Comparing pitchers directly is very difficult. Numbers like wins and losses are very easily shown to have little intrinsic value, since they are strongly dependent on how a pitcher's team performs. Earned Run Average is more useful, but favors pitchers with short, excellent careers, and also varies somewhat year to year, most notably in the 1960s when pitchers became so dominant that the height of pitchers' mounds was dropped from 15" to 10" to restore balance. For my comparison I used one of the newfangled nerd stats, WARP-3. It measures how many wins a particular player is responsible for based on their performance, compared to a hypothetical "replacement-level" player, taking into consideration the quality of opponents and the bias of the home and away parks, balanced over all seasons since 1890 including adjustments for different season lengths, and quite possibly incorporating the phases of the moon and ladies' skirt lengths.

This all means that WARP-3, or "wins" as I'll call it, has many nice properties. It's directly comparable across different decades or centuries; it's cumulative, so that a pitcher with 15 wins is better than one with 10. It's relative to a baseline, so that a pitcher hanging on past their prime will accumulate fewer and fewer wins even in full seasons, and can actually lose points if they perform badly enough. And it has a clear intrinsic meaning: if pitcher X had a WARP-3 of 10 this year, then his team won 10 more games than it would have with a generic replacement (eg, a journeyman or unremarkable rookie who is barely able to compete at the major league level).

With all that in mind, let's look at the first chart. It shows the career performances of the best currently active pitchers in baseball, Roger Clemens, Greg Maddux, Randy Johnson, Pedro Martinez, and Curt Schilling. I added the best major league pitcher ever, Walter Johnson, for comparison. (The "Big Train" pitched for the Washington Senators from 1907 to 1927 and is consistently considered one of the top five pitchers ever. By WARP-3, he's the best.)

First, we can see that Clemens is on the verge of overtaking Walter Johnson as the best pitcher ever. Second, we can that Maddux and Clemens have traded the lead a few times (Clemens is a few years older than Maddux; this chart lines them up by age, so Maddux is shown four years behind Clemens even though both are still pitching), but Maddux has been falling behind for several years. Maddux last had the lead at age 35; Maddux's 2002 season brought his total to 144.2 wins, Clemens was 35 in 1999 and his total was then 143.6. Since then, Clemens has rebounded while Maddux has declined.

Meanwhile, Pedro Martinez is only slightly behind either of them at age 33, but he'll have to maintain his current plateau for another six or eight years to reach their level. The Big Unit (Randy Johnson) and Curt Schilling are well behind by age, but Johnson has been able to log good years up through age 42, whereas Schilling has already lost most of 2005 to injury and is unlikely to catch Johnson, much less the others. Still, everyone but Schilling is certainly headed for the Hall of Fame, and Schilling is knocking on the door.

How do they look in the context of the best pitchers in history?

Clemens is in second place for all time. Maddux is in fourth, but unlikely to move up. Martinez will have to stay healthy and motivated for years to break into the top 20. Walter Johnson really stands out. Bob Feller had an even better start, but that flat line from age 23 to 26 (when he was serving in World War II) cuts him down brutally. He roared back in 1946 with the fourth-best season ever, but was ineffective by age 33. Randy Johnson's late bloom is the only thing keeping him in the second tier. Nolan Ryan, many people's pick for best, was good but not great for a very long time, but still doesn't break into the top ten. Phil Niekro started later than anybody else but played until age 48.

One name commonly found in this company but missing on this chart is Sandy Koufax. Although he was utterly dominant for four years, his peak coincided with Dodger Stadium (which strongly benefits pitchers) and with the pitcher-dominant 1960s. His career was short, and his peak, though anecdotally awesome, was less impressive in context. By WARP-3, Walter Johnson's 1912-1915 stands above all others, with Pedro's 1997-2000 and Maddux's 1992-1995 tied for distant second. (Looking at relative ERA, gives similar results.)

And that Dodger game where the Dodgers' rookie pitcher matched Clemens for seven innings? The Astros won 3-2 in the ninth. The National League Western Division is the weakest in the majors, and the Dodgers are sinking towards last in it. The Dodgers disabled list has claimed their entire starting outfield, half of the starting infield, the closer, and a big fraction of the rotation. If they could field the DL instead of the healthy players, they could easily win the division. Or at least tie with Barry Bonds solo.