Fun article, and more fodder for the greatness of that mid-90s to mid-00s group. Randy Johnson!

That was an outstanding article. For those that haven't yet RTFA, enjoy this factoid: In 1964, Dean Chance pitched 50 innings against the 99-63 Yankees. He gave up 18 hits and one earned run (a Mantle homer). Unreal.

Bill used to have 1-2 articles like this in each of his 1980s Abstracts. Great stuff, and a reminder of how dominant Pedro was (albeit briefly).

I was curious what Rick Wise's game score was on 9/18/71 – turns out it was 92, compared to his no-hitter that year of 89. B-Ref says that Wise threw 91 pitches in beating the Cubs in 12 innings. Possible, I guess. That would make his Game Score higher than his pitches thrown, which must be a select group (for starts over a certain length, naturally).

That would make his Game Score higher than his pitches thrown, which must be a select group (for starts over a certain length, naturally).

BR lists pitches thrown, consistently, since 1988. Since that year, Greg Maddux has 30 such games (26 if you set a minimum of 5 IP). His closest competitor has 9 (5 if you set the 5-IP minimum).

When was the last time a starter went more than 9 innings? It's got to have been done lately by someone (checking CC, my #1 suspect...nope. Halladay...negatory). I guess it's something that modern-day managers simply Do Not Do, given all the pitch-limit stuff-they simply don't let them go 10 even if they have been on their A game all night and still are only at c. 110 pitches.

When was the last time a starter went more than 9 innings?

Cliff Lee did it April 18th of this year: 10 shutout innings, Phillies lost 1-0 in 11 (Cain pitched 9 shutout innings for the Giants). Game Score of 85 (only 7 K's).

Halladay pitched 10 shutout innings against the Tigers in 2003. He had only used 99 pitches and was set to come out for the 11th but Bobby Kielty won it for the Jays in the bottom of the 10th.

When was the last time a starter went more than 9 innings?

Cliff Lee did it this year

Dean Chance against the Yankees on June 6, 1964, had the highest Game Score on record—116. 14 innings, 3 hits, 12 strikeouts, no runs.

and Bobby Richardson went 0 fer 6. Natch.

Dean Chance against the Yankees on June 6, 1964, had the highest Game Score on record—116. 14 innings, 3 hits, 12 strikeouts, no runs.

and Bobby Richardson went 0 fer 6. Natch.

but you gotta admit Repoz, his ass was magnificent

Halladay pitched 10 shutout innings against the Tigers in 2003. He had only used 99 pitches and was set to come out for the 11th but Bobby Kielty won it for the Jays in the bottom of the 10th.

Halladay's actually done it twice- the second was in 2007, also against the Tigers. Game score of only 78.

The other two pitchers to have done it since 2000 are Mark Mulder vs the Astros (in his last healthy season, 2005, game score of 87) and Aaron Harang, 2007, game score of 84.

Bartolo Colon didn't make it to ten complete innings, but he pitched into the tenth in 2005 against the A's.

For a second, I saw the header: "James: The Greatest of Games" and thought it was something about LeBron James and the 2012 NBA postseason.

Fun article, and more fodder for the greatness of that mid-90s to mid-00s group. Randy Johnson!

I noticed that too. I sometimes wonder if what's happening is that truly great pitchers would dominate in any circumstances, and therefore those who happened to pitch through sillyball are actually having their dominance magnified by the surrounding run context rather than masked by it. In other words, is it really a fair assumption that if you took Sandy Koufax and dropped him into the 2001 National League, he wouldn't pitch anywhere near as well as Randy Johnson? The way these advanced pitcher evaluations skew towards the Martinez, Maddux, Clemens, Johnson cohort along with their strong supporting cast of 2nd-tier HoFers like Smoltz, Mussina, Schilling, Brown, Glavine, Pettitte et al makes me think something is up here.

In light of all the Clemens threads recently, we should make note of this:

In 1996 Roger Clemens was 10-13 with a 3.63 ERA, and was criticized by his General Manager for being out of shape and not ready to pitch. This method shows Clemens to have been still the best pitcher in the American League (1. Clemens, 2. Pat Hentgen, who won the Cy Young Award, 3. Juan Guzman, 4. Kevin Appier, 5. Ken Hill.) The American League ERA in 1996 was 4.99, and the Park Factor in Boston was 113. Clemens raw ERA was the sixth-best in the American League.

Is it really a fair assumption that if you took Sandy Koufax and dropped him into the 2001 National League, he wouldn't pitch anywhere near as well as Randy Johnson?

I think you have a good point, but it goes both ways. If you put Pedro and Maddux and Johnson in the 1960s their numbers would be just as ludicrous as Koufax and Gibson. Can you imagine 1999-2000 Pedro in 1968? He'd have an ERA of 1. Fewer strikeouts probably, though still probably at the top of the league.

As for Clemens, does anyone from the dawn of sabrmetrics know what the word was on him back in 1996? I seem to recall hearing tell about how even back then people thought his 1996 was pretty good and that the Boston media were abominable to him. Of course, he seems like the guy that it's pretty easy to be abominable to.

There’s a simple pattern that predicts the data: each point of Game Score is 25 points of winning percentage. A four-point advantage in Starting Pitcher Game Score give the pitcher’s team a 100-point advantage in winning percentage.

To be clear, what he's actually saying is it's .025 WP, not a 25% improvement in chance to win. That confused me for a moment. I'd love to see the distribution of the curve, I'm kind of surprised that the biggest GS mismatch is only 54. It's not hard to come up with a seemingly plausible scenario that turns out worse than the Belcher/Walk matchup in the article, such as a pitcher going 7 IP with 10K, no walks, 1 ER, 1 unearned run, two hits, and a walk (GS 76) vs. a starter who gets hammered for 5 ER over 1.2 IP with 7 hits and 1 walk (GS 20) for a gap of 56. Just looking at play index, there've been 98 starts with a GS of 20 or worse and 122 of 76 or better. Being down 5 at the top of the 8th as the home team gives you only a 7.1% chance to win, so that's probably the hard part. Assuming* the (un)likelihood of a bullpen giving you 7.2 IP of 0 run ball, if you had 10-15 of those matchups you'd expect there to be a win for the pitcher with little to no chance.

I guess it's that 4 unlikely things have to happen to happen to get it to work- you need someone to pitch an excellent game, you need their opposite number to get shelled and pulled very early, you need the other team's bullpen to be lights out to keep the difference close, and then you need a big late inning rally against the good starter's bullpen. Each of those probably happens on a fairly regular basis, but having them all happen together is the real problem. Still, Given that there's been 2387 match-ups where you've had a 56 or greater GS difference, I'm surprised that there hasn't been one.

Really, really nifty article.

*Very much an "assume the can-opener" assumption.

**edited for clarity.

I think you have a good point, but it goes both ways. If you put Pedro and Maddux and Johnson in the 1960s their numbers would be just as ludicrous as Koufax and Gibson. Can you imagine 1999-2000 Pedro in 1968? He'd have an ERA under 1. Fewer strikeouts, though still probably at the top of the league.

This was the subject of intense debate a few weeks back. I'd sum up my thoughts on this by saying I think it'd be extremely likely that Gibson, limited to 210 innings, would be a jaw-dropping phenomenon. I think Pedro might face some difficulty in making 34 starts and pitching 300 innings.

As for Clemens, does anyone from the dawn of sabrmetrics know what the word was on him back in 1996? I seem to recall hearing tell about how even back then people thought his 1996 was pretty good and that the Boston media were abominable to him. Of course, he seems like the guy that it's pretty easy to be abominable to.

I probably have written about Clemens 1996 on this site a dozen times or so over the years. I'm not anyone but I was heavily active on the old AOL Saber boards that Bill James used to occasionally visit. I owned Clemens on a fantasy team at the time and my feeling was that he was as good as he ever was but suffered tremendously from bad luck and bad teammates:

1. The 1996 Red Sox had about as bad a defensive team as you can imagine. You basically have 5-6 DHs on the team. The catchers were Mike Stanley and Bill Haselman. Mo Vaughn at 1B. Jeff Frye at 2B; Tim Naehring at 3B; John Valentin at SS. The left fielders are bad: Mike Greenwell, Troy O'Leary and Reggie Jefferson. Center Field was OK, I think: Tinsley, Bragg, Cuyler, though not a Gold Glover among the bunch. RF was comical: O'Leary, Kevin Mitchell, Bragg, Rudy Pemberton. If I'm adding this up right, BBREF has the team at -77 runs on defense.

2. If you really need proof that the defense was bad, look at the ERAs of the other pitchers on the staff. The rotation was Tom Gordon, Aaron Sele and Tim Wakefield behind Clemens. Those three guys are all fine major league pitchers who had good, long careers. All were between 26-29 years of age, none were hurt. ERAs: 5.59, 5.14, 5.32. Jamie Moyer had a 4.50 ERA.

3. The ERAs for other guys on the team were downright frightening. Stan Belinda had a 6.59 ERA, twice his ERA the year before. Jeff Suppan 7.54. Vaughan Eshelman 7.08.

It was pretty clear to me that Clemens was keenly aware of the circumstances surrounding the team, and tried to make perfect pitches and strike everyone out. This had the predictable effect of leading to more walks than usual as well as a lot of strikeouts. And because he was on a team with no defense/pitching/bullpen, he was left in games too long, too often. Clemens had a 7.79 ERA after the 7th.

As you note, Clemens was the same pitcher in 1996 but the only reason it wasn't apparent is that he was in more difficult circumstances to succeed. Doesn't that clash with the assumption that Pedro Martinez was likely to have buckled under a Bob Gibson workload? A 1968 NL inning was not a 2000 AL inning. Just as Gibson's identical athletic performance would have been compressed into a smaller space if he'd been a 2000 Red Sox pitcher, while remaining the identical performance, transporting Martinez's identical performance to 1968 would have resulted in many more than 217 IP.

Thanks, baudib. That's enlightening, on both accounts.

No, it doesn't clash with that assumption at all. I'm well aware of the fact that conditions in the AL in 2000 were different than the NL in 1968.

By any real objective measure, Clemens was terrific in 1996. His ERA+ was 139. He led the league in strikeouts. He had a 20-strikeout game. He was a great pitcher, period. If you're looking for evidence that Roger Clemens was in decline, 1996 isn't what you're looking for. Beyond that, Clemens was actually far better than his 139 ERA+ suggests.

Incidentally, I never said he was the "same" pitcher, whatever that means. 1996 probably isn't one of Clemens' 10 best seasons. He was better in 1986-87 and far better in 1990-92 and far better in 1997-98.

I fail to see how acknowledging that Clemens was far better than a 10-13 W-L record suggests in 1996 translates into, "Pedro would totally pitch 300 innings in 1968."

Pedro doesn't suffer from the same problems with the Red Sox defense in 2000. BBREF has the 2000 Red Sox defense as +38, so that's a 115 run difference between the teams. Pedro wasn't durable even by the standards of his time. He missed starts every year. 2000 isn't even a durable year for Pedro standards. He made only 29 starts, he faced only 817 batters.

For most pitchers, moving from AL 2000 to 1968 would mean they would throw more innings. I don't think this would be the case for Pedro specifically. Most pitchers would allow fewer baserunners. Pedro allowed so few baserunners that he really can't go much lower. He had a .237 BABIP that year, and walked only 32 batters. Let's say he lowers that to .200 and 20. That's 29 fewer baserunners. He allowed 17 homers that year, let's say facing weaker lineups he allows 7. That's 39 fewer. If all of those baserunners turn into outs that's only 13 more innings.

Pedro in 2000 used only 14.6 pitches per inning. You can't go much lower than that, especially when you need at least 9 to strike out the side.

Pedro just doesn't have much room to improve. Let's say Strasburg watches R.A. Dickey, learns a knuckler which he can control perfectly, throws it 97 MPH, and pitches perfect games every start against MLB hitters. How much better would he be facing minor leaguers? How much better against my DC softball team (which does not include Bryce Harper)? The numbers would be exactly the same.

Could Pedro have pitched in a 4 man rotation and made 37-40 starts? I don't think he had the durability to hold up to that kind of workload.

I sometimes wonder if what's happening is that truly great pitchers would dominate in any circumstances, and therefore those who happened to pitch through sillyball are actually having their dominance magnified by the surrounding run context rather than masked by it. In other words, is it really a fair assumption that if you took Sandy Koufax and dropped him into the 2001 National League, he wouldn't pitch anywhere near as well as Randy Johnson? The way these advanced pitcher evaluations skew towards the Martinez, Maddux, Clemens, Johnson cohort along with their strong supporting cast of 2nd-tier HoFers like Smoltz, Mussina, Schilling, Brown, Glavine, Pettitte et al makes me think something is up here.

Game Scores like strikeouts :)