In our last two entries, we’ve discussed pitchers 100-51 and 50-11 according to GSDev career rating. Now, we’re going to get into rather more detail about the cream of the crop. It’s time for the top ten.
If you are a
dedicated student of baseball history who has for some reason followed these
rankings very closely, you may well already know which ten pitchers we’re going
to discuss. Even if you don’t fall into that category, getting the last ten
names dropped all at once feels rather anticlimactic. So let’s go through them one
at a time, taking the opportunity to expand on other topics as we go.
#10: Nolan
Ryan (1966-93, 50.50 RSS, 128.8 Weighted, 50.37 Combined, #16 in bWAR)
According to GSDev, Nolan Ryan was never the best pitcher in baseball. He had four top-3 finishes,
but was at least one full deviation back of the top spot in all but one of
those seasons. He has no seasons in the all-time top 100, and only three (1977,
1973 and 1989) in the top 300. So what exactly is he doing up here, ranked
ahead of pitchers who also had long careers and capped them off with much higher
peaks?
You are
likely aware that Nolan Ryan holds an enormous number of all-time records: most
strikeouts, most walks, most no-hitters, fewest hits allowed per 9 innings, and
so on. But let’s start at an even more basic level than that.
Most seasons
with 10 or more starts: Nolan Ryan, 26
Most seasons
with 15 or more starts: Ryan (tied), 24
Most seasons
with 20 or more starts: Ryan (tied), 22
So if you
want to explain Nolan Ryan’s placement here, start with the fact that he showed
up for work as often as anyone. He’s second all-time in games started, and the
guy ahead of him is Cy Young, whose career is only half-complete in our sample.
And he didn’t just show up, either; his adjusted Game Scores were above average
in 24 of the 26 seasons in which he made multiple starts, and at least 5 points
above average in 15.
You may have
noticed that Ryan’s RSS score is higher than his combined score, which would
normally be characteristic of high-peak, short-career pitchers. Ryan, as we
have discussed, is neither of those things. There’s an interesting phenomenon
involved in the interaction of RSS and peak-weighted sum. RSS will generally be
higher for pitchers with relatively short careers (say, less than 10 quality
seasons), and after that, weighted sum will dominate… for a while. In the rare case of a pitcher posting more than 20 good seasons, the weighted sum will all but ignore anything from year 21 onward, allowing RSS to pull ahead again. That’s what happens to Ryan, whose #20 season (the
injury-affected 1975) was worth a perfectly solid 6.28 GSDev. He had so many
years as a worthwhile pitcher that the career metrics push him back ahead of
several pitchers who reached loftier single-season summits.
Amusingly, this means that the pitcher whose approach to the game would be possibly the worst match imaginable for “slow and steady” somehow ends up as the tortoise passing a legion of hares at the finish line.
#9: Lefty
Grove (1925-41, 49.63 RSS, 131.5 Weighted, 50.46 Combined, #4 in bWAR)
Grove is the
pinnacle of the discussion on bullpen work by old school starting pitchers,
particularly in his early years. His relief outings from 1925-34 went 27, 12,
23, 8, 5, 18, 11, 14, 17, 10. His usage pattern changed after that, with only 13 total bullpen appearances in his last 7 campaigns. In spite
of the late-career shift, Grove ended up relieving in over 25% of his appearances, and those outings accounted for just under 10% of his total innings. As such, he probably suffers as much as anyone this side of Dennis
Eckersley from the fact that GSDev is a starts-only metric… and yet here he is
at #9 in the career rankings. If you can lead your league in ERA nine times and
FIP eight times, I would highly recommend it.
Grove also
makes a nice segue to a discussion of missed time. There are a few varieties of absence to cover. First and simplest is time missed due to injury or voluntary
departure (such as early retirement). If the pitcher is either physically unable
to perform or actively chooses not to, I don’t feel it’s wildly controversial
to say that he shouldn’t get any credit for that. The same would apply to
players who were banned from baseball while still active (such as Eddie
Cicotte, a Black Sox member who ends up at #117 in the GSDev career rankings). If you
throw the World Series or otherwise get permanently booted from the sport, I’m not pretending you were still pitching after that.
Second is
involuntary absence from the sport. This would primarily be caused by military
service, although early baseball also saw a few players get blackballed due to contract
jumping and the like. There is a much stronger argument to be made for credit
here, especially for players who were demonstrably excellent both before and
after their absences (such as Bob Feller). But one could also reasonably argue,
especially in the case of a pitcher, that being pulled away from the game
prevented additional strain on the pitcher’s arm and allowed him to perform
better once he returned than he otherwise would have. In Feller’s case, he had
been the best pitcher in baseball in 1939 and ’40, but slipped somewhat in ’41,
posting a career-low (to that point) in strikeout rate while his walk rate
climbed. After his 3.5-year stint in the Navy, he came back arguably better
than ever, with a 2.18 ERA and 348 strikeouts in 1946. Maybe he’d have done
that anyway, or maybe his arm would’ve given out by then. But it’s at least
worth considering before blindly handing out extra credit.
Third is non-MLB
play, whether in the minors, the Negro Leagues, Japan, or anywhere else high-level play can be documented. If the
pitcher is still pitching, the argument that he’s saving wear on his arm
dissolves, and therefore I think this is the subset that has the strongest case
for adding credit beyond the MLB record. The ultimate example here is Don
Newcombe, who started in the Negro Leagues as a teenager, spent three years in
the minors (pitching extremely well at every stop), and in his mid 30s even ended his career in
Japan (pitching decidedly less well by this point). Throw in a two-year absence
due to the Korean War, and you can argue that he should get a considerable bump
from his current placement just outside the top 200.
In Lefty
Grove’s case, he was famously held by the minor league Baltimore Orioles for
several years in the early ‘20s before their owner could be coaxed into selling
him to Connie Mack and the A’s. The young Grove’s minor league record shows a
pitcher who was undeniably wild (at least 5 walks per 9 innings each year from
1921-23), but still highly effective (going 108-36 and averaging 280 innings
per full season for Baltimore). Given that he also won his first ERA and FIP
titles in his second year in the majors, he was likely MLB-ready a year or two before
he actually arrived. Reasonable credit for even one minor league season would move him up
to #7 on this list.
#8: Curt
Schilling (1988-2007, 49.76 RSS, 131.8 Weighted, 50.58 Combined, #21 in bWAR)
Look, Curt Schilling was a terrific pitcher. But he also won no Cy Young awards and no ERA titles, so putting him in the top 10 is a leap. Let’s talk through how GSDev differs from a relatively conventional ERA-based approach to Schilling’s career.
First, postseason
play, which given how much I’ve harped on it should be no surprise. Schilling is genuinely
one of the greatest playoff pitchers of all time: 19 starts, 11-2, 133.1
innings, 2.23 ERA, 3.06 FIP, and the highest career postseason win probability
added of any starter (+4.1, trailing only Mariano Rivera among pitchers
overall). You probably remember his famous efforts in 2001 and 2004, but Schilling was also excellent
in the postseasons of 1993, 2002, and 2007, leading his teams to four pennants
and three championships out of five playoff runs.
Second, Schilling’s FIP was better than his ERA. He is ninth all-time in
strikeout to walk ratio (4.38), and the eight pitchers ahead of him pitched either in the distant past (before the number of balls in a walk was reduced to four) or extremely recently (after the league’s strikeout rates went through the roof). Only two other pitchers who appeared in the 1900’s could manage a K/BB ratio greater than 4; one is Rivera, and the other is yet to come on this list. All
of which is to say, in a measure that GSDev would value pretty highly,
Schilling would have retired at the top of the list for a sample starting in
1901.
Third is a
topic I’ve skated past a couple times in this series, and it’s time to do it
justice: unearned runs. The main reason that I join the run-based WAR systems
in counting unearned runs is that, despite the name, the pitcher still generally contributes to these runs scoring. More specifically, pitchers with certain characteristics are known to allow more unearned runs than they would otherwise. Let’s go through them
from most to least obvious:
Pitchers who commit a lot of errors allow more unearned runs than their clean-fielding counterparts. This
is pretty straightforward.
Bad pitchers
allow more unearned runs than good pitchers. If a fielder’s error puts the
pitcher in a jam, a good pitcher is more likely to escape.
Knuckleballers
allow more unearned runs than non-knuckleballers, thanks to the enormous number
of passed balls they inflict on their catchers.
Ground ball
pitchers allow more unearned runs than flyball pitchers. Infielders make more
errors than outfielders, both on a per-play basis and in total, and their
errors tend to be more damaging (they’re more likely to come at the cost of a
missed out and a new baserunner, while outfield errors often just involve
allowing existing runners to advance).
Looking at these four characteristics: Schilling made relatively few errors (22 in 3261 career innings,
on the low end of the range for modern pitchers with similar inning totals).
He was, obviously, a very good pitcher. Like most modern pitchers, he didn’t
throw a knuckleball. And finally, he was an extreme fly ball pitcher. If you
compare him to Kevin Brown (also an excellent pitcher with a near-identical
innings total and ERA+, but a ground ball specialist who made 49 career
errors), Schilling allowed 65 unearned runs to Brown’s 172.
65 unearned
runs in 3261 career innings gives Schilling a UERA (like ERA, but for unearned
runs) of 0.18. Baseball Reference doesn’t directly allow unearned run-based
searches in their Stathead service, but best I can tell from trying to run tangentially
related queries, Schilling’s UERA is the lowest of any pitcher with at least
2000 innings pitched. I’ve only been able to find two other pitchers (Jered
Weaver and Jake Peavy) who do better than 0.22. This is admittedly a niche category, but not too many pitchers can claim to be the greatest of all time at something with a direct effect on the scoreboard.
There are
other things we could explore for Schilling as well; I’ll briefly mention that
he broke into the league as a reliever for a few years, and if you ignore his
relief work (as GSDev does), his ERA gets a bit better and his K/BB ratio is
even higher. But the great playoff work, great fielding-independent numbers,
and historic unearned run avoidance hopefully help explain why he ranks so much
higher here than you might expect given his overall reputation.
#7: Grover
Cleveland Alexander (1911-30, 49.81 RSS, 131.8 Weighted, 50.61 Combined, #3 in
bWAR)
Within the
last couple of years, Baseball Reference has started moving away from referring to
players by nicknames. They began by removing the obviously rude ones (Chief Bender, nicknamed for his Native American origins, is now listed as Charles; the hearing-impaired Dummy Taylor is now Luther), but they’ve expanded to names that
weren’t clearly offensive or insulting (Home Run Baker is now Frank, Oyster
Burns is now Thomas). The policy is not universal (for instance, George Ruth, John Wagner, and
Jay Dean are still Babe, Honus, and Dizzy), but it covers a pretty
good chunk of players, especially older ones. My opinion on this change is mixed, as I do think it comes at a cost; a player such as Virgil Davis will be much more memorable to a casual observer if you call him Spud. But a major point in
favor of this policy is that it overwrites one of the least-inspired nickname
choices of all time. Finally, Pete Alexander can be Grover again.
Grover Cleveland Alexander. His name is quite literally poetic; it is spoken in trochaic tetrameter. It practically demands to be used in couplets, so let’s try a few.
Grover Cleveland
Alexander
Sold for
naught but lint and dander
Alexander changed
teams three times in his career, two of them rather famous. In mid-1926, toward
the end of his career, the Cubs outright waived him; the Cardinals picked him
up and he pitched them to a World Series win. Much earlier in his career, Alexander
was probably the best pitcher in baseball when the Phillies sent him to the
Cubs for a relatively uninspiring return. The Phillies were expecting
Alexander to be drafted to fight in World War I… and they ended up being
correct.
Grover
Cleveland Alexander
Sent away to
fight in Flanders
Despite the
tempting rhyme, Alexander’s military service appears to have been primarily in
France rather than Belgium. Either way, it cost him most of the 1918 season.
(For what it’s worth, war credit is unlikely to move him up much – maybe one
spot if you’re generous.) He returned from Europe with shell shock, epilepsy, and
alcoholism – and despite that, still grades out as an ace in eight of the next nine
years.
Grover
Cleveland Alexander
Had a wife
and couldn’t stand ‘er
Catchy though
it may be, this is a mischaracterization of Alexander’s marriage, although it’s
rarely a good sign to divorce the same woman twice. Alexander’s non-baseball
life was genuinely tragic, to the point that in his twilight years he once said “I’m in the Hall of
Fame… but I can’t eat the Hall of Fame.” He died all but penniless, never
really having recovered from the war.
Grover Cleveland
Alexander
At his
numbers, take a gander
Alexander’s 1915 season has the #13 GSDev of all time. This makes sense; it’s hard to argue with a 1.22 ERA and a 1.82 FIP in 376.1 innings, regardless of how favorable the circumstances are (especially when it’s followed by two excellent World Series starts). The surprise is that he has no other seasons in the top 100 when his 1916 and ’17 were nearly identical to ’15 at a glance. 1916 in particular featured 16 shutouts, the modern record by a healthy margin. And per bWAR, none of those three mid-teens years was his best; 1920 (27-14, 1.91 ERA in front of worse fielders) is preferred to the tune of 12.0 WAR. Oh, and don’t forget 1911, one of the greatest rookie years any pitcher has ever had (28-13, 2.57, 7 shutouts).
With all
that standout work, Alexander led MLB in GSDev a grand total of once (unsurprisingly in 1915).
And if you’re thinking “there’s another pitcher at the same time who was even
better, that’s probably why,” you would only be partly correct.
Alexander
finished second in GSDev five times in his career, behind five different #1
pitchers. Apart from the pitcher you likely thought of (who edged past him in
1916, 15.42 GSDev to 15.03), the other four aces to push Alexander to #2
were:
1911: Ed
Walsh (14.90 GSDev to 13.61). Walsh was a three-time #1 whose rubber arm was on
the verge of giving out, but it hadn’t done so quite yet. Alexander was a rookie up against an established star; he would get better chances later.
1917: Eddie
Cicotte (15.08 GSDev to 14.90). This one stings. Coming on the heels of a very close loss in 1916, Alexander was narrowly bested again, this time by a markedly
inferior pitcher. The especially bitter irony is that Alexander was ahead at
the end of the regular season; Cicotte improved his score from 14.50 to 15.08 with
two good starts in the World Series. Having the eventual ringleader of the
Black Sox pass you based on October heroics is unfortunate on a multiplicity of
levels.
1920: Stan
Coveleski (14.84 GSDev to 14.36). In his first full season back from the war, Alexander
ended the regular season with the GSDev lead again, an even bigger one than he’d had
in 1917. Coveleski then posted three of his seven best starts of the year in the World
Series (three complete games, only two total runs allowed), boosting his score
from 13.43 to 14.84. For a pitcher whose World Series feats were so famous that
they took up over half of his Hall of Fame plaque, it is remarkably unfortunate
that he loses two different seasonal #1 finishes due to the inclusion of
postseason stats.
1923: Dolf
Luque (16.15 GSDev to 12.33). At this point, Alexander was no longer what he
had once been. 1923 was his ninth and last 300-inning season, and his GSDev ranks in the other
years from 1921-25 were 12, 7, 22, 12. And yet, even approaching the sunset of his
career, he was still capable of hanging with the best the game had to offer.
Ultimately,
this is what you’d expect from the highest-ranked pitcher who only had one
season at #1: someone who just barely missed multiple other chances at the top
spot.
#6: Tom
Seaver (1967-86, 51.90 RSS, 137.1 Weighted, 52.68 Combined, #5 in bWAR)
Let’s talk
about arbitrary benchmarks. Earlier in the series when discussing the scale of
GSDev, I classified a score of 10 as an ace-level season. I said at the time that
the number of 10-point seasons has never matched or exceeded the number of
teams in the league, and that’s true. As such, I’ve been using 10 GSDev as a
cutoff for entry into certain sub-spreadsheets dealing with high-performance
seasons.
Tom Seaver has ten seasons of 10-plus GSDev, tied for tenth all time. If you instead add up each pitcher’s “GSDev above 10” (that is, subtract 10 from each 10-point season, to measure how high the pitcher’s peak is), he stands at #8. That all fits pretty nicely with this placement in the rankings. Seaver also has eight top-400 seasons, tied for fourth; the only three pitchers ahead of him in that category are also ahead of him on this list. So he does quite well by the arbitrary round number benchmark I selected.
But here’s
the catch: the original statement about number of pitchers hitting the cutoff
compared to number of teams in the league also applies if you use a GSDev of 9
to define an ace-level year. (There have been a couple of seasons in which
total 9-pointers exactly matched total teams in the league – 1992 had 26 and
2007 had 30. But it’s never exceeded the team count.) And if you use that
definition, Seaver benefits from the five seasons in which his GSDev landed between 9 and 10,
more than any other pitcher. That moves him up to 15 seasons total over 9, tied
for second all-time (with Nolan Ryan). He also has 16 seasons at measured ace
level (that is, ranking within the top N pitchers in baseball where N is the
league size); that’s tied for third, also with Ryan.
Moving the
cutoff from 10 to 9 also changes what the shape of Seaver’s career looks like.
Seaver’s ten years of 10-plus GSDev were consecutive, coming from 1968-77. He ranked
in the top 13 every year in that stretch, and in the top 4 seven times,
including three separate seasons as the best pitcher in baseball. All of that
is extremely impressive and not to be discounted. But if you add the
nine-pointers to the mix, you suddenly get to look at his closing act, in which
he was no longer a dominant force but still a highly capable pitcher, posting
ace-level years in ’78, ’79, ’84 and ’85 (and narrowly missing another
9-pointer in the strike-shortened ’81 season). That may not seem overly
important, but not every great pitcher can do it; many are felled by injury or
just lose too much ability all at once. Without those last nine years, Seaver’s
remaining score would place him ten spots lower on this list.
Having
reached Seaver, we’re done with his general era (‘70s to early ‘80s). It’s a
remarkable period for starting pitchers, having placed five entrants (Seaver,
Ryan, Carlton, Blyleven, and Perry) in the top 16. For any other period to
stand out beyond this cohort, it would have to produce something ridiculous,
like… four of the top five.
#5: Pedro
Martinez (1992-2009, 54.88 RSS, 144.6 Weighted, 55.63 Combined, #14 in bWAR)
With Nolan
Ryan, we talked about the extremely high number of seasons in which he pitched a significant number of
games as a strong point in his favor. This time…
Pedro
Martinez spent 18 seasons in the major leagues, but had fewer than ten starts
in four of them (1992-93 when he was first breaking in, and 2007 and ’09 toward
the end of his career). That leaves 14 seasons as a regular or semi-regular
pitcher. 14 years is not a small number for your garden variety major leaguer,
but when you get toward the top of the all-time list, it becomes an issue. Even
the other comparatively short-term pitchers in the top 20 (Grove, Kershaw,
Halladay, Scherzer, Mathewson) have more. #21 Bob Feller missed three and a
half years due to war and still had 15 ten-start years. #23 John Smoltz has a five year mid-career hiatus from the rotation, and still got
to 16. The first pitcher below Pedro to have 14 or fewer ten-start seasons is
Dazzy Vance, who stands at #30 on the career list.
All of which
is to say, career length is a significant impediment to Martinez’s position in
these rankings – and yet, here he is in the top 5. Of his 14 ten-start seasons, 12 of them
have him rated as an ace-level pitcher. 10 were worth at least 10 GSDev. Seven
were among the top 300 seasons of the 1901-2022 period, and that’s not counting
2001, which ranks #302 despite including only 18 starts (only three seasons of less than 20 starts make the top 800). And
as noted in the top 100 seasons list, Pedro’s 1999 and 2000 seasons were better than
any season any other pitcher has had in the last century-plus. Overall for his
career, Pedro’s average adjusted Game Score was 13.8 points above league
average; the second-highest figure in the database so far is Jacob deGrom’s
12.0, in barely half as many starts. The distance from Martinez to #2 in this
category is slightly larger than the distance from #2 to #11 (Bob Gibson) among
pitchers with at least 100 starts.
He may have
only had 14 seasons as a regular starter. But for those 14 seasons, Pedro
Martinez was better than just about anyone has ever been.
#4: Greg
Maddux (1986-2008, 55.67 RSS, 146.5 Weighted, 56.41 Combined, #6 in bWAR)
To some
extent, the discussion of Greg Maddux picks up where Nolan Ryan’s entry left off. Maddux spent 23 seasons
in the majors. In 22 of them, he had 25 or more starts. Yes, that means he was
one of the pitchers tied with Nolan Ryan for most 20-start seasons. If you count 25-start seasons, Maddux stands alone at the top of the list. He trails Don
Sutton 20-19 in seasons with at least 30 starts, but if you account for labor
disputes, they would be tied at 21 apiece.
Of course,
I’m approaching this on a seasonal level because that’s how GSDev does it. But
it’s also simple enough to pull total career starts within our sample, and in
that category, Ryan is first (780, including the playoffs), and Maddux’s 770 starts tie Sutton for second. The obvious question (what separates Maddux from the
other two extremely long careers?) has an equally obvious answer. Sutton is
famously peakless; his best season (1972) is #157 all time, and it’s one of
only two times he lands in the top 700. Ryan does somewhat better, with two seasons
in the top 150, but still only four in the top 400. Maddux, as seen earlier in
the series, placed six seasons in the top 100 overall.
Greg Maddux
wasn’t as unassailable on a per-game basis as Pedro Martinez; almost nobody
ever has been. For his career, Maddux’s average adjusted Game Score was “only”
8.3 points above average (still #25 among pitchers with over 200 starts). But if you take his seasons from 1991-2002, a stretch
including 425 starts (2 more than Pedro’s career total, including postseason
for both pitchers), Maddux’s Game Scores cleared average by a Pedroeqsue 13.3. The extra 340-plus starts he added at the beginning and end of his career weren’t wildly impressive in their own right – but they were good enough to nudge him ahead.
#3: Walter
Johnson (1907-27, 58.25 RSS, 152.6 Weighted, 58.89 Combined, #1 in bWAR)
We’ve spent
quite a bit of time talking about why the GSDev system isn’t a particular fan
of pitchers from the deadball era. There are 12 total pitchers in our top 100
who debuted before 1920, and only one of them ranks higher on this list than in total
bWAR. (That was Dazzy Vance, who shouldn’t really count; despite his 1915 debut, he didn’t stick in
a major league rotation until 1922.) There are another 11 pitchers who started their careers before 1920 and made bWAR’s top 100 but missed the GSDev list entirely.
Walter
Johnson faces the same challenges that his contemporary rivals do in these ratings: the league’s
deviations were higher, he pitched in relief a lot (136 times in his career,
plus a famous one in the 1924 World Series), and he put up higher volumes than
modern pitchers (nine seasons of at least 36 starts, more than any pitcher has
had in a single regular season since 2003). And yet, here he is in the overall
top five. How?
The years 1901-1920 accounted for 12 of GSDev’s top 100 seasons. The breakdown of those seasons is: Walter Johnson 6, everyone else 6. If you’re thinking that might be a fluke and you want to go past #100 to check, the next two entries from this period (#110 and #117) make the score 8-6 in favor of the Big Train.
As has been
mentioned previously, Johnson’s six top-100 seasons are tied for the
second-most. That is also where he falls in other related categories, such as top-400
seasons (10, tied for second) and seasons of at least 10 GSDev (13, second). If
you switch to seasonal ordinal rankings, he does even better: 11 years as a
top-quartile ace (second), 13 as a top-half ace (second), and 18 as an ace
(tied for first). Those 18 ace years are out of a total of 21, which is quite
remarkable; Johnson had as many ace seasons as he did seasons with at least 20 starts.
The number that towers over the rest, though, is this: Walter Johnson rates as the best pitcher in baseball six times (1912, ’13, ’14, ’16, ’18 and ’19). That stands alone as the most of anyone in the sample.
The GSDev
system does not rate Walter Johnson as the greatest pitcher of all time. I
happen to think this evaluation is probably correct. But if you disagree, I
cannot confidently tell you that you’re wrong.
#2: Randy
Johnson (1988-2009, 58.70 RSS, 156.4 Weighted, 59.84 Combined, #7 in bWAR)
Here are some
things that have been shared already in this series:
Walter
Johnson is the #3 ranked pitcher by career. He had six seasons in the overall
GSDev top 100, and also led the majors in six separate years. (They were not
the same six years; two of his top-100 efforts finished in second place seasonally.)
Randy Johnson
is the #2 ranked pitcher by career. He had six seasons in the overall top 100,
all six of which were also in the top 40; four of those six ranked in the top
20. When considering a sample of over 120 years, a top-20 overall season would
in theory have at least 5/6 odds of leading the league in its year.
Randy Johnson
led the league in GSDev a grand total of two times. This is a respectable
total, but not what would be expected from the second-best pitcher of all time.
There are a dozen pitchers with more, ranging from Johnson’s rivals in the top
10 (Maddux, Seaver, Grove) to pitchers just outside the top 10 (Bob Gibson, Clayton
Kershaw) to pitchers who finished well back (Robin Roberts, Ed Walsh, half of
Cy Young). Another 16 pitchers had exactly two league leads, and a couple of
them (Mario Soto and Ewell Blackwell) didn’t make the top 100 at all. So how
exactly does a pitcher who usually couldn’t even beat his own contemporaries
end up in the overall #2 spot?
The two
seasons in which the Big Unit did lead the league were 2001 (#3 in the overall
rankings) and 2004 (#20). In 1999, he posted the eleventh-highest GSDev ever
and finished second to Pedro Martinez; in 1995, he finished #19 overall, but
came in behind Greg Maddux. In his other top-100 seasons, he finished behind Martinez
again in 2000 (#33 overall) and Curt Schilling in 2002 (#37). Overall, the Big Unit
finished second six times (also in 1994, behind Maddux, and in ’93, behind
Kevin Appier). That stands alone as the most second-place finishes to date,
edging out Alexander and Grove who are tied with five each. Add in the two
victories, and he has eight top-2 finishes, which ties him for second overall
(Grove has eight as well, and Walter Johnson squeaks by him with nine).
Even with all
of that, Unit’s overall #2 finish surprised me. Just among his contemporaries, he
has a lower peak than Pedro (so does everyone else), and a shorter career than
Maddux (so does almost everyone else), and beating both of them with those
limitations is no easy feat. And that’s not to mention the guy with the same
last name who’s one of the two most common choices for greatest pitcher ever.
Let’s go back
to the method we used to compare Maddux and Pedro, adjusted Game Score over average
measured across an extended prime. Maddux’s 12-year consecutive peak put him
basically even with Martinez’s career start total, with a distance from average
that was within shouting distance of Pedro’s career mark (13.8, the highest
ever). Johnson had a couple of injury years and wasn’t in the playoffs as often
as Maddux was, so it takes him 14 prime seasons (1992-2005) to clear Martinez’s total
starts; that forces us to include a couple of decidedly off-peak efforts.
Randy
Johnson’s adjusted Game Score from 1992-2005 was 15.2 points above average,
blowing past his rivals with room to spare. If you have a subset of your career
that’s as long as the career of Pedro Martinez and can be argued to be better
than the career of Pedro Martinez, that makes a pretty compelling case in your
favor.
Ultimately,
it’s enough to propel the most common second-place finisher to one final
second-place finish. Surprisingly, given the extent to which their careers
overlap, none of the Big Unit’s seasonal second-place finishes were behind the
pitcher who beats him for the #1 overall spot:
#1: Roger
Clemens (1984-2007, 62.94 RSS, 161.5 Weighted, 62.96 Combined, #2 in bWAR)
I’ve avoided
using his name to this point in the post due to potential spoilers. Despite that, I doubt
anyone is surprised here. Clemens has been peppered all over this series,
whether I’ve been mentioning him or not. Going through the categories I’ve used
in discussing other pitchers: Clemens has seven top-100 seasons, twelve of the
top 400, fifteen of 10 GSDev or higher, and seventeen of 9 GSDev or more. All
of those totals rank #1. He has 18 ace seasons, 15 top-half ace seasons, and 12
top-quartile ace years; those totals are either in first, or tied for first.
In Game Score
above average, we already know that Pedro Martinez is #1 all time. #2 is Jacob
deGrom (pending the final numbers from 2023-25), #3 is Curt Schilling. Clemens
sits in fourth, 11.6 points above average. The difference is that Clemens is
also fourth in our overall sample in total starts (741); he’s so far ahead of
the few pitchers with better rate stats that you could give either Martinez or
Schilling all of deGrom’s starts to date and they still wouldn’t catch up. In
total distance above average (instead of the per-game measure), Clemens easily
leads again.
Years at #1, we already know Walter Johnson is the all-time leader with six. Clemens, pitching in a larger league and facing possibly the greatest peer group the sport has ever seen, led the league five times (1987, ’88, ’90, ’91, and ’97), and finished second thrice more (to Mike Scott in ’86, Maddux in ’92, and Kevin Brown in ’98), placing him in a tie with Randy Johnson and Lefty Grove with eight top-2 finishes. If you extend the top-N finishes search beyond 3, Clemens either wins, ties for first, or is second behind Walter Johnson for every value between 4 and 15, and that’s with the disadvantage of a league that has both integrated and nearly doubled in size since Walter was around.
There are arguments you could make against Clemens as the greatest pitcher of all time. He slips a bit if you value the pitcher’s best years being consecutive, and there is always the discussion about the underlying reasons that he was able to rebound to peak level after backsliding in the mid ’90s. GSDev doesn’t put any emphasis on consecutive peak, and has no opinion of Clemens’s extracurricular activities. And beyond those considerations, his results are overwhelming enough that nobody else comes close.
And that’s
the end of our top 100 pitchers of (almost) all time! Up next will be a
postscript to this list; inconceivable though it may seem, there are a few
topics I didn’t touch on in the last three posts that I’d like to cover. Who just missed? Which active pitchers might be about to join? Who does bWAR include that GSDev omits? We
No comments:
Post a Comment