Sometimes, you sit down to write a column with an idea of where the column is going, only to find that the evidence veers you off course. You had an axe to grind, a vendetta to fulfill, but the data simply didn’t cooperate.
What do you do? If you’re, I don’t know, Gerry Fraley or something, you just plow on ahead, logic and facts be damned. But the whole creed of sabermetrics is that you go where the evidence takes you.
I sat down last night planning to rip on Trey Hillman for his bizarre usage of Joakim Soria. Here’s what I wrote:
“I was a fan of the decision to hire Trey Hillman, and I still think that he has the makings to be a fine manager, the best we’ve had in a long time. But more and more I’m becoming concerned that, while he may be an asset in the clubhouse, he’s not a competitive advantage when it comes to making tactical on-the-field moves. In order to have a competitive advantage on your opponents, you have to do things differently than your opponents. Hillman has shown some signs of that, most notably his decision to occasionally play the entire outfield shallow in the hopes that the doubles and triples which get past them will be more than made up for by the many would-be singles that don’t fall in.
When Andruw Jones was at his absolute defensive peak – at which time he was probably one of the five greatest defensive centerfielders of all time – the key to his success was that because he was so good at reading balls off the bat, and had such a quick first step, that he could play 30-
I have no idea how much of Jones’ brilliance came from his talents and how much came from his positioning, but since then I’ve always wondered whether teams really have positioned their outfielders in an ideal manner. We now have enough data on batted balls to figure out whether moving outfielders
I digress. For as much as Hillman might be willing to think outside the box with his defense, his handling of his pitching staff is pure, unadulterated conventional wisdom. I speak specifically to his philosophy on using Joakim Soria, a philosophy that to this point seems predicated on extracting as little value from his best pitcher as possible.
Joakim Soria has pitched in 11 games this season. In all 11 games he came in at the start of the inning. This in itself is a little bothersome – if you’ve got a pitcher that never gets into a jam of his own, why not use him to clean up other people’s messes either? If a first-and-third, one-out jam in the eighth doesn’t call for Mr. Incredible, what does?
Anyway, Soria completed all 11 innings, and in every game he was the final Royals pitcher – 11 games, 11 games finished. He has faced 38 batters this year; not only have none of them scored, but only one of them – the aforementioned Clete Thomas – has reached third base. Aside from Thomas, the only one to reach second base is Milton Bradley, who reached on a Grudzielanek error last night. Soria has retired 33 of 38 hitters, including 24 straight at one point. He’s given up a double, two singles, a walk, and the ROE. Opposing hitters are hitting .083/.108/.111 against him this year.
So if you’re the manager, and you have the perfect fail-safe weapon in your arsenal, how would you deploy it? You’d use that weapon as much as possible, and you’d use it in the most critical situations, right?
Soria has thrown 11 innings in 26 games; he’s on pace for 68.2 this season. That’s a little lower for a 24-year-old reliever, but I’ll let that slide.
Let’s break down those innings. Nine of the 11 appearances were in the ninth inning, one was in the 11th inning (Opening Day), and one was in the 8th. As we said before, Soria has been used exclusively to close out games. That’s what closers do.
Now look at the score when Soria came in:
Leading by four: 3
Leading by three: 2
Leading by two: 1
Leading by one: 3
Losing by one: 0
Losing by two: 1
Losing by five: 1
Twice Soria has been used when the Royals losing, including once in a pure mop-up role; both appearances were purely to get Soria some playing time after he hadn’t pitched in a while.
Six times Soria has pitched in a save situation, although two of those were when the Royals were leading by 3 runs, the softest save situation of all and one that drives us analyst types batty.
And three times Soria has pitched with the Royals leading by four runs, including last night.
I got this information from Soria’s page at baseball-reference.com, which also lists Soria’s “leverage” factor for each game. Leverage is a statistical tool that measures how much the game was on the line when a pitcher came into the game. A leverage score of 1.00 means that the leverage for that appearance was the equivalent of the start of the game, i.e. a 0-0 score in the first inning. The whole point of having a closer is that, while a closer might only throw 60-80 innings all year, he’s being used to protect small leads in the late innings, where the impact of a run allowed is huge. Typical closers will have Leverage scores around 2.
Four of Soria’s 11 appearances this year – basically, the four times he came in protecting a lead of one or two runs – have been high-leverage situations. His other seven appearances all had leverage scores of less than one. In nearly two-thirds of his appearances this year, Soria was used in a situation that had so little on the line that he would have been more helpful starting the game and then leaving after an inning.
That’s a criminal mis-use of resources. Why the hell do we need Soria to protect a four-run lead in the ninth? Yasuhiko Yabuta will protect a four-run lead 90% of the time. Hell, Hideo Nomo would get three outs before surrendering four runs 60-70% of the time. I’m not asking for Hillman to taunt the opposition by using the last man in his pen – just go with Leo Nunez, or Jimmy Gobble, and save Soria in the event they get into a real jam.
So far this year, Soria’s overall leverage – using the Baseball Prospectus measurement, which probably differs from the bb-ref stat a little – is just 1.41.”
Everything I wrote above is factually correct, but when I started looking at the facts a little deeper, I realized that as inefficient as Soria’s usage has been, I’m not sure how Hillman could have done much better.
I went through the box scores of every game this season, looking for high-leverage situations from the 7th inning on in which Soria could have pitched. In particular, I wanted to figure out why, from April 8th until April 24th, a span of 14 games, Soria pitched just twice – on the 15th and 16th. He had six days of rest, pitched in consecutive games, then had seven days of rest again. There had to be situations in which Hillman could have made better use of his best pitcher, right?
On April 4th, the Royals trailed by one run from the fourth inning until the game ended. The next day they would trail by two runs in the seventh and eighth. Soria had pitched on April 2nd and 3rd, and would pitch on the 6th, so it makes sense that he wouldn’t pitch in these two games.
On April 12th, the Royals trailed 1-0 going into the seventh, and 2-0 thereafter.
On April 19th, the Royals were tied 4-4 going into the bottom of the 7th before the A’s scored two runs; the Royals would score a run in the ninth but fall short.
And…that’s pretty much it. The Royals have now played 27 games, and in only four of them did they have a one or two-run lead to protect in the late innings. Soria pitched the ninth, and got the save, in all of them.
Just five of the Royals’ 27 games have been decided by one run; they’re 3-
To get an idea of how often the Royals should be playing one-run games, I went back to the 1985 NL, which averaged 4.10 runs per game. The 12 NL teams averaged 53 one-run games, ranging from 42 to 64 per team. So the Royals should expect to play a one-run game about a third of the time, but so far they’ve played far fewer than that.
(I ran these numbers before the mini-slugfest in
They have now played six two-run games, but they’re 1-
For whatever reason – or more likely, no reason at all – the Royals simply haven’t had many close leads to protect in the late innings. When they have had the lead, they haven’t coughed it up, Soria or not. Do you know how many times the Royals have lost a game in which they were leading after four innings?
Twice. On April 18th, when Bannister nursed a 2-1 lead into the sixth before coughing up four runs, and the next day, when the Royals had a 4-0 lead after four innings, but Greinke gave up three in the fifth, Ramon Ramirez gave up the tying run in the sixth, and the A’s scored the winning runs in the 7th. That’s the only loss the Royals’ bullpen has had all season.
There are only two other games the Royals have lost in which they were tied after four innings, and in both of those games it’s because the Royals didn’t score any runs after the fifth inning. On April 27th the Blue Jays broke a 2-2 tie with a run off Meche in the fifth, and won 5-2; on April 24th the Indians and Royals were scoreless until the Indians got to Bannister for two runs in the 7th.
The reason Soria hasn’t pitched more in high-leverage situations is simply that the Royals haven’t had many high-leverage situations. I mentioned before that Soria’s leverage score is 1.41. Well, Leo Nunez’s score is 1.56, which is higher than Soria’s (it’s not unusual to have a set-up man with a higher Leverage score than his closer) but not particularly high either. More importantly, every other reliever on the team has a Leverage score below one. The Royals have a great bullpen – well, a great top-half of a bullpen – and no opportunity to really use them so far.
And it’s hard to argue that Hillman’s usage of Soria has cost the Royals ballgames, because the Royals have yet to lose a game they were leading after 7, and have lost only two games were losing after 4.
It’s a lot more emotionally satisfying to throw stuff at my TV set when Hillman brought Soria in to pitch the ninth in a 9-5 game on Tuesday night. But looking at this rationally, we simply can’t evaluate Hillman’s use of his closer fairly at this point, because the season has so far conspired to give him precious few opportunities to do so.