To this point in the weighted WAR series, we’ve introduced the weighting system, the schedule length adjustment, and the method for classifying players by position. Most recently, we looked at the top shortstops through the perspective of the timeline adjustment – in particular, the adjustments for segregation and minor league control.
In the timelining discussion, I mentioned performing a sanity check by examining the total number of top-100 players at each position who come from each decade. (More specifically, I classified each player by the season in which their median plate appearance occurred and broke those down by decade.) In general, the number of representatives from each decade increased over time, but there were two major exceptions to this. First, the 1940s trailed behind the ‘20s and ‘30s; this is easily explained by World War 2 cutting short the careers of many players from that decade.
The second anomaly was much more unexpected: the 1880s were significantly
ahead of the 1890s. After considering this discrepancy, I decided to implement
a third timeline adjustment, this time for league expansion.
Baseball fans are generally familiar with the history of
modern expansion, which began in 1961 after a lengthy period of the AL and NL
having eight teams each. Here is that history in table form:
|
Season |
Size change |
Final size |
|
1961 |
+2 |
18 |
|
1962 |
+2 |
20 |
|
1969 |
+4 |
24 |
|
1977 |
+2 |
26 |
|
1993 |
+2 |
28 |
|
1998 |
+2 |
30 |
In summary, the league expanded by 50% over the course of a decade, and then by a further 25% over the next 30 years. In isolation, that sounds like a large change, and the effects of such seasons on individual players are well-known, since 1961 and 1998 both saw the single-season home run record broken. Fans will frequently bring up expansion as a caveat in assessing player quality in the relevant years.
What does this have to do with the 1880s? To explain, let’s
briefly step through the history of 19th-century league structures.
The first openly professional baseball team was the 1869
Cincinnati Red Stockings. By 1871, enough other local clubs had gone pro that
they were able to form the first professional league, the National Association.
Assessing league size in the NA is trickier than you’d expect, because the
teams played wildly different numbers of league games; in 1872, the Baltimore
Canaries played a league-high 58 games, while the Washington Olympics managed just nine. Attempting to
account for teams that disappeared and were replaced during the season, my best estimate is that there
were effectively eight teams in the first four years of the NA, and 10 in the fifth
year.
Speaking of disappearances, the NA itself folded after 1875, replaced by the National League. The NL debuted in 1876 with eight teams, shrinking to six in 1877. The league re-grew to eight teams in 1879, and stayed at that number for over a decade, which would seem likely to make the discussion of expansion in the 1880s rather dull. However, the NL was joined by the American Association in 1882. The AA played its inaugural season with six teams, expanding to eight in 1883 – and that was the warmup act. In 1884, the AA jumped to 12 teams, which on its own would represent a 150% increase in the number of major league teams over a three-year span. But also in 1884, the Union Association sprang onto the scene, with eight teams of its own (roughly – the UA makes the NA look like a bastion of stability). The 1884 season saw an approximate total of 28 major league teams, a total that would remain unmatched for decades.
It would remain unmatched, of course, because the UA folded
after one season, and the AA dropped back down to eight teams at the same time. The
two leagues would remain comparatively stable for the next few years, a stretch that ended when a labor dispute caused a number of players to break away and found their own
league (sensibly named the Players’ League). The PL was higher-quality than the
UA (it was probably the strongest professional league in 1890), but shared
the same fate, collapsing after a single season. Not only
did the financial fight doom the PL, it also crippled the AA, which ceased
operations after 1891. However, the NL opened its doors to the AA’s four best
franchises, expanding to 12 teams, a total it would maintain through 1899.
The 1899 baseball season merits a book-length discussion; its issues are too extensive to explore in detail here. For our purposes, the result of that disastrous season was that
the NL discontinued four teams at the end of the year, going back to a total of eight,
and making 1900 the first season since 1881 with a single-digit number of major
league teams. It would also be the last such season to date, as the American
League made the jump to major league status in 1901. This set
the total number of major league teams to the 16 that would remain standard for
six decades… sort of. The Federal League tried to duplicate the AL’s genesis in
1914, leading to two years of three-league, 24-team majors, before the FL
folded and the 16-team setup resumed.
(Yes, the Negro Leagues also merit discussion in this sense.
We will cover them in more detail in a future article; for now, suffice it to
say that because they did not share a player pool with the segregated AL/NL
organization, their initial formation and later changes in their league
structure did not have a direct effect on the AL/NL quality of play until
integration, and that effect was already folded into the segregation adjustment
discussed previously.)
As before, here is the history of pre-1920 MLB league size
in table form:
|
Season |
Size change |
Final size |
|
1871 |
+8 |
8 |
|
1875 |
+2 |
10 |
|
1876 |
-2 |
8 |
|
1877 |
-2 |
6 |
|
1879 |
+2 |
8 |
|
1882 |
+6 |
14 |
|
1883 |
+2 |
16 |
|
1884 |
+12 |
28 |
|
1885 |
-12 |
16 |
|
1890 |
+8 |
24 |
|
1891 |
-8 |
16 |
|
1892 |
-4 |
12 |
|
1900 |
-4 |
8 |
|
1901 |
+8 |
16 |
|
1914 |
+8 |
24 |
|
1916 |
-8 |
16 |
The late 1900s brought several instances of coordinated, measured expansion. The 1800s were the comparative wild west, with aggressive expansion followed quickly by rapid, corrective contraction, over and over again. And while the older expansions have some differences from the newer ones, in that the existing league structures usually didn’t endorse them or directly facilitate the building of the new teams’ rosters, they were often characterized by players jumping their contracts from the existing teams to join new leagues, so the quality of play in the 1901 NL was definitely still affected by the birth of the AL.
Granting that expansion has an effect (and an especially
large one in much of the 19th century), how do we account for it?
Once again, I am sadly bereft of a systematic way to make the adjustment, and
have resorted to an arbitrary method with sanity checks applied. As before, all
adjustments are applied to replacement level on a per-600 PA basis.
The basic framework of the method I’m using is as follows: In
an expansion year, adjust replacement level by 0.1 wins per team added. For
each year after expansion, reduce the adjustment by 0.1 until it reaches 0. (So,
in 1969, the league added four teams, and 1969 WAR values are thus reduced by 0.4
WAR per 600 PA. This adjustment diminishes by 0.1 per year moving forward, finally vanishing in 1973.)
If an expansion is counteracted by contraction afterward,
remove the expansion effect to the extent that the contraction would explain.
(The formation of the Federal League in 1914 is an eight-team expansion, so WAR values are cut by 0.8 per 600 in 1914 and 0.7 in 1915. The FL folds before
1916, and thus the adjustment resets to 0.)
Two special cases are worth noting. First, as discussed
above, the first professional team formed in 1869, so I’m treating 1871 as an
eight-team expansion over the course of three years, therefore starting with an
adjustment of 0.6 per 600 rather than 0.8. Second, the reduction from 12
teams to eight in 1900 is the only one that had not been recently preceded by an
expansion; in fact, the majors had already shrunk twice in the preceding
decade, from 24 to 16 to 12. As such, I actually boosted WAR values in 1900 by
0.4 wins per 600 PA, and counted the 1901 expansion as effectively being from
12 teams to 16, staging the adjustment in from 0.4 in 1901 and reducing it to 0
by 1905.
So, with all of that applied, how does the sanity check come
out? Here are the totals for top-100 positional players by decade of median
plate appearance (plus second basemen, since that’s who we’re ranking in
this article):
|
Decade |
Players |
2B |
|
1870 |
5 |
1 |
|
1880 |
21 |
2 |
|
1890 |
24 |
2 |
|
1900 |
31 |
5 |
|
1910 |
28 |
4 |
|
1920 |
48 |
7 |
|
1930 |
42 |
6 |
|
1940 |
38 |
4 |
|
1950 |
52 |
7 |
|
1960 |
55 |
4 |
|
1970 |
71 |
6 |
|
1980 |
77 |
10 |
|
1990 |
86 |
11 |
|
2000 |
91 |
11 |
|
2010 |
108 |
16 |
|
2020 |
23 |
4 |
Outside of a few exceptions, that is a steadily increasing trend. (1870 sees the beginning of pro baseball, so players who were mid-career at that point may be underrepresented; it also had shorter schedules than the 1880s which skews median plate appearances forward. 1920 is the debut of the officially recognized Negro Leagues; some players listed in the ‘20s may well have been placed in the teens if their stats from those seasons were counted. The 1910s and particularly the 1940s were affected by world wars. And finally, the 2020s are barely halfway in the books, and some players will reclassify from 2010s to 2020s by the end of their careers; some have already done so as a result of the 2024 season.) The second base totals follow the same general trend, albeit with the increased inconsistency you’d expect from a smaller sample size.
Speaking of second basemen, let’s get to our tables. First,
as always, active second basemen (as of 2023) who are either in the top 100 or within
reasonable distance of breaking in:
|
Player |
Rank |
Years |
WAR |
aWAR |
wWAR |
2024 WAR |
Rank Change |
|
Jose Altuve |
19 |
2011-24 |
52.8 |
53.0 |
42.3 |
3.4 |
+3 |
|
Marcus Semien |
26 |
2013-24 |
45.8 |
46.3 |
39.0 |
4.1 |
Pos Chg |
|
DJ LeMahieu |
43 |
2011-24 |
29.9 |
35.2 |
29.2 |
-1.6 |
-1 |
|
Ketel Marte |
47 |
2015-24 |
31.3 |
33.0 |
28.6 |
6.8 |
+22 |
|
Jeff McNeil |
88 |
2018-24 |
20.8 |
22.5 |
20.2 |
1.6 |
+5 |
|
Ozzie Albies |
89 |
2017-24 |
21.7 |
22.5 |
20.0 |
1.6 |
+6 |
|
Andres
Gimenez |
97 |
2020-24 |
18.6 |
19.8 |
18.6 |
4.0 |
+37 |
|
Tommy Edman |
105 |
2019-24 |
17.9 |
19.5 |
17.9 |
|
|
|
Whit
Merrifield |
109 |
2016-24 |
18.4 |
19.3 |
17.3 |
|
|
|
Brandon Lowe |
113 |
2018-24 |
16.0 |
18.8 |
17.0 |
|
|
|
Jorge Polanco |
114 |
2014-24 |
18.6 |
18.9 |
16.9 |
|
|
|
Luis Arraez |
126 |
2019-24 |
16.0 |
17.0 |
15.6 |
|
|
|
Nico Hoerner |
138 |
2019-24 |
15.2 |
16.0 |
14.9 |
|
|
|
Gleyber
Torres |
144 |
2018-24 |
16.1 |
16.1 |
14.6 |
|
|
|
Jake
Cronenworth |
147 |
2020-24 |
13.2 |
15.0 |
14.0 |
|
|
Gimenez leapfrogged four active players to join the top 100 this season, but there’s plenty of space available for anyone currently outside who plays well; second base has the lowest barrier to entry we’ve seen so far. By comparison, the top end of the group is pretty thin, with only four members of the top 80 (would have been three but for Semien’s position change – which, by the way, is why LeMahieu’s ordinal ranking went down). The list of active top-100 second basemen might well increase in size by 50% or more in the next two years.
Finally, here are the top 25 second basemen ever by weighted
WAR, plus the usual addition of 30 to 100 by tens:
|
Player |
Rank |
Years |
WAR |
aWAR |
wWAR |
|
Rogers Hornsby |
1 |
1915-37 |
127.1 |
121.5 |
85.9 |
|
Eddie Collins |
2 |
1906-30 |
124.6 |
110.9 |
71.8 |
|
Joe Morgan |
3 |
1963-84 |
100.6 |
100.4 |
69.4 |
|
Nap Lajoie |
4 |
1896-1916 |
107.2 |
96.8 |
66.6 |
|
Rod Carew |
5 |
1967-85 |
81.3 |
81.7 |
57.7 |
|
Charlie Gehringer |
6 |
1924-42 |
84.7 |
76.5 |
56.8 |
|
Bobby Grich |
7 |
1970-86 |
71.4 |
73.5 |
54.5 |
|
Robinson Cano |
8 |
2005-22 |
68.0 |
70.7 |
53.7 |
|
Jackie Robinson |
9 |
1945-56 |
64.5 |
65.3 |
52.9 |
|
Ryne Sandberg |
10 |
1981-97 |
68.1 |
68.4 |
51.9 |
|
Chase Utley |
11 |
2003-18 |
64.5 |
64.6 |
51.4 |
|
Lou Whitaker |
12 |
1977-95 |
75.0 |
77.2 |
50.3 |
|
Craig Biggio |
13 |
1988-2007 |
65.2 |
68.7 |
50.1 |
|
Roberto Alomar |
14 |
1988-2004 |
66.7 |
67.8 |
49.7 |
|
Frankie Frisch |
15 |
1919-37 |
72.0 |
63.1 |
47.2 |
|
Willie Randolph |
16 |
1975-92 |
65.7 |
66.9 |
45.6 |
|
Ross Barnes |
17 |
1871-81 |
28.1 |
48.6 |
43.1 |
|
Dustin Pedroia |
18 |
2006-19 |
52.0 |
53.4 |
43.0 |
|
Jose Altuve |
19 |
2011-24 |
52.8 |
53.0 |
42.3 |
|
Ian Kinsler |
20 |
2006-19 |
54.2 |
54.9 |
42.2 |
|
Jeff Kent |
21 |
1992-2008 |
55.5 |
56.1 |
41.7 |
|
Joe Gordon |
22 |
1938-50 |
55.6 |
51.1 |
41.6 |
|
Chuck Knoblauch |
23 |
1991-2002 |
44.5 |
46.6 |
39.3 |
|
Tony Phillips |
24 |
1982-99 |
51.3 |
53.1 |
39.2 |
|
Ben Zobrist |
25 |
2006-19 |
44.6 |
46.6 |
39.1 |
|
|
|
|
|
|
|
|
Jim Gilliam |
30 |
1946-66 |
45.4 |
45.3 |
35.6 |
|
Buddy Myer |
40 |
1925-41 |
48.6 |
41.1 |
31.4 |
|
Bid McPhee |
50 |
1882-99 |
52.8 |
38.1 |
27.7 |
|
Orlando
Hudson |
60 |
2002-12 |
30.9 |
30.9 |
25.9 |
|
Bobby Avila |
70 |
1949-59 |
28.5 |
26.9 |
23.8 |
|
Jimmy
Williams |
80 |
1899-1909 |
32.5 |
25.0 |
21.6 |
|
Billy Goodman |
90 |
1947-62 |
27.5 |
25.1 |
19.9 |
|
Daniel Murphy |
100 |
2008-20 |
20.3 |
21.6 |
18.3 |
Despite four of the top six having debuted before 1950, the total number of older debuts from the top 25 is only eight. One of them is Ross Barnes; ranking Barnes ahead of, say, Pedroia, is likely to be a controversial take. I’ll offer two caveats. First, Barnes led his leagues in position player WAR five times in a six-year span; 1870s or not, that is impressive. And second, I’m not claiming these as my own personal rankings. I would have to do considerable research on early baseball to feel comfortable with any assessment of Barnes or his contemporaries.
It’s also worth noting that Hornsby and Collins have very
similar career WAR totals, and their careers have significant overlap, but
Hornsby loses less than 6 WAR to adjustments while Collins loses over twice
that. The reason is Hornsby’s peak, which is one of the highest ever.
Timelining adjustments are additive based on playing time, while schedule
length adjustments are multiplied by the player’s WAR total; this effect causes
a giant peak season (Hornsby has six over 10 WAR, with a maximum of 12.1) to lose
comparatively less value than a merely excellent season, of which Collins has a
larger number.
Speaking of peak adjustment, look at Lou Whitaker, who is
sixth in total adjusted WAR and drops to #12 when weighting is
applied. Comparing him to Chase Utley, Whitaker leads by 12.6 adjusted WAR, but
falls behind due to his famously flat peak/prime compared to Utley’s collection
of high-value years.
That wraps up the discussion of timelining. Next time, we’ll
head out to center field and look at our final adjustment to WAR totals:
negative seasons.
