang's disappointing season was a direct result of Baker's relief experiment on May 25, analysts tend to break his workload into the following three segments: Pre-relief Outings (covering the beginning of season through to May 22); Post-relief Outings (May 29 through an arbitrarily selected August 16); and Arbitrary 8/16 Marker through to the end of season (August 22 through September 30). Breaking it up in this way seems to indicate that Harang bookended his season with very solid performances, yet struggled mightily in his starts from 5/29-8/16, posting a 9.06 ERA.
Period GS IP HR/9 BB/9 K/9 ERA 3/31-5/22 11 74.2 1.2 2.2 7.6 3.50 5/29-8/16 10 51.2 2.8 3.3 7.5 9.06 8/22-9/30 8 54 1.5 2.0 6.3 2.83
There's a problem with this tidy arrangement, however. A slight adjustment needs to be made in order to more accurately reflect what happened to Harang in-season, because he suffered an injury near the flexor mass in his right forearm, shelving the ace from July 9 through August 9. In the segment markers noted above, the middle period of ineffectiveness includes two of his starts that occurred after the time spent on the disabled list. The more relevant divisions would be: Pre-relief Outings (4/4-5/22); Post-relief outings to DL (5/29-7/9); and Post-DL (8/10-9/30). This shift takes two hot-off-the-DL starts that some prefer to cobble together with his earlier mid-period struggles (making the mid-period appear worse and the end-season better), and adds the 16 earned runs given up in 7
Here are Harang's peripheral statistics with the latter two segments corrected:
Period GS IP HR/9 BB/9 K/9 ERA 3/31-5/22 11 74.2 1.2 2.2 7.6 3.50 5/29- 7/9 8 44.1 2.1 2.9 7.3 7.31 8/10-9/30 10 61.1 2.2 2.3 6.6 4.84
In his first 11 starts Aaron looked much like his old self, the guy that averaged 220 IP with a 3.7 K/BB, 3.75 ERA, and 3.70 FIP over the previous three campaigns. It's no wonder that Baker routinely finds himself blamed for Harang's season of disappointment and mediocrity. After the relief outing and up until he landed on the disabled list, Harang barely averaged more than five innings per start. His K/BB dropped by approximately one point, and his run-prevention skills eroded to the tune of an ERA over 7.30. After returning from injury, Harang's peripherals fell somewhere in between. He did not return to the 3.8 K/BB, 3.70 ERA pitcher, but he definitely looked better than his post-relief/pre-DL segment.
These statistics do not explain the entirety of the situation, however, as Harang pitched in front of shoddy defenders while having to call a bandbox of a ballpark home. It would be very easy for a pitcher of even his caliber to go through an awful stretch under those circumstances, regardless of the relief outing. Were there underlying factors aside from the relief appearance that caused the decline in peripheral statistics? After the relief outing, did Harang pitch the majority of the next stretch at home, at Great American Batting Practice Stadium? Did he happen to face teams like the Phillies, Cubs, Cardinals, and Marlins right away? Along similar lines, were his opponents prior to the relief appearance in the bottom tier of offensive prowess? Did interleague play start around then, necessitating Harang to pitch against the likes of the Red Sox and other dominating junior circuit teams? This isn't to say that Dusty should be blame-free, but before officially wagging the finger, these questions merit investigation.
Following his dominating four-inning, nine-strikeout relief performance on May 25, Harang's next eight starts were evenly split between home and road, eliminating that as a distinct possibility for the statistical decline. His opponents did include the Phillies, Red Sox, and Marlins in that stretch, but they also featured the Pirates (twice), Blue Jays, and pre-Manny Dodgers. Overall, the average team EqA that Harang faced stayed virtually identical throughout the season. In the first segment, he faced teams with an average EqA of .263, a mark that remained the exact same during his poor post-relief/pre-DL stretch. After recovering, his opponents improved to a much more potent .264. It would seem then that the underlying factors potentially capable of getting Baker off the hook are unequal to the task.
The really interesting statistical findings with regards to Aaron's splits deal with his Pitch F/X data. Take a look at his pitch frequencies by segment:
Period FA CU SL CH 3/31-5/22 66.9 6.1 23.0 4.0 5/29- 7/9 59.0 5.7 21.0 14.3 8/10-9/30 69.4 4.3 19.1 7.2
Primarily a fastball/slider righty, Harang traded in a high percentage of his heaters for extra changeups during that bad stretch of starts. He also threw a higher percentage of pitches right around the middle of the plate. How about the movement, velocity, and release point of his fastball by segment?
Period Velo Horiz. Vert. Release 3/31-5/22 89.4 -3.7 11.2 6.96 5/29- 7/9 89.4 -3.4 9.9 6.95 8/10-9/30 89.4 -4.7 10.4 7.18
The first aspect that really sticks out is the eerily consistent velocity throughout the season. Harang averaged close to 90.6 mph on his fastball from 2005-07, but it appears that he noticeably lost velocity this past season. With as much movement as Harang gets on his pitches, a velocity dropoff like that can be hidden. But what if his movement also suffers? As the numbers above show, he lost both vertical and horizontal movement on the pitch during the middle segment. So, Harang did not throw as hard as in the past, he also lost a substantial amount of movement, and he threw a higher percentage of pitches near the middle of the zone. That's not exactly a formula for success.
Following his DL stint, even though his starts on Aug 10 and Aug 16 produced lines even Radhames Liz would scoff at, Harang's pitch data improved. He gained over an inch in horizontal movement and a half-inch in vertical movement, two numbers very important for a pitcher who needs that tail on his tailing fastball. With consistent velocity, increased movement, and improved control, Harang also threw his highest percentage of fastballs. His higher release point reflected his ability to get on top of the ball, a facet of pitching that those with tailing fastballs and sliders strive to perfect. Scanning the numbers once more, something very interesting stands out: over the course of the season, Harang's velocity and release point showed consistency, for the most part, while his movement, well, moved.
According to Will Carroll:
The basics of biomechanics are in play here. Broadly, the shoulder is responsible for velocity while the elbow tends to focus more on control. We see this in players coming back from Tommy John surgery who have to re-learn that proprioception, that feeling of "connectedness" between the hand and elbow. Of course, the forearm (pronation or supination) and fingers (grip) factor into it as well and any breakdown in the kinetic chain will lead to a loss of energy and hence a loss of velocity.
The above contention really boils down to the idea that, even though all parts of the kinetic chain can adversely affect one another, the shoulder exerts the majority of its influence over velocity and release point. Likewise, the elbow and forearm deal much more with control and movement. Since Harang's injury came near the flexor mass in his right forearm, it would be expected that his fingers would also be affected, vastly hindering his ability to grip or turn pitches. Due to this, his movement theoretically should have been down, just like his control should not have been expected to be typical of the Harang we know and love. Since the problem surfaced in the forearm, though, minimal, if any, differences should have been expected in his velocity and release point. Basically, do not be fooled by the fact that his velocity and release point stayed the same for most, or all, of the season. His movement and control suffered, and even though these numbers are not as sexy as the velocity, they are direct results expected and predicted based on his injury.
How did Harang get hurt, though? Was it the fact that he had thrown a minimum of 99 pitches in the 11 starts leading up to the relief outing? Did the bullpen experience tip him over the edge? Is Harang simply not effective when fatigued? The most likely answer is that Harang was overworked, developed fatigue, and instead of getting rest, toed the rubber whenever possible, especially on short rest. If fatigued, and a breakdown in proprioception occurs, the worst action a pitcher can take is to continually force himself to pitch. Or, in Harang's case, be forced by his manager to pitch. Harang threw 103 pitches on May 22, 63 in the relief outing two and a half days later, and then 73 in his May 29 start that also came on short rest. Did Dusty have a choice, though? Was Harang really the only viable candidate to log those relief innings?
The May 25 game between the cellar-dwelling Reds and Padres lasted 18 innings and saw Baker call upon every available designated relief pitcher. In fact, he had used his entire bullpen up after the first 12 innings of action. Matt Belisle could not escape the fifth inning, necessitating the use of the bullpen early in the contest. At that juncture of the season, the team generally used seven relievers: Mike Lincoln, Bill Bray, Josh Fogg, Jared Burton, Jeremy Affeldt, David Weathers, and Francisco Cordero. Examining the previous four games for the Reds, it seems that the only reliever who perhaps merited the day off was Cordero. Of course, Cordero entered the ninth inning of this game looking for the save, and blew it; he pitched the tenth inning as well. Josh Fogg started the eleventh inning but only recorded two outs before being lifted in favor of Bray. Though Bray pitched 1
Instead, the entire relief corps finished their daily work, and as things turned out, there was a third of the game yet to play, meaning starting pitchers would need to enter the ballgame. Harang had started on May 22, the first game of this series. Volquez started the following day, and Arroyo got the nod after that. Belisle, their fifth starter, began the game in question. This meant that only one starter would realistically be relatively fresh at that point: Johnny Cueto. Granted, I am not a major league manager, but in that situation I would have the equivalent of a flowchart in my head. If all relievers used up, go to your long reliever. If your long reliever gets used up, go to freshest starter. If freshest starter used up, go to your next freshest starter. Unfortunately for Harang, Baker opted to go with: "if long reliever used up, skip freshest starter, go right to second freshest starter because he is an innings-eater." Or, perhaps he remembered the past ire directed at him for ruining young arms and wanted to go with an older pitcher in this situation. Either way, it still doesn't help us sort out why the long reliever wasn't utilized properly.
A game lasting as long as this is not commonplace; perhaps we see one such game every two or three years. Still, burning the bullpen out is a problem that has become commonplace, especially in today's micro-managed game. Considering that Harang received ample time to rest and heal over the winter, he should return to form next season. This game, however, should serve as a reminder that teams need to have a mop-up man or long reliever on their 25-man roster to be used as a just-in-case-pitcher; the guy that goes out there and throws five innings of relief when no other relievers are left. Only after he can no longer pitch should the starters begin to enter the game. That Baker used his equivalent of the just-in-case-pitcher for less than a full inning, when he was also fresh, and the team had a day off on the horizon, is extremely irresponsible. That he opted for his ace over the much fresher Cueto adds to the irresponsibility.
As far as other takeaways, Harang does not seem capable of being effective when used while fatigued. He should not be asked to throw a multitude of pitches on short rest as often as he was this past season. This seems like an easy concept to grasp, but something in the back of my head suggests a similar overusage-while-fatigued situation will present itself with the Reds next season. The axiom mentioned at the beginning of this article-that a manager should put his players in a position to succeed-can be offered in tandem with another, which is that a manager should also learn from his mistakes. Baker not only struggles with putting players in the positions to succeed, but it also looks like he has not shown that he understands the mistakes he's made in the past, let alone learned from them.