Comparing players between historical eras part two: pitchers

In my last article, I introduced a method of comparing batters from different historical eras to each other, by attempting to measure how dominant they were in the time period they played in. Here’s the pitcher counterpart.

To recap, rather than using the usual “plus” stat (like OPS+ or ERA+) approach, I tried using a normalized standard deviation-based method, the z-score. “Plus” stats use simple percentage points, which work fairly well, but if the distribution of talent changes over the years, players from certain time periods will get an unjustified boost or an unfair punishment. In the deadball era of the 1910s, hitting performance fell, coupled with a decrease in the spread of talent. Not only did hitters perform worse, they also fell closer together in talent to each other. OPS+ will adjust for the deadball environment as a whole, but it won’t adjust for the change in the talent spread.

For pitchers, I took every qualified player from 1871 to 2011 and calculated each year’s league average and standard deviation for FIP. Graphed over time, it mirrors the changes in wOBA fairly closely. Click to enlarge.

image

And here’s the top ten pitcher seasons by z-score.

NameSeasonZ-Score
Pedro Martinez19993.64
Dwight Gooden19843.05
Pedro Martinez20002.73
Randy Johnson20012.66
Bob Gibson19702.56
J.R. Richard19792.49
Randy Johnson19952.48
Nolan Ryan19872.42
Pedro Martinez20032.39
Mike Scott19862.38

Some thoughts:

  • Pedro, Pedro, Pedro. By basically any metric, Pedro Martinez’s 1999 season ranks as one of the greatest full seasons by any pitcher in baseball history, but this really helps put it in perspective. His z-score is the greatest out of any qualified pitcher, and it’s not close.
  • Interestingly, Cy Young doesn’t appear on the list until rank 88, with his 1905 season. He’s exactly the kind of pitcher who gets a little boost with normal “plus” stats as I mentioned before. He was at his best in the middle of the deadball era, when the spread in batting talent fell to its lowest level in baseball history. The change in batting performance is coupled with a change in pitching performance, as not only did pitchers get better in the deadball era, they exhibited a wider spread of performance. Cy Young’s best seasons came when pitchers were more spread out, so his raw numbers indicate that he dominated the deadball era more than he actually did.
  • The top 10 is a lot more varied than the top 10 batters list (Bonds, Ruth, Bonds, Ruth, you get the idea). No pitcher appears on the top ten more than three times (Martinez), and no one appears on the top 20 more than four times (Randy Johnson).
  • With a FIP of 5.21 in 1960, Mudcat Grant holds the worst z-score for a qualified pitcher, at a staggering 3.48 standard deviations below the average. The worst semi-recent season would be Matt Keough in 1982, with a FIP of 5.88.
  • The last 20 game loser was Mike Maroth of the Detroit Tigers, who went 9-21. He actually doesn’t fare too badly by the z-score method, sitting only 1.67 standard deviations below the average.
  • Like with the batters, I’ve provided a sortable leaderboard in the References and Resources section. Play around with it however you wish, and feel free to use it for anything you’d like.

References & Resources
Leaderboard here. Like last time, click over to list view if you want to sort it through Google Docs, or you can just download it and play with it in Excel or any other spreadsheet software. All data from Fangraphs.


11 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Marc Schneider
12 years ago

Fascinating article.  Where would you consider the cutoff for an “elite” season.  I understand the top ten are truly extraordinary, but if you were going to go down farther, how far would you go.  Top 50?  Top 100?

Chris Brune
11 years ago

Is it possible that there is a bias that favors power pitchers?  Or a bias that favors recent pitchers?  The entire TOP 10 fit into both of these categories…..

A Mets Fan
11 years ago

Are you sure you don’t mean Gooden’s ‘85 season?

Detroit Michael
11 years ago

Because Pedro Martinez often pitched fewer innings each season than his contemporary elite starters, and even more so compared to the starters of yesteryear, doesn’t that make it easier in a sense to have higher Z-scores (both for better and for worse)?  Pitching more innings makes it more difficult to maintain a truly exceptional Z-score.

It’s sort of like how we know that it is easier for a MLB batter to bat .400 in 100 PA than it is over 600 PA.

Dan Lependorf
11 years ago

@Marc

Well, Top 100 is an average of a little more than one “elite” season every year and a half or so, which seems kinda high to me. Maybe top 50 is a better bet.

@Chris

It would simply depend on the stat used. I used FIP, but you could just as easily use ERA, tRA, or even pitcher wins if you really wanted to for some reason.

@A Mets Fan

Nope. Gooden’s ‘85 season was really good, but not quite as good as ‘84. I have Gooden’s ‘85 season at rank 36. I used FIP, which likes 1984 Gooden more than 1985, even though ‘85 Gooden had the ridiculous ERA.

@Detroit Michael

Yes, it would, but I used a cutoff of the usual qualification threshold of 1 IP per team game. So you have to have 162 IP to qualify for this list. Pedro in 1999 had 213.1 IP, which is pretty normal for a full-time starter.

rubesandbabes
11 years ago

With all the modern guys, seems like Maddux is missing. And Eckersley (and perhaps even Koufax, too).

Mike Scott was a ball doctor. Ha!

Nolan Ryan’s 8 – 16 1987 year showing up so high tests this list’s cred.

rubesandbabes
11 years ago

Yes, Ryan’s 8-16 1987 does not compare to Hershiser’s 1988, where Hershiser goes 23-8 with one save.

Hank G
11 years ago

Ryan led the National League in ERA in 1987, so the high ranking is not quite as ridiculous as it looks at first glance. Whether it deserves to be ranked as high as it is, I can’t say, but he pitched much better than the W-L record would indicate.

Dan Lependorf
11 years ago

Nolan Ryan’s 1987 season had not only a fantastic ERA, but a ridiculously low FIP. That’s why he ranks so high. The poor W-L record is irrelevant.

Robert Haymond
11 years ago

Bob Gibson had an amazing year once (well, more than once) with the lowest ERA and many shutouts while winning over 20 games.  where does that amazing season set?  I am a bit baffled by all the extraordinary seasons by modern ballplayers.  Why are they favoured?  Are they just better, as a whole, than those of yesteryear?  If so, why?

Michael
11 years ago

Great stuff. Over at SBNation, Rob Neyer wrote about the difference in WAR leaders, with many playing before WWII. I would be interested in seeing something like this for WAR.