20+ decisions with 0 starts was done 49 times from 1960-83 but only 9 times in other years

Click here to see the list from Stathead.

Before 1960 it was done only 6 times with the first being Clint Brown of the White Sox in 1939. Before 1939, the most decisions for a guy with 0 starts was 14 in 1937 by the same guy, Clint Brown, and the famous reliever Firpo Marberry in 1925. So Brown's 1939 season was a big leap. 

Also, there were just 3 cases of exactly 19 in 154 game seasons. A 162 game season is about 5% longer than 154, so by lowering the cutoff to 19 things are being equalized. But it does not affect the numbers that much.

The first case of 10 decisions was Slim Sallee in 1921. But there were still just a total of 9 cases of 10+ decisions and 0 starts from 1921-1938.

The last pitcher to have 20+ decisions with 0 starts was Scott Shields in 2005. The last before that were Roger McDowell & Mark Eichhorn in 1986. 

Pitchers who did it more than once are

Bill Campbell (1976-77)

Mark Clear (1980, 1982)

Rollie Fingers (1972, 1976, 1980)

Gene Garber (1975, 1979)

Rich Gossage (1977-78)

Mike Marshall (1972-75, 1978) 

Dick Radatz (1963-65)

Kent Tekulve (1980, 1982)

Wilbur Wood (1969-70)

In 46 of the cases, the guy had 100+ IP with a high of 208.1 for Mike Marshall of the Dodgers in 1974 (15-12) to a low of 72 for Bob Grim of the Yankees in 1957 (12-8). Both teams came in first place and made it to the World Series.

Here are the number of occurrences over the years 1960-83

1960	1
1961	2
1963	3
1964	2
1965	3
1967	1
1969	3
1970	1
1971	1
1972	2
1973	1
1974	3
1975	3
1976	3
1977	3
1978	5
1979	3
1980	3
1982	4
1983	2

There were 29 cases from 1974-83, averaging 2.9 per year with each year having at least 2 cases. 

There has been only 1 case of 19 since 2005, 19 by Tyler Clippard in 2010. There have been just 1 case each of both 18 & 17 and they came last year, in 2021. Since 2005 there have been a total of 16 cases of 15+ or just about 1 per season.

So the 1960-83 period (especially 1974-83) is a big outlier.

But, if we drop the cutoff to 10+ decisions along with keeping the 0 start criterion, there have been 345 cases since 2006. That is way more than what we saw from 1921-38, when it was just 9. So things have scaled back but they still look much different than they once did. 

From 1960-83, there were 509 cases. But I dropped the 5 from 1981 due to the strike.So I divided 504 by 23 and got 21.91. For years since 2006, I excluded both this year and 2020 (because of Covid) and only 1 guy this year has reached 10. So I divided 344 by 15 and got 22.93.

That is just a bit more than the 21.91 for 1960-83. The latter period has had 30 teams every year while the earlier period at most had 26 and for most of it had 24 or less and was as low as 16.

That earlier period had about 22.7 teams per season, well below the 30 per season since 2006. That means that the latter period has about 32% more team seasons. If we raise up the 21.91 of the 1960-83 period by 32% we get 28.92, a much higher number. So even at an intermediate level, things are different now than from 1960-83.

Adam Cimber of the Blue Jays is the guy with 10+ this year. He is 8-3 so he has a shot at reaching 20 as long as he does not start any games.

Update July 22:

Here are the number of cases for zero starts and 14+ decisions in the decades before 1960

1930-39) 2 (both Clint Brown)

1940-49) 15

1950-59) 20

So this kind of thing was growing before 1960, but still not alot.  I used 14 decisions since it is 5 less than 19 and I used 19 for games before the 162 game schedule since it is 5% less than 20 and 154 is about 5% less than 162.

But if I use 15 decisions and zero starts we can see that things really explode in the 1960s.

1960-69) 55

Now there were more team seasons in the 1960s due to expansion (372). That is 2.325 times as many as the 1950s. If things had held steady, we would have seen 34.875 seasons of 15+ decisions and zero starts. Yet it was 55, which is 57.7% higher than 34.875 (and I used 15 decisions for the whole decade which is not quite right because the AL in 1960 and the NL in 1960-61 were still at 154 games, but that would only change things a bit, showing an even bigger % increase).