In case you haven't noticed, relief pitchers are taking it in the shorts so far this season, and there's no help in sight. Tony LaRussa's Cardinals alone blew a half-dozen leads in the ninth inning in April, and the late-inning meltdowns are becoming a daily staple.
The fact is that more leads are blown in the eighth inning than in any other inning, partly because managers are hoarding their closers--i.e. their best reliever--for ninth-inning duty, leaving inferior pitchers for the earlier innings. A great example of the pitfalls of this approach occurred over the weekend, and I feel compelled to harp on it here. I've written about this phenomenon before (see "My Favorite Box Score of the Month," archived in December 2009), but Marlins manager Edwin Rodriguez didn't get the link I sent him because he fell right into the same trap.
The Marlins played at Cincinnati on Saturday (April 30), with Josh Johnson starting. Johnson has been the best pitcher in the National League so far this season, with a 4-0 record and 1.06 ERA going into the game. It was business as usual for him--seven shutout innings, only five hits allowed as he nursed a 1-0 lead through the whole outing. The Marlins got him two more runs in the top of the eighth, at which point he had a 3-0 lead and an ERA of 0.88 for the season. You had to like his chances, right?
Edwin Rodriguez cared about only one number: Johnson had thrown 117 pitches, so out he came. The decision was an easy one. A pair of left-handed hitters--Jay Bruce and Joey Votto--would start the Reds' half of the eighth, and Rodriguez had just the man for the job. Randy Choate has become one of today's managers' favorite type of specialist, the southpaw who is particularly tough on lefties. In his career, lefties have hit just .217 against him, and only .202 in 2010. Sure enough, Bruce and Votto were no match for him; he struck them both out on an impressive seven pitches.
That brought up Brandon Phillips, and also brought Rodriguez heading from the dugout to the mound. Let's pause for a second. One of the problems with making substitutions is that you don't know how the new player will perform. This uncertainty is especially true for pitchers. An outfielder inserted as a defensive replacement might not have a ball hit to him, but a relief pitcher will be the center of the action from the time he steps on the mound. There is less margin for error. So the manager has to--or should, since apparently not all of them do so--consider the likelihood that the pitcher entering the game will do a better job than the pitcher you have out there right now.
In this case, Rodriguez had some solid information to go on: Choate looked terrific. He was throwing strikes that the batters couldn't hit, including the reigning NL MVP. His effective stuff was a known quantity compared to the uncertain stuff of the pitchers warming up in the bullpen. He faced a .279 career average by right-handed hitters like Phillips, though it was over .400 in 2010 (which means that earlier in his career, when he faced righties and lefties about the same number of times, he was more effective than righties than he has been recently, when his managers have been more reluctant to let him face righties). In other words, in 2010 righties were twice as likely to get a hit off Choate as lefties were--which sounds pretty compelling until you remember that even with that disparity, Choate was still a sizable favorite to get him out. Yet Rodriguez took him out in favor of right-hander Edward Mujica.
Here's my point. Why did Rodriguez feel such an urgent need to remove Choate? He had a three-run lead, two outs, and nobody on base. What was he afraid of? That Phillips was twice as likely to get a single or a double? So what? Even though the Reds had several righties in a row due up, they would all have to get on base in order to present Rodriguez with the crisis he thought he was averting by removing a pitcher who had just struck out two batters on seven pitches. Why not see just how good Choate was today? Why not give him a chance to make an out pitch against Phillips, or even against the hitter after him? Why take out an effective pitcher and bring in a new one who might be throwing like a batting practice pitcher today?
Was there some compelling reason to have Edward Mujica in the game at that point? Let's see--career average by righties against him: .272; by lefties: .274. In 2010, righties actually hit 41 points higher against him than lefties did. So even if you accept a manager's prerogative to make game decisions based strictly on percentages, allowing the conclusion that Choate needed to come out of the game, there was no statistical basis for thinking Mujica would do better. Righties hit him about the same as they hit Choate. Throw in the uncertainty factor--the vague odds that Mujica's stuff be as good as Choate's--and it's hard to see why Rodriguez was in such a hurry to match up Mujica against Phillips.
You know what happened, of course--I wouldn't be blogging about it if Mujica had mowed down the Reds. Phillips doubled, about what Rodriguez feared he might do against Choate, except that now he had a pitcher on the mound who suddenly didn't look anywhere near as good as Choate did for his seven magical pitches. Johnny Gomes singled, scoring Phillips, and went to second on the throw home. Miguel Cairo singled in Gomes, and the score was 3-2.
Pause here a second. First we saw Rodriguez take out a pitcher who had just struck out two straight hitters. Now he has a pitcher on the mound who has given up three hits in a row. And this is the guy Rodriguez chooses to face another hitter! Not the guy with good stuff, but the guy who can't get anybody out. It wasn't until Ramon Hernandez followed with another single that Rodriguez got the message and took out Mujica.
In came Ryan Webb, another denizen of the lower reaches of the Florida bullpen. He promptly surrendered a single to Paul Janish on which Cairo scored the tying run. Even though Mujica was the chief culprit, the "blown save" was charged to Webb. That made three appearances in a row in which he blew a save, the previous two in the seventh inning. Meanwhile, closer Leo Nunez languished in the bullpen, waiting for the Marlins to take a lead so he could mop up another save. It never happened, as the Reds won in the tenth off yet another unproven reliever nobody has heard of yet, Mike Dunn. Rodriguez wound up using four relievers, including he three worst, with the game on the line, and succeeded only in getting his best reliever a day of rest.
As a Reds fan, I loved it. I was following the action on the computer, was delighted to see Choate exit and kept waiting for Mujica to come out. But he didn't until he gave up four straight two-out hits. That's what I hated to see, as a pitching analyst. The Marlins website has a "depth chart" indicating that Mujica is the bottom-ranked reliever, another way of saying that he's the worst pitcher on the staff. That's why so many leads are being blown in the seventh and eighth innings--managers are putting in their worst pitchers in the game first and hoping they'll hold on long enough to allow their best reliever to pitch a carefree ninth inning. More and more often, these managers aren't arriving at the ninth inning with a lead at all.
The hitch for the managers is that they want to avoid being second-guessed. They don't want to face the post-game press conference and have to explain why they went against the percentages and managed by gut feeling. The percentages can be defended, and if events overcame expectations, that's just part of the game and why they play on the field and not on paper. What I want to see is the Edwin Rodriguezes of baseball give themselves more credit. I hope that at some gut level, Rodriguez had an impulse to leave Choate in to face Phillips. I want him to think, "if I have the balls to let Randy get the third out here, I'll look like a freakin' genius!" Instead of being afraid of what some reporter might ask later, managers need to look past the percentages and see the evidence right in front of their eyes. Unless they're playing my Reds, of course. Then they can make as many short-sighted, horrible decisions as they like.
* * * * * * * * * * * * * *
I posted the above two days ago. If I had waited a day, I could've had an even starker example of narrow-minded managing. I'll leave out the names of the relievers, to protect the innocent, but the manager was Brad Mills. His starting pitcher, Aneury Rodriguez, was making his first major league start, which would account for Mills removing him after 83 pitches even though he had allowed only one hit through five innings. No matter. He was done for the day. In came a reliever who retired the side in the sixth with only a walk allowed. He left for a pinch-hitter. In came reliever #2, who retired both hitters he faced, striking out one. But he was a lefty and the next batter was a righty, so out he came. The new guy, reliever #3, retired the righty, and Mills' team took a 2-0 lead to the eighth inning.
Even though Mills had used three relievers who had recorded six outs while giving up nothing more than a lone walk, that wasn't good enough. Without having to, he brought in yet another reliever, replacing the one with the 1.72 ERA. Reliever #4 mowed down the other team in order on ten pitches, including a strikeout. He was the fourth reliever employed by Mills who didn't give up a hit. That wasn't good enough either. Mills had one more trump to play, his closer, Brandon Lyon, who had already blown three saves in the first month of the season and had a 4.77 ERA. His job is to close and by golly he's going to close, figured Mills.
So in came Mills with a 2-0 lead heading to the bottom of the ninth. He promptly walked the first man on four pitches. No problem. Next came a single, and still another single, but that didn't trouble Mills either. So what if his closer had already given up more hits than the other five pitchers yielded in the first eight innings? He was the closer, it was his job, and he stayed out there--long enough to give up a third straight single that tied the game, 2-2--and very quickly after that, a game-winning double.
Let me cut to the chase. This kind of game is why I could never have lasted as a beat reporter. After the game, I would've asked Mills, "What was there about [#4's] performance which persuaded you that he had to be replaced?" If by some chance I got a civil answer to that one, I would've continued with, "What was there about [#3's] performance which you thought #4 could improve on?" and so on up the line. Then I would've tracked down those relievers and asked them, "while you watched Lyon blow the game, were you thinking 'jeez, I could've done better than that if he had left me in?"
The only encouraging thing about this sequence of events is that once again my Reds were the beneficiaries. Held to one hit over eight innings by five pitchers, they finally found a pitcher they could hit--and the opposing manager obliged by deciding that this was the guy he would allow to keep pitching no matter what, until the game was completely lost. That's two reprieves in four days. Maybe if enough National League managers either don't read this blog or don't learn from its lesson, the Reds will get the whole pennant handed to them.
I just know that I'm going to continue my vigil for senseless bullpen breakdowns, especially when the Reds are playing. Will Dusty Baker fall into the same trap? Or has he learned from what he has seen the last few days?