Leveraging Across the Eras (SP Lev., Part 8)

Welcome back to this series of articles on starting pitcher leverage. Normally here’s the part where I cut and paste the same opening paragraph I’ve been using for a while to explain what I’m doing. But this is article #8. I don’t have that many new readers. For those that are new, read the beginning of any of these articles. For the math behind it, just scroll to the bottom of this article.

The Question

To date, I’ve generally looked at the starting pitcher leverage era, which lasted from the earliest days of baseball until the 1960s, as one big monolithic block. Oh sure, I might make a passing mention to the fact that leveraging had its ups and downs within that nearly century long period, but I’ve never seriously looked at when leveraging peaked or when it reached its low ebb in those years. I’m going to fill that gap by finding out the whens (and hopefully whys) of the rises and falls of leveraging.

Fortunately, figuring this out is mercifully easy. The last article introduced the notion of LPA—Leverage Points Average, a very simple concept. The further a pitcher’s AOWP+ is from 100, the more Leverage Points it’s worth. Figure it out for all the pitchers with at least 20 starts in a season, and that’s the LPA for that year. (The bottom of the page has the exact formula for LPA and AOWP+). I’ve already figured out AOWP+ for a slewload of guys, so this should be pretty easy.

The Peak of Leveraging

Before giving the highest single-season LPAs of all time, please remember this isn’t based on a full sample size but “only”182,000 starts, including two-thirds of all GS from 1876 to 1969. Also, I’m tossing out the years 1871-5 and 1884 because the leagues from those seasons were so poorly organized that their AOWP+s are tainted.

Qualifications aside, here’s the highest single-season LPAs, including the number of pitchers with at least 20 GS AOWPd from that year:

    Year       #SP   LPA
1.  1954       46    7.17
2.  1908       40    5.73
3.  1939       47    5.60
4.  1953       43    5.58
5.  1942       41    5.51
6.  1880       10    5.50
7.  1910       40    5.48
8.  1879       7     5.43
9.  1955       49    5.39
10. 1909       40    5.30
11. 1946       42    5.10
12. 1890       47    5.09
13. 1919       43    4.98
14. 1886       33    4.88
15. 1889       41    4.88
16. 1907       50    4.86
17. 1936       42    4.76
18. 1937       50    4.70
19. 1949       49    4.49
20. 1887       35    4.43

First, there are some patterns there. This list represents every decade from the 1880s to 1950s except the 1920s. Interesting. Some eras really dominate. Four straight years from 1907-10 make the list, as do three straight from 1953-5. Meanwhile, three out of four seasons make it from both 1936-9 and 1886-90. Lastly, in the early days of 1879-80 also make it. I’m not sure if that part means anything because their sample size is so small that one AOWP+ can make an important impact, and both of those 1879 and 1880 had one of the lowest single-season AOWP+s of all time.

Second, 1954 blows the competition away. They had 15 pitchers with an AOWP+ of either 106 or higher, or 94 or lower.

Name                AOWP+
Whitey Ford          112
Eddie Lopat          109
Murray Dickson       109
Billy Pierce         108
Early Wynn           108
Johnny Klippstein    107
Mickey McDermott     107
Vic Raschi           107
Billy Hoeft          106
Frank Sullivan       106
Gene Conley          106
Bob Porterfield       94
Mike Garcia           94
Paul Minner           94
Harry Byrd            92

For perspective, in the last 40 years I have zero AOWP+s higher than 105, and only Andy Pettitte’s 2002 had a 94. I don’t entirely know why it’s that much greater than any other season. I imagine it’s a combination of flukerifical aligning of the stars in conjuncture with the fact that the early-to-mid 1950s were a golden age for leveraging.

Looking at five year blocks is another way to get a sense of the ups and downs of leveraging would. It also has the advantage of smoothing out any oddities like 1958’s insanely high score. When looking at five year chunks, however, it turns out the 1950s really were the all-time peak for leveraging. The best five-year stretch was from 1952-6, with 1951-5 right after it. That’s surprising. I would have guessed the Deadball era would have topped all else. It was easier to rejigger the rotation back then than any other time because pitchers could take it easier more often. The three highest single-season AOWP+s of all-time happened then. Plus, you had more extreme differences in team winning percentages then, which should make AOWP+s more extreme.

So why the 1950s? My hunch is that Casey Stengel caused it. Last article showed that Stengel was an all-time great leverager. And he won titles every damn year. I’m sure others aped him. Frankly, that’s all I can imagine. It also means that leveraging died out completely barely a decade after its golden era. Amazing.

The Deadball era, especially 1907-12, was leveraging’s silver era. Aside from reasons already given, that was the start of the 154 game schedule. (Yes, I know there were two non-consecutive 154 game schedules in the nineteenth century. This was when it stuck, though). Facing opposing teams 8-10 series a year gave teams more opportunities for leveraging.

Fighting for the bronze it’s essentially a toss up between the late 1880s and late 1930s. In 1886, both the NL and Double-A expanded their schedules. The increased number of series against rivals led to higher AOWP+s. It dropped down when baseball became one 12-team league in the 1890s. You didn’t see opposing teams as often, causing, the late 1880s to be the 19th century peak. Platoon leveraging, a central component to leveraging, reached a crest in the late 1930s, causing the other bump in AOWP+.

Leveraging Lulls

I already know when leveraging bottomed out: 1966-present. 1965 had an LPA of 2.66. Next year it fell to 1.64. It’s been like that ever since. There’s no real leveraging in any of those years. Lows are mostly random happenstance caused by the much smaller percentage of games AOWPd, and happenstance. I’d rather look at leveraging’s years when leveraging was dormant, not dead. In that spirit here are the lowest L#s from 1876-1966. For perspective, I’ll toss in the last 40 years’ cumulative LPA:

    Year         #SP     LPA
1.  1958          47    1.40
2.  1923          46    1.65
3.  1900          29    1.76
4.  1944          40    1.78
   1966-2005    1285    1.80
5.  1959          49    1.94
6.  1914          64    2.30
7.  1943          41    2.37
8.  1918          37    2.38
9.  1922          48    2.46
10. 1926          50    2.62
11. 1965          65    2.66
12. 1917          51    2.67
13. 1902          46    2.67
14. 1896          26    2.69
15. 1963          63    2.71
16. 1898          34    2.74
17. 1945          35    2.74
18. 1940          45    2.76
19. 1920          56    2.79
20. 1928          51    2.92

Again, some eras dominate. There are several years from around the turn of the century, a slew around 1920, three straight in WWII, and several from final eligible years including 1965. The real story might be that damn near every single season scored higher than modern standards. With rare exceptions, it’s never even close.

Most amazingly, just four years after the greatest season ever for leveraging, you have the worst season until the complete end of leveraging. Here’s a comparison of the AOWP+s for individual pitchers from those years:

1954      AOWP+      1958
15      106+/94-        0
4        105/95         1
6        104/96         1
5        103/97         4
4        102/98         3
8        101/99        21
4          100         17

The top row refers to AOWP+s of 106 or higher, and 94 or lower. It looks like leveraging died in four years. But I’ll be damned, it came back in 1960-2, fell down in 1963, revived in 1964, before declining again in 1965, then falling completely. Leveraging’s death throes was not a neat, orderly process, but 1958 was the beginning of the end. The reasons for its death frankly deserve an article for themselves (and that’ll come later on in this series). The most important fact lies with the rise of a new pitching philosophy: standardized roles.

A Hardball Times Update
Goodbye for now.

Baseball for about a century managers fluidly used their starting pitchers. Rather than lock themselves into a fixed ABCDABCD rotation, they constantly reworked their rotation, even when everyone was healthy and effective. Also, virtually all pitchers were swingmen. As late as the 1960 AL, all pitchers with at least 20 starts had at least 1 relief appearance. That was common back then. Now 75% have zero relief appearances.

Around 1960, baseball teams began locking their pitchers into a fixed ABCDABCD rotation and making a sharper distinction between starters and relievers. Fluidity gave way to rigidity. Leveraging depended on that lost fluidity and hence its decline came as a result of this change forty years ago.

Breaking it into five year bunches, each five-year period from 1962-1966 scores lower than every five-year period before it. No exceptions. Making the break even more impressive, in the last 40 years, the two highest periods are 1962-6 and 1963-7. Well yeah, that’s the transference period.

The next lowest five-year periods are 1920-4, 1922-6, and 1921-5. Clearly the leveraging didn’t roar in the 1920s. The key lay in platoon leveraging. As previous articles showed, platoon leveraging was a key component in overall leveraging. Well, the 1920s were the heydey of offensive leveraging. That would lead to a decline in pitcher platooning, which would hurt pitching leveraging altogether. Makes sense.

The Course of Leveraging

Looking back, you can see a sensible rise and fall emerge. Leveraging became big with the expanded schedules in the 1880s. It declined in the 1890s, as teams didn’t face each other as often in the 12-team NL. It picked up again rising to new heights in the Deadball era with the creation of the 154 schedules in eight-team leagues. It began to decline as offensive platooning caught on later in the century, falling to new lows in the 1920s. The decline in that offensive philosophy reignited pitcher leveraging in the 1930s, remaining fairly strong for decades. It came to a whole new prominence when Casey Stengel widely used this strategy while guiding the Yankees to five successive titles. Finally, in a back-and-forth, decade-long process standardization slayed leveraging. It has never risen from the grave.

But why take my word for it? I got the numbers, and there’s no fun in sitting on them. Here’s the number of AOWP+d pitchers for every year, and the LPA for each season from 1871-2005. Next article I’ll look at franchises and leveraging.

Year     #SP      LPA
1871      1          0
1872      3       1.33
1873      3       4.33
1874      4       0.25
1875      5       22.6
1876      5        2.2
1877      2          4
1878      3       3.33
1879      7       5.43
1880      10       5.5
1881      10       1.9
1882      16      2.13
1883      22      2.73
1884      28      5.46
1885      26      2.46
1886      33      4.88
1887      35      4.43
1888      35      3.46
1889      41      4.88
1890      47      5.09
1891      40       4.1
1892      36      3.08
1893      33      3.42
1894      29      4.34
1895      28      3.21
1896      26      2.69
1897      26      3.54
1898      34      2.74
1899      39      4.15
1900      29      1.76
1901      47      3.04
1902      46      2.67
1903      44      3.73
1904      50      3.62
1905      54      3.69
1906      56      3.73
1907      50      4.86
1908      40      5.73
1909      40       5.3
1910      40      5.48
1911      42      3.83
1912      42      4.38
1913      41      3.12
1914      64       2.3
1915      63       3.7
1916      48      3.48
1917      51      2.67
1918      37      2.38
1919      43      4.98
1920      56      2.79
1921      45      2.93
1922      48      2.46
1923      46      1.65
1924      45      3.44
1925      53      3.21
1926      50      2.62
1927      52      3.08
1928      51      2.92
1929      53      3.68
1930      53      4.06
1931      50      3.52
1932      56      3.55
1933      48      3.13
1934      46      3.93
1935      46      4.09
1936      42      4.76
1937      50       4.7
1938      46      4.22
1939      47       5.6
1940      45      3.18
1941      44      3.95
1942      41      5.51
1943      41      2.37
1944      40      1.78
1945      35      2.74
1946      42       5.1
1947      46      3.65
1948      51      4.41
1949      49      4.49
1950      48       4.4
1951      42      3.74
1952      45      4.02
1953      43      5.58
1954      46      7.17
1955      49      5.39
1956      45      3.84
1957      48      3.46
1958      47       1.4
1959      49      1.94
1960      51      3.67
1961      53      3.72
1962      58      3.12
1963      63      2.71
1964      58         3
1965      65      2.66
1966      64      1.64
1967      61      1.59
1968      64      1.22
1969      57      1.98
1970      54      2.63
1971      48      1.75
1972      41      2.56
1973      40      1.48
1974      35      1.31
1975      32      2.16
1976      31      1.58
1977      27      2.15
1978      30       1.6
1979      29      1.76
1980      30      1.63
1981      28      2.79
1982      32      2.41
1983      31      1.68
1984      31      0.87
1985      29      2.24
1986      36         1
1987      37       1.7
1988      32      1.41
1989      33      1.52
1990      32      1.47
1991      30      1.63
1992      33      1.85
1993      35         2
1994      25       2.2
1995      25       2.4
1996      25      1.48
1997      23      0.91
1998      23         2
1999      23      2.04
2000      16      1.81
2001      13      1.46
2002      15      3.27
2003      13      2.46
2004      11      1.82
2005      11      1.82

References & Resources
What the heck is AOWP+?: The stat I invented to judge pitcher leveraging. It’s AOWP/TOWP*100. AOWP is Average Opponent Winning Percentage. TOWP is Team’s (Average) Opponent Winning Percentage. To figure AOWP for a single season, you take the number of starts a given pitcher had against each opposing team, and multiply that by the team’s winning percentage. After doing this for all rival squads, add up the products and divide by the pitcher’s total GS. The result is his AOWP. The same logic applies to TOWP, only here you look at how many games the team played against all rivals. If a pitcher’s used evenly, his AOWP will be the same as the TOWP, and he’ll have an AOWP+ of 100. If he’s used more against better teams, he’ll have a higher AOWP+. I calculated AOWP+ for 659 pitchers who started 182,000 games, including over two-thirds of all games from 1876-1969.

What the heck is LPA? Leverage Points Added tells you how common leveraging was in a given subset. For example, when looking at how common leveraging was in a given year, you figure award a certain number of points to every AOWP+. An AOWP+ of 100 is worth 0 points. 99 & 101 are worth 1 point each. Then 102 & 98 are worth 3. Each successive integer AOWP+ moves away from 100 is worth an additional two points. So an AOWP+ of 110 or 90 is worth 19. LPA can show how common leveraging was for a year, for a manager, for a franchise, and various other subsets.


Comments are closed.