15 September 2004

"...I Knew Lefty, and Sir, You're No Lefty!"

Much attention this season has been paid to the plight of Arizona's Randy Johnson.

randy Posted by Hello

Naturally, it's hard to think of anything a world-famous man making bajillions of dollars a year to play a children's game in front of thousands of adoring fans would experience as a "plight". Nevertheless, it must be something of a drag to see your best efforts kicked to the curb like yesterday's hot dog wrappers, or, perhaps more poignantly, like an infield grounder to Chad Tracy.

ESPN Radio's morning show Friday mentioned that The Big Unit would pitch Friday night against San Francisco, and that despite his losing record, which stood at 12-13 entering the game (Note: Johnson won that game, pitching seven shutout innings to even his record at 13-13.) Mike Greenberg, a ~35ish sports reporter, proffered the idea that a case could be made for Johnson as the best pitcher in the NL this season, and that the only reason for his lackluster record was poor run support from a poor team, one that has already lose three more games to this point in the season than the Philadelphias did in '72, with three weeks yet to play. I tended to agree.

But Bob Pecose, a ~55ish newsman and self-declared purist, proved unwilling to concede the point. Bob still believes in Wins and Losses as a useful measurement tool for pitchers, and mentioned in support of his position the case of Steve Carlton, who somehow managed to go 27-10 and won the Cy Young Award for the 1972 Phillies, who lost 97 games for the worst record in the NL that year. Greenberg wondered aloud whether it might be determined that Carlton's record was indebted to unusually high run support, and suggested that they ought to look into it some time when they weren't so busy.

So I did.

I got to thinking about how this could possibly have happened, and whether or not something can be understood about the nature and/or usefulness of pitching wins and losses from comparing 2004 Johnson to 1972 Carlton. There are several factors here to consider here when cmparing two pitchers across eras like this.

Lefty Posted by Hello

For one thing, the starting pitcher's role has changed significantly in 32 years. Steve Carlton pitched 346 innings, while starting 41 games that year. He completed 30 of those starts, eight of which were shutouts. He easily led the league in starts, innings, complete games, ERA, strikeouts, and some more obscure stats. Johnson is second in the NL in innings pitched this season, but probably won't come within 100 innings of Carlton's total. At the rate at which Johnson consumes innings, he would have to start 50 games to equal Carlton's 346 innings pitched, and he would win 21 games...but would also lose 21. Carlton's 30 complete games were seven more than the next closest pitcher in the NL, but there were ten other guys with at least 13 complete games. This year Livan Hernandez will likely lead both leagues in that category without getting into double digits.

Carlton's 1.97 ERA was the lowest in the NL that year, whereas Johnson's 2.75 is currently pacing the NL, nearly a run higher. The NL ERA in 1972 was only 3.45, making CArlton's league-leading 1.97 was 75% better than the league average, whereas Johnson's on "only" 58% better this season, because as runs become more scarce, they also become more valuable.

This factor surprised me the most in my research. The 1972 NL average team scored 3.91 runs per game, as compared to 2004, when the NL is averaging 4.7 runs/game right now, again, almost a run higher. Carlton's horrendous team scored 3.22 runs per game on average, and about 3.28 when Carlton pitched, roughly 83% of the NL average. This was the big surprise for me. I figured that if he won 27 of the team's measley 59 victories, it must have been because they scored more runs when he pitched, but that wasn't the case. They scored almost exactly the same number of runs per inning innings in Carlton's starts as they did for anyone else, on average, which was substantially less than the NL average.

Again by contrast, the 2004 D-Backs average 3.88 runs per game, 3.92 when Johnson pitches, pretty much the same. The NL average this year is about 4.67 runs per game, which means that the 3.92 runs they score for Johnson are (wait for it...) 83% of the NL avearge! It should be noted that the D-Backs score slightly more than a run over the number of earned runs Johnson allows, but also that runs become "cheaper" as scoring increases. Johnson has allowed ten unearned runs this season as well, which hasn't helped. Carlton only let eight unearned runs score all season, while pitching almost 130 more innings than Johnson has to this point.

So clearly the expectations have changed. Starting pitchers rarely complete their starts, and if they allow three earned runs in six innings, it's called a "Quality Start" even though a pitcher who did that every time out would have a 4.50 ERA, which is basically mediocre.

That's the background you need to understand before we can really say anything about how good Randy Johnson is or is not. First, let's look at how well Johnson has pitched this year:

He's pitched 215 innings, starting 31 games so far and completing four. Both the innings and complete games totals are second in the NL to Hernandez. Of those 31 starts, 22 have been "Quality Starts" meaning that he pitched 6 innings or more and allowed three earned runs or fewer, theoretically giving his team a chance to win. Except that it doesn't win, because eight of his 13 losses have come in these so-called "Quality Starts." The team has scored a grand total of seven runs in those eight starts, and no more than two runs in any one of them, so Johnson is only 10-8 with four no decisions in those QS. Not surprisingly, the Diamondbacks proceeded to lose all four of those no-decisions as well.

In 1972 a Quality Start would have been assessed differently. The NL ERA was 3.45, so an appearance which leads to a 3.50 ERA might be aptly called "Quality", such as 7 innings and two earned runs, or eight innings with three earned runs. Carlton had 30 such appearances that year, and went 25-4 with only one no-decision, a 10 inning shutout that was lost when the first batter in the 11th inning homered off the relief pitcher who replaced the pinch runner who replaced the pinch hitter who replaced Carlton in the top of the inning. If I were Steve, I wouldn't ever give up the ball either.

One reason Carlton won so many games is that he almost never gave up the ball. He completed nearly 3/4 of his starts, was only replaced in the middle of an inning once all season, and even pitched the tenth and eleventh innings in three of his starts. As I mentioned, his run support was just about as feeble as Johnson's is, but he allowed so few runs to score, and refused to let a relief pitcher screw up his work unless he absolutely had to, that he managed to win.

Eight of his games, as I mentioned, were shutouts. Nine other times he pitched a complete game and allowed only one run, earned or otherwise. Five of his other complete games entailed only two runs by the opposition, and two of those were 11-inning jobs. In adidtion, he also had that 10-inning shutout bid ruined in the top of the eleventh, as I mentioned earlier. So that's 22 of his 30 complete games that included 2 runs or fewer, not all of which were wins. Johnson has only had 16 games of any length this year allowing 2 or fewer runs, and only seven of those have been eight innings or more, none more than nine. Pitchers just can't do the same type of job with any kind of consistency.

After 31 starts in 1972, Carlton was 20-7, with a 2.10 ERA and had just completed nine straight and 14 of 15 starts, including two 11-inning, 2-run jobs. And though he had pitched about 50 more innings than Johnson in those 31 starts, he had only one more decision, with seven more wins and six fewer losses. Carlton was, more often than not, so stingy with runs, and so reluctant to surrender the ball, that he usually managed to win.

In 1971 and 1973, Carlton pitched very similarly. He had an approximately league-average ERA, and pitched a lot of innings. But the 90-win, '71 Cardinals gave him enough run support to provide a 20-9 record at year end, whereas the 91-loss Phillies let him take it one in the "L" column 20 times in 1973. The difference in run support was less than two runs per game.

And for the most part, Johnson has done everything you could ask from a pitcher: he gets batters out, prevents runners from scoring, fields his position, etc. What he doesn't do is bat, at least not very well. Hey, if you had a strike zone the size of a screen door, you might have a little trouble making contact too. But Johnson's .114 average, with one Run scored and 4 RBI, certainly doesn't help his cause much. Carlton, though not exaclty Babe Ruth with a batting average that echoed his ERA (.197), did have 5 extra base hits, including a homer, 6 runs and 8 RBI. There were four times that season when a run that he scored or drove in made a difference in the game, four wins that (presumably) would have garnered the Big Unit a Big L.

Now we get into "what if" scenarios? What if Johnson got league-average or better run support? Well, looking at his game log, if the team could score runs to average out to 4.67, depending on how things worked out, Johnson might still only be something like 16-12, or with some breaks, 17-10, which would certainly put him under serious consideration for the 2004 Cy Young Award, but still isn't anywhere near as good as Carlton was.

Amazingly, though, Carlton did not win an inordinate number of games given his run support. The Pythagorean projection for his ERA and Run Support

[ERA^2/(ERA^2 + RS^2)]

suggests that 27-10 is exactly where Carlton's record should have stood in 1972. His team barely scored more than 3 runs per game, but he allowed only two, and didn't allow a relief picher to get in the way of victory. The margin for error was so close, but Carlton managed to be ever-so-slightly within that margin 27 times that season. Johnson's just had hard luck and bad relief pitching, as Pythagoras tells me that he should be 17-9 at this point, not 13-13.

This seems to make it fairly clear that though Johnson may be the best or one of the best pitchers in the league, he's still nowhere near as good as Carlton was that year. I doubt anyone would really argue with this. But you don't win a CYA by proving that you're better than some guy who pitched thirty years ago, you win by being better than anyone else pitching in your league now, theoretically. People like Bob Pecose and Joe Morgan need to wake up and smell the 21st century. Pitchers are asked to get batters out, in an effort to try to win games for their team. If the team doesn't score and they lose, 1-0, it's hardly the pitcher's fault. The fact that he has 13 wins on a team that might not win 50 is fairly impressive to begin with.

Let's not put a damper on his accomplishments by complaining that he's not another person in another time. That's a standard against which no one should have to be measured.

Stumble Upon Toolbar