Ten things I didn’t know about bullpens

God bless Retrosheet. Baseball-Reference gets the most attention, but Retrosheet is an amazingly cool sight, that includes some stuff B-Ref doesn’t yet have.

For example, B-ref currently lists team splits going back to 1954. Meanwhile, Retrosheet has all that – plus 1952-53, 1945-48, and 1920-39 for all teams (plus the AL for 1940-41). Not bad, eh? There’s also the 1911 NL and some National Association info as well.

For me, this is a lot of fun, because I can update and expand my Tendencies Database, which was something I created for my new book: Evaluating Baseball’s Managers, 1876-2008 (which recently won rave reviews from Baseball America and Yahoo’s Big League Stew blog).

Lately, I’ve added all the splits available from 1920-onward (there’s too much fragmentary stuff in 1911, and I don’t care about the NA): info on 1,810 teams in all. Right now I just want to share some of the most interesting things I’ve discovered about the bullpen stats I have on hand. So here it goes:

1. Best bullpen EVER: 2003 Dodgers

While I dumped in bullpen split stats, I also added in park factors, and league ERA, allowing me to determine (by my own rough mathematical approximation) the squads with the best bullpens ever, as rated by ERA+. They are:

Year	Team	ERA+
2003	LAD	164
1981	NYY	159
1990	OAK	158
2002	ATL	158
1995	STL	156
1926	A's	155

(I put the sixth team on because I love the fact that such an old bullpen holds up so well).

Admittedly, there are problems with using ERA+ to gauge bullpen quality. Most notably it overlooks inherited runners. But check out that gap!

This team is most famous for Eric Gagne in full bloom. He converted all 55 save opportunities given to him while compiling a 1.20 ERA. He fanned 137 batters in 82.1 innings and allowed only 37 hits and 20 walks. Yeah, he was good.

He was good, but his middle relievers were almost as good. Paul Quantrill and Guillermo Mota both had ERAs under 2.00 while combining to throw more than 180 innings. Paul Shuey and Tom Martin also pitched well, though nothing as superhuman as the bullpen core.

The entire bullpen had the best strikeout rate and walk rate of the season, en route to a 2.46 ERA overall. Even without adjusting for park and era, that’s the 15th best bullpen ERA on record.

If anyone’s curious, the best unadjusted bullpen ERA is a tie between the 1967 White Sox and 1968 Dodgers at 2.14. The next lowest is the 1972 Pirates at 2.25.

2. Worst bullpens ever

Every yin must have its yang, so let’s look at the opposite end. What are the worst bullpens ever? Without looking, my guess would be the 1930 Phillies. Their bullpen ERA was 8.15, the only one on retrosheet that’s higher than 8.00. In fact, it’s the only one higher than 7.00. OK, that’s bad. Added bonus: throw in unearned runs and they allowed more than a run per inning. Yikes.

Still, adjust for park and era (both of which kill the 1930 Phils), and someone else lands on bottom: it’s a tie between the 1920 Cubs and 1938 Pirates, each with an ERA+ of 60.

That Pirate team is a fairly (in)famous squad. They paced their season like they did their ballgames: blowing a seven-game lead in the season’s final month. The flop culminated in Gabby Hartnett‘s famous “Homer in the Gloamin’” which the Hall of Famer wrapped up a comeback win for the Cubs with a walk-off homer in the ninth.

Naturally, Hartnett’s homer was off a reliever. It was a game-winning homer because the Pirates pen allowed a pair of runs the previous inning. Yeah, those relievers weren’t very good.

Still, it’s not entirely fair to call those teams the worst bullpens ever. The 1920 Cubs relievers tossed only 202 innings and the 1938 Pirates were well under 300. The worst bullpen with more than 400 innings was the 1953 Tigers (66 ERA+ in 402.1 IP).

Working upward, the 1967 Astros had an ERA+ of 70 in 427.1 innings. The 1955 A’s threw 482 reliever innings, with an ERA+ of 72. That was their first season in Kansas City – a helluva way to get started. The 2007 Devil Rays had a 75 ERA+ in 497 innings. That’s the worst bullpen of recent times.

Depending on how you weigh innings with ineptitude, one of the above is the worst ever.

3. The 1964 Kansas City A’s bomb squad

Actually, the reference to the 1955 A’s bullpen brings up the strangest finding I came across.

If I were a betting man, I would’ve wagered a pretty penny that the relief unit that allowed the most homers came from recent times. After all, modern baseball has seen an increase in bullpen usage and a massive surge in power.

Good thing for me I’m not a betting man, then, because the most homer-prone relief unit of them all came during the height of Beatlemania – and it’s not even close. Here’s the “leader” board:

Year	Team	HR
1964	KCA	92
1996	DET	88
2006	BAL	86
2004	CIN	85
2000	KCR	83

Until 1995, they had a 18-home run lead over second place (the 1982 Twins, at 74). Prior to 1964, the known record by a bullpen was 65 – by the 1955 Kansas City A’s.

Improbably, the 1964 A’s do not have the worst HR/IP. (That’s the 2001 Rockies.) The old KC gang is tied for 10th worst at 1.41HR/9IP. That brings up their second memorable feature: They threw an insanely large number of innings for their era.

In fact, based on the info in my database, the following teams set the record for most IP by a bullpen, from 1920-onward. (I’ll include managers because I care about that stuff):

Year	Team	IP		Manager
1920	STL	365.1		Rickey
1921	STL	382.2		Rickey
1924	PHI	394.2		Fletcher
1928	PHI	418.2		Shotton
1931	CWS	443.0		Bush
1935	STB	487.1		Hornsby
1956	WAS	505.0		Dressen
1957	NYG	508.2		Rigney
1962	LAA	528.1		Rigney
1964	KCA	588.1		McGaha
1977	SEA	599.1		Johnson, Drl
2003	TEX	601.1		Showalter

You know how modern bullpens always throw so many more innings than older ones? Well, the 1964 A’s are still sixth highest on the all-time bullpen IP board. They were third highest until 2003.

4. Bullpen IP going backward

Let’s flip around the chart from above: Looking backward, what’s the fewest innings thrown by any bullpen in history? We’re going to flip history on its head here, as if 2009 was the beginning and 1920 was the end, and act as though MLB history has been the story of ever-declining relievers. Again I’ll include managers, to see who has the least bullpen interest here, compared to his peers:

Year	Team	IP	Manager
2009	STL	437	LaRussa
2008	TOR	425	Gaston
2007	CHW	424.2	Guillen
2006	CHW	407	Guillen
2005	STL	397.2	LaRussa
2003	NYY	396	Torre
2002	ARI	387.1	Brenly
1998	ATL	364	Cox
1988	TEX	331.1	Valentine
1980	OAK	210.1	Martin
1946	DET	184.1	O'Neill
1941	CHW	167	Dykes
1923	NYY	160.1	Huggins
1922	NYY	159.1	Huggins
1920	CHW	159	Gleason

The above only looks at full seasons. That Billy Martin sure was something, wasn’t he? There’s a gap in the record, but that only affects three years between the 1980 A’s and 1946 Tigers.

5. Dick Radatz: an appreciation

Here’s another, similar list: best bullpen K/IP rate, by decade. (Look at it by decade because K-rates have gone up over the years.) Keep in mind that this info is fragmentary for the 1950s and especially 1940s:

Decade	Year	Team	K/9IP
1920s	1927	WAS	4.26
1930s	1931	WAS	5.23
1940s	1947	NYY	5.16
1950s	1959	NYY	6.78
1960s	1964	BOX	8.32
1970s	1976	PHI	7.75
1980s	1986	TOR	8.09
1990s	1999	NYM	9.40
2000s	2001	CHC	9.83

One thing leaps out at me in this list: the 1964 Red Sox. In previous decades, no team even topped seven, and they topped eight. While, admittedly, Ks increased in the 1960s, they were still the only team over 8K/9IP until 1986. No team topped their 8.32 mark until the 1990 Reds (which themselves had a widely heralded “Nasty Boys” bullpen featuring Rob Dibble, Randy Myers, and Norm Charlton).

The Red Sox’s secret: Dick Radatz, who was aptly nicknamed “The Monster.” In 1964, he threw 157 innings – all in relief (!) – with 181 strikeouts. Not bad, eh? The year before he whiffed 162 in 132.1 innings.

I’ve seen the numbers before and knew he was great, but this reinforces it. In his prime, Radatz was about as dominating as the Nasty Boys trio put together. Not bad.

6. The bicentennial bullpen of brotherly love

As noted above, strikeout rates over time have changed over the years, so the best way to look at this is to take every bullpen, divide by league K/9IP, and see who has the fanningest of them all.

Doing so would actually aid teams in low-K environments (it’s easier to be well over 5K/9IP than over 7K/9IP, but no answer is perfect). My hunch would be that the 1964 Red Sox would win, but again I’m wrong. They weren’t even tops in 1964: the Reds were at 139, while the Red Sox were at 136. It turns out the AL as a whole fanned an unusually large number of batters in 1964.

This isn’t meant as a knock at Radatz, as his achievement was remarkable. The Reds had a deeper well of fireballing relievers, including Sammy Ellis (125 Ks in 122.1 IP), Billy McCool (87 Ks in 89.1 IP), and an aging Ryne Duren 39 Ks in 43.2 IP). None of whom were The Monster, but when the ‘pen is that deep you can live without him.

The point here is to laud the team with the greatest era-adjusted K-rate: the Bicentennial Phillies. Their relievers fanned 359 batters in 417 innings, a rate of 7.75 in a league that averaged less than 5K/9IP.

The 1976 Phillies bullpen possessed a strikeout rate 57 percent higher than the league average, which is the greatest mark ever. Admittedly, this approach favors older teams (it’s easier to be 57 percent over five than over seven), but that Phillies bullpen was tremendous.

They had three relief aces: Tug McGraw, Ron Reed, and Gene Garber. All posted more than 10 saves that year and each of whom had fine, long careers. Their fourth reliever was Wayne Twitchell, whose 1.75 ERA was best in the bullpen. Their fifth reliever, Ron Schueler, fanned 43 men in 49.2 innings with a 2.90 ERA. Yet he was still just the fifth man in the bullpen.

If anyone’s curious, the worst adjusted strikeout rate ever came from the 1927 White Sox bullpen. Despite pitching in a league that averaged only 2.78K/9IP, the South Siders could manage only 59 percent of the league rate. They fanned 43 batters in 236 innings – 1.64K/9IP. If it’s harder to top the 1976 Phillies in a high-K era, then it should be equally hard for a team in a low-K environment to have the worst adjusted rate. But those 1927 White Sox pulled it out anyway. That’s some sort of an achievement.

The 1927 Sox also are last in unadjusted K-rate, and second-worst (behind the 1930 White Sox) in total batters fanned.

7. Ground ball specialists? The early 1920s Senators

I can do the same thing for homers allowed as I did for strikeouts. In theory, at least. In reality, it doesn’t work because bullpens pitched fewer innings, so some teams happened to have three or four homers in a season, giving them marks three or four times better than the league average. There’s also a far greater park factor, which I can ignore in strikeouts.

Still, the 1920s Senators bullpens are beyond belief. They have the two best league-adjusted homer marks ever: the 1924 Senator relievers had a HR rate one-fourth as low as the league rate, and the 1921 Senators were nearly as impressive.

A lot of that is surely park, but that doesn’t explain it all. In 1921, the Senators allowed 51 homers total: 48 by the starters and three by the relievers. How the hell do you do that? Starters allowed a gopher ball once every 23 innings but their relievers did it less than once in 80 innings. It’s like a bullpen of Greg Mintons. In 1924, the bullpen allowed a homer once ever 104 innings (!?!?) while the starters were at once in 35 innings. It’s just craziness.

If anyone’s curious, the bullpen with the best league adjusted walk-rate was the 1984 Royals. That’s not surprising given that Dan Quisenberry was there. The 1929 Tigers were the worst. They’re also the only bullpen to average more than six walks per nine innings.

8. The virtue of consistency

This one isn’t terribly important, but I sure found it interesting when Retrosheet dumped the 1945-48 info up last week. Total relief appearances by the Boston Braves in the mid-1940s:

Year    RA
1945	166
1946	167
1947	167
1948	167

They were to relief appearances what Albert Pujols used to be to at-bats in a season.

9. The pride of Anaheim

Here’s another take: as long as I’ve looked at park and league-adjusted bullpen ERA, I can see what teams had extended stretches of quality relief work. Several franchises have had impressive stretches, but the one that I found most striking was the Angels in recent decades (yes, decades):

Year	ERA+
1977	109
1978	125
1979	 82
1980	109
1981	112
1982	119
1983	 96
1984	104
1985	121
1986	107
1987	119
1988	104
1989	132
1990	108
1991	121
1992	101
1993	102
1994	 90
1995	129
1996	102
1997	129
1998	120
1999	121
2000	120
2001	134
2002	146
2003	135
2004	132
2005	119
2006	116
2007	102
2008	121
2009	101

Admittedly, this is comparing apples and oranges. League-wide bullpen ERA is routinely better than starter ERA, so a score of 100 isn’t quite league average. That said, if a bullpen is only slightly over 100, they ain’t off by much. A unit that is only a few/couple points below perfectly average is an about-average bullpen.

This is a long way of saying the Angels haven’t had back-to-back really bad bullpens in over 30 years. And they’ve had a helluva lot of really good bullpens, too. They had 10 straight years over 120. Neat trick. They’ve only had a pair of truly lousy bullpens: 1979 and 1994. They’ve been 96 or higher in the remaining 31 campaigns over the last 33 seasons, with 30 seasons in triple digits.

For perspective, from 1977-2009, there have been 916 bullpens. About seven-tenths (645) are in triple digits. The Angels have impressively continued to produce good to great bullpen performances.

10. The best on the worst and the worst on the best

Last query (well, queries, really): what’s the worst bullpen on a good team or best bullpen on a bad team.

Let’s start with the second part. The 1995 Cardinals had the best bullpen ERA+ by any team with a losing record. In fact, their 156 ERA+ makes the leaderboard at the top of this article for the fifth-best bullpen ERA+ ever.

Then again, they were “only” 62-81. What’s the best bullpen on a truly dreadful team? The best bullpen on a sub-.400 squad was the 1977 A’s: a 129 ERA+ for a 63-98 team. Among 100-game losers, the standout was the 1964 Senators bullpen, which posted a 122 ERA+.

Looking at the other end, the 1923 Indians had a bullpen ERA+ of 71 while going 82-71 on the year. In more recent times, the 1971 Yankees and 1980 A’s barely snuck over .500 despite bullpen ERA+s of 72. Hey – there’s the 1980 A’s! That’s famous as the starting staff Billy Martin beat to death. Halfway across this article we saw that every full-season team since 1980 has had their relievers throw at least 100 innings more than the 1980 A’s. I think we just discovered one reason Martin leaned so hard on his starters: His relievers were horrible.

Among great teams (.600 percentage or better), the worst reliever ERA+ was the 1933 Yanks, at 75. In the last half-century, the worst bullpen on a great team was the 977 Orioles, who had an ERA+ of 83 from their relievers.

That was fun. Schedule permitting, maybe I’ll have something similar for starting rotations next time.

References & Resources
I should note, my ERA+ calculations might be a tad off as I didn’t bother looking up the official calculations. I divided league ERA by team ERA and then adjusted for park factor. One other oddity for recent teams: about a year or two ago, Sean Forman noted on his blog that he had set up recent park factors (1996-on?) a little differently than previous years, but has decided to correct that. I’ve got the old park factors in my database. I could fix it, I guess, but I’m a little iffy on park factors. They’re valuable and useful (so I use them), but I think their precision is a bit overrated. Plus it’s a pain for me to add in the new park factors, there’s that, too.

Print Friendly
 Share on Facebook0Tweet about this on Twitter0Share on Google+0Share on Reddit0Email this to someone
« Previous: Yet another pitching metric
Next: Clone Wars: Jonathon Papelbon and Daniel Bard »

Comments

  1. kevin said...

    Hey Professor, the word in the first sentence should be “site,” not “sight.”  And what’s with the unnecessary comma after the incorrect word?

  2. Chris J. said...

    Kevin,

    Good catches. I have some sort of brain lock on site/sight. 

    It’s the second sentence, by the way.  There’s a period after Retrosheet.

  3. Gilbert said...

    Since we have editors out today, fix the year on “the worst bullpen on a great team was the 977 Orioles”.

    I couldn’t figure out how to get number of reliever appearances (rather than innings) per team out of retrosheet or bb-ref to see whether TLR was low on innings but high on pitcher swaps.

  4. Northern Rebel said...

    This is why Baseball is the greatest of all sports.

    Nobody cares what the record FG pct of a twelvth man on a conference champion is.

  5. Johnny said...

    Gagne was an outstanding pitcher in his prime, too bad he was ‘roid aided. That ‘03 Dodgers team had a devastating 7/8/9 with Mota/Shuey, Quantrill, then Gagne.

  6. MikeS said...

    Interesting that Tony LaRussa, a man accused of overhandling his bullpen by making too many switches and slowing the game down, has used his bullpen less than any other manager twice in the last 6 years.  Maybe the fact that he uses his starteres more heavily than most allows him to use relievers for shorter stretches.

    Also, as a White Sox fan, I can tell you Guillen did the right thing in ‘06 and ‘07 keeping the relievers down in the bullpen and off the mound.

  7. Cliff Blau said...

    Chris,
    In #2, you don’t mean the 1938 Pirates.  Their bullpen had an ERA of 3.09 vs. their starters’ ERA of 3.62.  Maybe the 1938 Phillies?

  8. Chris J. said...

    Wha? ? ?

    (checks)

    Yipes!  – You’re right (as usual), Cliff.  You even have the correct answer: the 1938 Philies had the ERA+ of 60.  Pittsburgh had a good bullpen.  Jeez, that’s a howler.  Far worse than any missplaced comma . .

    The 1938 Pirates, far from having the worst bullpen ever, had the fourth best bullpen in the league taht year.  Actually,it was a strange year for bullpens.  Four teams had bullpens over 120, and the other four were 94 or lower.  It was feast or famine.

    THanks again for pointing that out to me.

  9. Chris J. said...

    mccombe35,

    I have them at 138.  (Note: these ERA+ numbers ain’t official.  I didn’t check b-ref’s formula when I did it.

  10. Gen3blue said...

    Don’t mean to pile on but I can’t find the Angels streak of 10 years over 120 (ERA+). Interesting stuff.

  11. mccombe35 said...

    OK Thanks

    Seems like it should be better with 3 of their main bullpen guys (Hermanson, Politte, & Cotts)all had an ERA+ of 220 or better & combined for over 180 IP.

  12. Chris J. said...

    Without looking too closely, I guess the 40 ER in 50.1 IP by the bottom of the ‘pen is what hurt them so much.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>