The point about it depending on what you're asking for is a good one. If it's one game, I won't object to Marichal.
Marichal topped out at 10.3 Wins Above Replacement (WAR), which is very good, comparable to Mike Trout's rookie year. He managed that in 295 innings, which gives him .035 WAR/inning.
Perry's best year was worth 11 WAR, in 1972. He pitched 342 innings that year. And so comes in at .032 wins per inning. At his best, very close to Marichal.
Niekro's best season was 1978, when he managed 10 WAR on the nose, in 334 innings. Comes out to .03 wins per inning. Slightly behind Marichal.
So yes, for one game, Marichal is the right answer. Perry isn't far behind.
For one season, Perry is slightly ahead, but it's pretty close.
But if you want to know who had the best career, that's when Marichal isn't in the same class as Niekro or Perry. Basically, he just didn't pitch enough innings to have had a comparable career. Our loss, a 5000 inning career from Marichal would have been something to see.
Finally, a word of warning about using FIP in this context. Fielding independent statistics are predictive statistics, and they carry more significance (relative to ERA or similar measures) the fewer innings a pitcher has already pitched. FIP assumes that pitchers have no control over what happens to a ball once it's put in play. Most of the time this is a safe assumption: the year-to-year correlation between balls-in-play outcomes is very weak for most pitchers, and it takes a really long time to figure out which pitchers do have an ability to induce weak contact and so affect what happens to balls in play. But when we're talking about pitchers with 5000 innings pitched, it's not a safe assumption any more. After 5000 innings we know pretty well if a pitcher can control what happens on balls in play. For these kinds of discussions a statistic that takes ordinary ERA and adjusts it for context (like ERA+) is a better choice. (ERA+ compensates for the era that players played in, as well as their home park. Since scoring was higher in Colorado in 1999 than in Chavez Ravine in 1968, an ERA of 4.00 will give you a much better ERA+, better than average in fact, in Colorado in 1999 than for someone playing for the Dodgers in 1968.)
By ERA+ Perry and Niekro are basically tied. (Perry has a 2 point lead, which is essentially nothing.) Marichal does have a better ERA+ than the other two, although in far fewer innings. Marichal had an ERA+ or 126 in 3500 innings. Over his first 3500 innings Perry's ERA+ is about 124. So, basically tied. Niekro's, over his first 3500 innings, is also 124. Higher ERA+s are better.
|