if this becomes a steroid thread, i'm going to cut some #############.

there's a real problem with this question, btw, in that people have varying definitions of "greatest"- it really depends on what you mean by it. if you mean you get to see one game or one season of that pitcher at his peak, i'd have to go with Millennial Pedro. if you're talking watching someone over the full span of their career, you end up picking someone like Maddux.

there's a real problem with this question, btw, in that people have varying definitions of "greatest"- it really depends on what you mean by it. if you mean you get to see one game or one season of that pitcher at his peak, i'd have to go with Millennial Pedro. if you're talking watching someone over the full span of their career, you end up picking someone like Maddux.

Ah, this generation's version of the old peak/career value Koufax/Spahn debate (see the Bill James Historical Baseball Abstract).

Check out the innings pitched line, then pretend the difference in innings pitched between Koufax and pitcher X was pitched by a mediocre pitcher. Koufax pitched many more innings each year than Pedro.

Check out the innings pitched line, then pretend the difference in innings pitched between Koufax and pitcher X was pitched by a mediocre pitcher. Koufax pitched many more innings each year than Pedro.

Here are their 4 best seasons by ERA +:

Koufax:
190 ERA+ 323.0 ip
187 ERA+ 223.0 ip
160 ERA+ 335.7 ip
159 ERA+ 311.0 ip

Pedro:
291 ERA+ 217.0 ip
243 ERA+ 213.3 ip
219 ERA+ 241.3 ip
210 ERA+ 186.7 ip

Now, here is the ERA+ of Pedro, plus a league average pitcher filling in the difference in ip between Pedro and Koufax:
228 ERA+
237 ERA+
186 ERA+
166 ERA+

So Pedro plus a guy league average guy is still better than Koufax. Plus, the fewer innings pitched helped keep Pedro's arm attached a few years longer than Koufax's did.

Is Koufax' 1963-1966 better than Randy Johnson's 1999-2002? Johnson pitched 162.2 fewer innings, but he was more effective (at least by adjusted ERA).

Now, here is the ERA+ of Pedro, plus a league average pitcher filling in the difference in ip between Pedro and Koufax:

What is the basis for adding a league average pitcher to Pedro's total? Are we assuming he would be average over the remaining innings, or that the difference in innings was being filled in by an actual league average pitcher? Neither of these is a likely assumption.

I was just doing what depletion asked, to imagine that the difference in Pedro's ip and Koufax's ip was made up by a mediocore (league average) pitcher.

Koufax wouldn't be throwing 300+ IP if he pitched in Pedro's era.

those who drive slower than you are idiots, those who drive faster, maniacs.

et tu, Joe Po?

Koufax wouldn't be throwing 300+ IP if he pitched in Pedro's era.[/q80te]

That's a key difference, pitchers now throw more pitches per innings so they can't throw as many innings. And their innings count is managed better for health reasons. Perhaps if Koufax's innings were managed better he would have had a longer career.

But also, most of these statistics (ERA+) are measuring distance from mean performance. It's more likely you'll find bigger outliers in larger populations, i.e. after expansion. It's also likely you'll find bigger outliers if the talent pool is diluted, i.e. after expansion or with increasing competition for athletes from other pro sports. That's not saying Pedro isn't the freak he is, but just that I wonder if ERA+ really is a great historical benchmarking mechanism given the changes in those factors.

if you mean you get to see one game or one season of that pitcher at his peak, i'd have to go with Millennial Pedro

I'd go with Big Unit. Watching something that big and ugly just dominate the hitters, racking up unreal strike out totals was good enough. But there were times when there was an overmatched lefty they left in against him, and it was like watching a horror movie -- you had to feel sorry for the batter.

I'd go with Big Unit. Watching something that big and ugly just dominate the hitters, racking up unreal strike out totals was good enough. But there were times when there was an overmatched lefty they left in against him, and it was like watching a horror movie -- you had to feel sorry for the batter.

Roger Angell's description of watching Koufax pitch, collected in The Summer Game, gives a remarkably similar feeling:

"It is almost painful to watch, for Koufax, instead of merely overpowering hitters, as some fastball throwers do, appears to dismantle them, taking away first one and then another of their carefully developed offensive weapons and judgements, and leaving them with only the conviction that they are the victims of a total mismatch."

I was just doing what depletion asked, to imagine that the difference in Pedro's ip and Koufax's ip was made up by a mediocore (league average) pitcher.

Interesting definition of "mediocre" - a 100 ERA+ would be good for the #2 spot in the rotation on the average team, from what I remember. A 97 ERA+ is league average for a starting pitcher. Mediocre to me would be no higher than a 90 ERA+ - somewhere between 85 and 90.

What a shame … in our time we may have seen the greatest hitter and the greatest pitcher of all time. And it seems like nobody even believes it.

Nobody believes it because part of the reason these men were so successful was "allegedly" because of illegal performance enhancing drugs.

How many guys have had better (non-consecutive) peaks than Koufax? Even deflating his numbers, it isn't that many. So, you have Walter Johnson, Clemens, Maddux, Pedro. Maybe Randy and Lefty Grove. Pete Alexander? Basically, I think Koufax is definitely top 10, and maybe top 5. He's still a truly great pitcher.

Robin Roberts, Christy Mathewson, and Kid Nichols deserve a mention. Maybe Hal Newhouser, based on how you discount him for the war years.

you could make an argument for Randy pretty easily. but Pedro was interesting in the opposite direction because he wasn't that big a guy who just utterly dominated other teams, and the effect he had on that teams fans when he got any type of lead. I remember seeing a game at Safeco he started in 2000, when the Sox went up by 4 in the first and the Mariners fans totally gave up on the game. (it was this game- http://www.retrosheet.org/boxesetc/2000/B08020SEA2000.htm).

over the course of 1999 and 2000 he gave up more than 3 ER four times and had game scores under 50 four times. he was the most dominant pitcher i've ever seen during that course of time. to put that into perspective CC Sabathia in 2007, when he was the Cy Young winner, had 8 games where he gave up 4 or more runs.

And here is a comparison of Babe Ruth's home run+ with Kingman

Babe

59+
60+
45+

Kingman

1000+
990+
870+

Kingman and Torres
1001+
990+
870+

Kingman was a much better player even when Hector Torres is added in.

There are as many "greatest" pitchers as there are definitions of "greatest." And if your definition of "greatest" is "the pitcher who demonstrated during a short peak that he could dominate in big games like no other pitcher before or since, and do so in a manner that left opponents feeling beaten almost before the game began," then Koufax is a reasonable choice for "greatest pitcher," if not the only one.

By any other definition, of course, he isn't. But who gets to frame the argument, and why?

Interesting definition of "mediocre" - a 100 ERA+ would be good for the #2 spot in the rotation on the average team, from what I remember. A 97 ERA+ is league average for a starting pitcher. Mediocre to me would be no higher than a 90 ERA+ - somewhere between 85 and 90.

"Mediocre" and "average" are synonymous, although many people use "mediocre" to mean "below average."

Koufax's 311 IP in 1963 placed 3rd in the NL.
Pedro's 241 IP in 1997 placed 4th in the NL.

Pedro was in the league top 10 in IP 6 times, to only 4 for Koufax.
But Sandy led the league twice and Pedro never did.

Randy Johnson, between 1993 and 2002, had as dominant a decade as any pitcher since 1900: 175-58, 2,928 strikeouts(!), and an ERA+ of around 170.

Randy Johnson, between 1993 and 2002, had as dominant a decade as any pitcher since 1900: 175-58, 2,928 strikeouts(!), and an ERA+ of around 170.

And yet, during roughly the same 10 years (1992-2001), Greg Maddux was just as good: 182-82, only 1785 strikeouts, but a mere 375 walks, and an ERA+ of 172 in 180 more innings.

Koufax's 311 IP in 1963 placed 3rd in the NL.
Pedro's 241 IP in 1997 placed 4th in the NL.

Pedro was in the league top 10 in IP 6 times, to only 4 for Koufax.
But Sandy led the league twice and Pedro never did.

All this is the point about context that Pos correctly cites, and of course Pedro's the "greatest" pitcher under the definition of "at his peak, or in a relatively short career, most dominant relative to his peers." But then there's also this:

Pedro: 46 complete games in 389 starts
Koufax: 137 complete games in 314 starts

And in Koufax's last five years, it was 100 complete games in 176 starts.

No matter what the context, there's a lot of extra value in that. Just think of that game 7 of the 2003 LCS for a good example of why bullpens can't always be counted to bail you out.

All of which speaks to Pos's point: Context is everything. And there's no single answer to "Who's the greatest?"

Poz's standard - number of great seasons - is a little bit odd; shouldn't we also care by what margin the seasons in question exceeded his minimal threshold for "great"?

But I'll try playing the same way. I have the records of a large number of pitchers' records turned into RA+ PythPat equivalent records, season by season. I can then turn that into equivalent FWP. For the purpose of this, I'm going to set the threshold for "great" at a single season 18 FWP for the years 1920 to the present, 20 FWP for 1905-1919, and 22 FWP for 1893-1904. Here's what I get:

14 seasons: Cy Young

12 seasons: Roger Clemens

11 seasons: Lefty Grove

10 seasons: Christy Mathewson, Walter Johnson, Kid Nichols, Bert Blyleven

9 seasons: Pete Alexander, Greg Maddux

8 seasons: Bob Gibson, Jim Palmer, Tom Seaver, Randy Johnson

7 seasons: Pedro Martinez, Ferguson Jenkins

6 seasons: Ed Walsh, Eddie Plank, Red Ruffing, Carl Hubbell, Hal Newhouser, Warren Spahn, Whitey Ford, Jim Bunning, Kevin Brown, Mike Mussina, Curt Schilling

5 seasons: Stan Coveleski, Wes Ferrell, Dazzy Vance, Bob Feller, Bob Lemon, Robin Roberts, Juan Marichal, Tom Glavine

4 seasons: Vic Willis, Rube Waddell, Urban Shocker, Ed Reulbach, Babe Adams, Burleigh Grimes, Herb Pennock, Lefty Gomez, Lon Warneke, Sandy Koufax, Early Wynn, Billy Pierce, Gaylord Perry, Luis Tiant, Steve Carlton, Don Sutton, Dave Stieb, Frank Viola, David Cone

In case you're wondering about Nolan Ryan, I only get two such seasons for him.

And different people with the same number of great seasons don't necessarily rate the same overall. For instance: Ferguson Jenkins hit the threshold 7 times, but in most of those years he was barely over the mark. And saying Walter Johnson exceeded 20 (or 18) equivalent FWP 10 times doesn't capture the fact that he had seasons of 53 and 47 equivalent FWP.

[I've brought this up here before]

Everybody assumes that ERA+ can be extrapolated in a linear fashion across eras, for comparison's sake. Or, in other words, we confidently say that a 1.60 ERA is harder to do in an offensive era than it is in a defensive era, using a precise 25% factor (or whatever). But what if there's a point beyond which it becomes virtually impossible to drive that ERA down further, no matter in what environment someone pitches? We all assume that we can take Unit 2001 or Maddux 1994, and extrapolate their hits allowed downward to a 1966 environment using a simple proportion, but what if it doesn't work that way, and it's parabolic or something?

I was just running around at BBRef, trying to figure out how to examine this, but it probably deserves more than the quick-and-dirty approaches I was brainstorming (like looking at what offensive eras the single season leaders in Hits/9 pitched in).