Ten things I didn’t know about starting rotations

Two weeks ago I wrote a column here called “Ten things I didn’t know about bullpens” in which I took the splits info at Retrosheet to examine the best and worst of relief units of all time. Since I can just as easily study the starting pitchers, it makes sense to look at that as well.

Both this column and the previous one are outgrowths of work on my book, Evaluating Baseball’s Managers, since I first started collecting info from Retrosheet to gain an idea about managerial preferences. These articles have one added feature: Because Retrosheet keeps updating, I can draw on a wider well of material from there.

In fact, since the bullpen column, Retrosheet has updated yet again, giving me splits info for the AL from 1942-44. I now have everything from 1920-onward the except the NL from 1940-44 and both leagues from 1949-51. That’s it in the lively ball era. I have info for 1,834 different starting rotations.

That’s fragmentary info, but a heckuva fragment! It’s over 95 percent of all lively ball rotations, and almost three-fourths of all teams since 1876. The following statements are based on that supersized fragment.

1. The best rotation of them all

Let’s hop to it. According to ERA+ (and I’m adjusting for park and league myself here), the best starting rotations that I know of are as follows:

Year	Team	ERA+
1998	ATL	139
1997	ATL	138
1931	PHA	138
1981	HOU	135
1926	PHA	135
1993	ATL	134
1922	STB	134

Well, I guess it shouldn’t be too surprising that the 1990s Braves do so well. After all, they had Greg Maddux and Tom Glavine, and John Smoltz. It still is s bit surprising because as great as they were, there have been a whole herd of other fantastic rotations over the years. For that reason alone, it should be tough for one unit to dominate the top slots.

From the above list, the real stunner—to the point where I double-checked the results—is the 1922 Browns. They had an overall team ERA+ of 123, but did much better as starters. Actually, looking at personal ERA+s at Baseball-Reference.com disguises their success because they almost all pitched out of the bullpen at times. Ace Urban Shocker posted a 1.88 ERA as starter, but “only” 2.45 in relief. The No. 2 pitcher, Elam Vangilder, had a 3.39 ERA as starter and 3.71 in relief. Others posted better marks in relief than as starters, but Shocker and Vangilder bore the main loads as starters.

Way back when, SG of the Replacement Level Yankee Weblog ran some computer sims for me showing the 1922 Browns were one of the best teams to miss the World Series. Well, this shows one reason why they were so good.

The 1981 Astros had a 2.43 ERA, which is the best ever. Among teams pitching full seasons, the Bob Gibson-led 1968 Cards top them all at 2.49.

2. 1990s Braves: an appreciation

As impressive as the Braves’ result is, it actually underestimates how dominant they were. Look, I know saying the 1990s Braves had great pitching is hardly original, but the more I look at the results, the more amazed I am.

They pulled off three of the top seven marks (and three of the top six from full seasons) despite pitching in an era when starter ERA+ on the whole has declined. Since the bullpen has become an established part of the game, relievers lodge better ERAs than starters. Back when the 1931 A’s played, a starting rotation ERA+ of 100 would actually be below average, now it’s above average.

In the last 50 years, only seven rotations had an ERA+ of 130 or better. Four were by the Braves. In the last 22 seasons, 20 rotations have an ERA+ of 120 or higher. Bobby Cox managed half of them.

You know those great 1970s Orioles rotations? Well, they maxed out with an ERA+ of 119 in 1972. That’s a hell of a mark—in fact it’s the best mark by any AL rotation in the 1970s. But Cox’s Braves topped it 10 times from 1992-2005.

Here are the best rotations of the 1990s:

Year	Team	ERA+
1998	ATL	139
1997	ATL	138
1993	ATL	134
1995	ATL	130
1994	ATL	129
1996	ATL	127
1992	ATL	126
1999	BOS	125

How ’bout them Red Sox!

If you’re curious, the 1999 Braves had an ERA+ of 118, and the 1991 squad was at 113. The best non-Atlanta staff was not quite as good as an average Atlanta staff—for an entire freaking decade! Dear lord.

3. Best by decade

Let’s look at it another way: What were the best starting rotations in each decade? Let’s see (and remember: info is fragmentary for the 1940s and 1950s). In case of ties I’ll list both teams from the decade:

Dec	Year	Team	ERA+
1920s	1926	PHA	135
1930s	1931	PHA	138
1940s	1948	CLE	133
1950s	1957	BRK	127
1960s	1964	CWS	130
1960s	1969	NYM	130
1970s	1970	CHC	127
1980s	1981	HOU	135
1990s	1998	ATL	139
2000s	2003	MON	130

That 1971 Cubs team was the only rotation in the 1970s over 123. The highest full season of the 1980s was 129 by the 1985 Dodgers. That was one of only two times all decade a full-season performance topped 122.

This makes Atlanta’s accomplishment in the 1990s that much more impressive. It wasn’t like the Braves competed against unusually middling competition: 125 had been a seemingly unbreakable ERA+ barrier for a generation, and they took up permanent residence there.

I never would’ve expected to see the 1957 Dodgers on top of the decade (they narrowly beat out the 1954 Indians), but when you look at their roster, it makes sense. All their pitchers had great years.

4. Worst ever

Let’s briefly take a gander at the other side: What are the worst rotations on record?

Year	Team	ERA+
2005	KCR	72
1984	SFG	73
1964	KCA	74
1977	SDP	74
1939	PHI	74
2003	CIN	74

Hey look: It’s Lima time!

On the Royals in 2005, Jose Lima allowed 219 hits (including 31 homers) in 168.2 innings while walking 61 batters and whiffing only 80. As a reward for this horrible pitching, manager Buddy Bell decided to leave him in the rotation all year. Yeah, the Royals didn’t have many quality pitchers that year, but Bell benched guys doing worse. In fact, Lima tied a record that year for lowest ERA+ (63) for any pitcher in MLB history with more than 30 starts.

It’s also fun seeing the 1964 A’s up there. I mentioned them in my column on bullpens because their relievers still hold the record for most homers allowed in a season. One reason they achieved it: They threw so many innings because the starters were horrible.

The 1984 Giants had four men start more than 10 games. They had ERA+s of 81, 77, 77 and 66. Now that’s impressive.

The “award” for worst unadjusted ERA goes to the 1996 Tigers at 6.64. No one else is worse than 6.34. Those Tigers were managed by Buddy Bell in his first season as manager. Yeah, he had bad staffs, but I can’t help thinking he had no idea what to do with them.

5. Control kings: 2000 Twins

Here’s something I did in the bullpen article that works here as well: Take the rate stats (walks, homers and strikeouts per inning) and adjust them by league wide rates. After all, throughout its history both the way the game has been played and its environment has changed, causing those things to drop and/or rise. Adjusting for league rate allows us to better look at teams across eras.

To figure the best control staff, I divided league BB/9 by a team’s rate (that way higher equals better, and lower worse). Based on that: here are the best control staffs ever:

Year	Team	BB+
2005	MIN	200
2000	ATL	169
1979	MIL	167
2003	NYY	161
2002	ARI	160

I can only assume there’s some sort of era bias causing really recent teams to dominate the list. More importantly, one team breaks from the field by a downright comical extent.

Like the 2005 Royals, the 2005 Twins had a pitcher who embodied their style on the mound: Carlos Silva. In 188.1 innings, despite fanning only 71 batters, he walked only nine guys. NINE. I have no idea how he did it. Its one thing to rely on your defense but batting practice pitchers have higher walk and strikeout rates. He didn’t fool anyone, but still refused to move off the plate.

By this approach the worst starting rotation was the 1971 Indians. Steve Dunning walked 109 guys on that team while staff ace Sudden Sam McDowell issued 153 free passes. The team as a whole issued 303 more walks than its batters attained, the worst walk differential by any team since World War I. The team traded McDowell in the offseason.

6. The 1946 AL flamethrowers

Apply the same approach given above to strikeouts, and the best staff of flamethrowers ever was the 1924 Dodgers. That doesn’t surprise me in the least. They had Dazzy Vance, who was personally responsible for every 13th strikeout in the league that year. Yes, you read that sentence correctly. He fanned 262 batters in a league where 3,408 Ks occurred.

For me, the memorable part was learning about the 1946 Tigers, who had the second best era-adjusted K-rate. That rotation recorded 5.89 strikeouts per nine innings. No previous squad I have information for topped 5.00K/9.

That was Hal Newhouser in his prime. On my list of articles to do is a piece about Newhouser, and 1946 will have a prominent place in it. As great as his end-of-the-year numbers look: 26-9 record with a 1.94 ERA and 275 strikeouts (please note he led the league in Ks the year before with only 212), they arguably minimize his accomplishments. He won his 20th game on July 27, when Detroit was 52-39. No pitcher since, not even Denny McLain, has made it to 20 in so few games for his team.

Beyond Newhouser, teammate Virgil Trucks was fourth in the league in strikeouts. Dizzy Trout was fifth, and Fred Hutchinson tied for sixth. It’s not too often that four of the top seven strikeout artists are on the same team.

Even still, this tidbit says “1946 AL flamethrowers” not 1946 Detroit flamethrowers. Remember how I said no known rotation before 1946 topped 5.00K/9? Well, another team did that year: the Indians, at 5.54. They were the only teams until the 1950s to break five, and they both shattered it.

It was less of a collaborative effort with the Indians. That was the year of Bob Feller. He threw an insane 371.1 innings, and whiffed 348 batters. He won his 20th game of the year in Cleveland’s 100th game (the Indians were 47-52 with a tie). Denny McLain didn’t win his 20th game until Detroit’s 101st contest in 1968 (though like the 1946 Indians, those Tigers had a tie).

7. Across the board brilliant

Only twice has a starting rotation topped all others in league in park-adjusted ERA, BB/9 and K/9: the 1996 Braves and 1928 A’s.

I suppose if it’s going to be any squad, it would be the 1990s Braves and Mack’s second coming in Philadelphia. After all, those were the two teams that made the best ERA+ list multiple times back at the top of this article.

Still, neither one of these teams were listed in that bunch. That’s what happens when you have a great corps of pitchers: It’s hard to spot what their most impressive season ever was.

(Quick qualifier: I haven’t yet ranked all the 1940s teams by their walk and strikeout rates—it’s hard to keep up with Retrosheet these days—but odds are no third team from then will enter the mix.)

8. 1930s St. Louis Browns: a non-appreciation

Six times a starting rotation has finished last in the league in park-adjusted ERA, as well as both walk and strikeout rates: 1932 Red Sox, 1936 A’s, 1937 Browns, 1952 Pirates, 1982 A’s and 2006 Royals.

The Browns are the staff I find most interesting. Not only were they last in all the above categories, they also allowed 111 homers, the most by any starting rotation up to that point in time. As an added bonus, the Browns kept up their horrible pace for the next several years.

In 1938, their starters walked 590 batters, over 50 more than the next highest team. I don’t mean the next highest team in 1938,; I mean of the 1,800+ starting rotations I have this info for.

In 1939 they posted an ERA of 6.34. Only Buddy Bell’s 1996 Tigers have ever had a worse starter ERA. Speaking of Bell, please note he managed the 2006 Royals, who were last in BB/9, K/9, and ERA+.

9. Homers allowed over time

The last point makes me wonder: How has the record for most home runs allowed in a season by a starting rotation changed over time? Well, with the continual qualifier that I don’t have some info from the mid-century, here are the record setting (or record tying) staffs:

Year	Team	HRA
1920	NYY	47
1921	NYG	79
1922	PHI	89
1929	PHI	90
1930	PHI	107
1934	CWS	108
1937	STB	111
1938	PHA	119
1953	CIN	134
1956	PHI	134
1961	CLE	134
1962	KCA	143
1963	DET	146
1986	MIN	160
1987	CAL	169

I’m rather impressed with the 1987 Angels. I remember 1987 being one of the game’s occasional Silly Ball years where homers flew out with abandon, but I would’ve guessed that a more recent season would hold the record.

One could argue the Angels’ continued dominance came because current staffs throw so few innings, but that doesn’t work. The Angels also hold the record for most HR/9 at 1.60. The runner-up is Buddy Bell’s 1996 Tigers, at 1.59. Man, you can’t swing a dead cat without running into a historically inept pitching staff he managed.

10. Homers going backwards

Let’s flip around the above list: What’s the fewest homers allowed by a rotation starting with 2009 and standing the calendar on its head? Since starters often throw fewer innings now, I’ll look at HR/9:

Year	Team	HR/9IP
2009	STL	0.72
2008	TOR	0.71
2005	FLO	0.68
1997	ATL	0.64
1993	LAD	0.63
1992	ATL	0.45
1984	LAD	0.43
1981	HOU	0.33
1945	WAS	0.27
1944	DET	0.22
1927	CIN	0.20
1923	CIN	0.17
1920	PIT	0.16

More appearances by Bobby Cox teams. He is the yin to Buddy Bell’s yang.

If you’re wondering, the list doesn’t change too much if you toss out the strike-shortened 1981 Astros. Then the 1984 Dodgers lead holds until 1976, when the Giants averaged 0.42HR/9. That was the fewest (that I know of) in a season since 1946, when the Pirates were at 0.36.

No pitcher on the 1976 Giants allowed a dozen homers that year. Meanwhile, MLB’s other teams featured 72 men who allowed at least that many. Their starting pitchers allowed 45 homers, 10 fewer than any other team in the entire 1970s. Aside from the 1981 season, that’s the fewest by any team that I know of in the last 60 years.

References & Resources
Thank you, Retrosheet!