March 24, 2006
Aim For The Head
New Relief Categories
In my previous mailbag article, I ran through some reader suggestions about a term for a "big lead squandered," and settled on a "choke" as the most promising term. Having put a name to it, and prompted by several reader inquiries, I decided to run some numbers on chokes in the Prospectus play-by-play database.
Interestingly, only 19 pitchers had a "choke" during 2005, and only one pitcher had more than one--Chad Bradford, with just two chokes. On July 26th, he came into a 5-1 game with two runners on base, and lost the lead on an Aubrey Huff grand slam. Then on August 9th, he came in with the bases loaded and a 7-3 lead, and gave up two hits, an intentional walk, and two run-scoring groundouts, and left with the game 7-7 and runners still on second and third.
Bradford is actually one of just 18 pitchers to have choked at least twice in a season, which has been done a total of 19 times since 1960. Two pitchers have choked three times in a single season, the others have done it twice. Gene Garber choked two times in two different seasons (1977 and 1979). Any guesses as to the identity of the two thrice-in-a-year chokers? The answer is at the end of this column.
The career leader in chokes is 6 by Dave Righetti, followed by Gene Garber with 5, and five pitchers with 4 each--Dale Mohorcic, Dan Quisenberry, Stu Miller, Jose Mesa, and Rollie Fingers. I admit to being surprised that chokes were not more common, and that the career leaderboard didn't even reach double-digits.
The biggest blown lead in the database was a 7 run lead, blown by Dyar Miller in the first game of a double header on June 13th, 1979. Miller came in to an 8-1 game with the bases loaded, but allowed all three inherited runners to score, plus was charged with four runs of his own while pitching less than an inning. Then, just to add insult to injury, he pitched the second game of the double-header as well, allowing four more runners to cross the plate (two inherited runners plus two of his own). A very bad day all around for Mr. Miller.
The most runs allowed in a choke since 1960 is 8, given up by the usually-great Dan Quisenberry on September 4th, 1980. He came into the game in the 9th inning with two runners on and a 4 run lead. He gave up seven hits and an intentional walk while retiring only one batter. Both inherited runners scored, plus six more charged entirely to Quiz, and the Royals lost to the Brewers 9-5.
As I alluded to in the mailbag, I started thinking about the whole "save-hold-blown save" system for classifying relief appearances, and wondered if we could do better. To begin with, I wanted to set some parameters on what we're trying to accomplish. My intent is to create a classification system for short relief appearances. Long relief stints--such as a pitcher who comes in during the 2nd inning and pitches until the closer takes over in the 9th, are more akin to starting efforts than the context-sensitive value of short stints, particularly towards the end of a game. I arbitrarily set the threshold at 3 innings--the maximum length of an outing that we'll concern ourselves with here.
Next, we need to identify the conceptual categories we want to characterize relief outings into. As the "Save/Hold/Blown Save" system focused on preserving leads, our parallel system should at least have counterparts for each of them. The categories I came up with are:
1) Preserve a small lead (equivalent to a Hold)
But why stop there? Relievers are also used to keep a game close enough to rally back, and to soak up innings in lost causes. We could conceivably concoct terms for the following cases that aren't recognized today:
6) Keep a game with a small deficit (or tied) from getting out of hand
I don't think there's a real reason for designating the reverse of a choke--a pitcher coming into a game with a big deficit and having his team rally to win. The rallying is more to the credit of the offense. The pitcher was just a bystander in a situation where he was expected just to eat up innings.
I took a first pass at coming up with terms and definitions for these scenarios. I'm making no claim that this is flawless or even well thought out. I'm just putting it out there for comment and feedback. Maybe this is overkill, and I'm solving a problem that no one is interested in (it wouldn't be the first time). Or maybe the combined BP community has even better ideas. But you've got to start somewhere... so, matching up to the category numbers above, here we go...
Begin -> Did he pitch 3 innings or less? NO = N/A YES -> Did he have the lead when entering the game? YES -> Was the tying run on base, at bat, or on deck when he entered the game? | YES -> Did he ever lose the lead at any point? | | YES = "Cough" | | NO -> Did he finish the game? | | YES = "Slam" | | NO = "Shutdown" | NO -> Did he ever lose the lead at any point? | YES = "Choke" | NO = "Cinch" NO -> Was his team down by two or fewer runs when he entered the game? YES -> Did he allow the lead to exceed two runs at any point? | YES = "Letdown" | NO = "Stem" NO = "Mopup"Well, there you go. Have I taken it way too far? Is there any value in trying to categorize relief appearances like this? Or are we better off just calculating each appearances's WXRL and letting the numbers speak for themselves? I'm deliberately not drawing any conclusions in this article, but leaving it open-ended for your feedback and comment. Tell me what you think. As we figure out what the best approach is here, I'll run some numbers on these new definitions, and see what relievers come out on top. Then I will contrast the results with the current SV/HD/BS stat trio.
Trivia Answer: The two pitchers who choked away 3 big leads in a single year are Bob Stanley in 1984, and Rollie Fingers in 1979.