Wit and Wisdom, Dept. (Redux)
When do we swap out Blake Beavan for James Paxton?


Reprinted from May 24, 2013.


At BJOL two weeks ago, James published a fun-and-fascinating discovery about baseball.  It was the kind of discovery that could easily be one of a franchise's Ten Core Values -- you know, sign on the wall, We Believe In ... college pitchers, position scarcity ... yada yada ... This One.  

And yet, it was just buried in the letters column.  Another day, another bedrock insight about baseball.  And it goes exactly to the issue that Gordon brought up in his last series.


Just the Fa'ax, Ma'am

We'll publish the letter below our sigline ... right now we'll give the Exec Sum.   And this article is just an opener -- a chance for you to ask, "Now that I know that, what implications are there here?"  We'll follow on that later.

The facts:

A reader asked James, "If you have a rookie pitcher with a GREAT season -- Fidrych, Kerry Wood, Candelaria, etc -- does that guy go on to a better career than a rookie pitcher with a GOOD season?  Maybe we need to be skeptical of splashdowns like Strasburg's and Harvey's."

James replied by doing a study.  He found two things.  First, he found, YES, if a rookie pitcher has an epic season, it is one signal that he's going to be a great pitcher.  

But second, he found that except for Epic Rookies, it doesn't much matter how good a rookie pitcher's season is.

:: he stops short ::

If that were true, what decisions would we make as a result?!


The Study

James took every rookie pitcher from 1960 to 1990, who started about 30 games in their rookie years.  (Actually anywhere from 20 to 40.)  This was 327 rookie starters.

Then he cut them into 10 layers, from best ERA's to worst ERA's.   And he asked how each layer did for the rest of their careers.  This is what he got...  the ERA's after Tier 1, unless they have an *, are rounded for clarity, and to illustrate the study.  If a number does have an asterisk, that ERA was specified by James in the reply.


Tier Rookie ERA, first 20-40 games Rest of Career
1 (Gooden, Guidry, Rozema, Zachry, etc) 2.81* 3.74*
2 3.20* 4.18
3 3.40 4.18
4 3.60 4.18
5 3.75 4.18
6 3.90*


7 4.15 4.18
8 4.40 4.18
9 4.75 4.18
10 5.00 4.18


So the "superstar" 10% did do better, a lot better.  But the other 90% "regressed to the mean" -- in essence, it didn't much matter whether your ERA matched Michael Pineda's or Blake Beavan's; you projected to have the same career from then on.

Those of you who are sharp-eyed** will see some interesting caveats here; for example, all 327 pitchers, their managers believed in them enough to start them 30 times.  But still.  That is a lot of innnings, and a very interesting effect.

And the implications are thunderous.  We followed on with a question to James about organizational philosophy, and his reply was very important.


Dr D

**that's everybody



Matt Harvey-inspired question, which we've been batting around in "Reader's Posts": Are the careers of young superstar pitchers (after their rookie years) any better than the next-best group of young pitchers? IOW, is it an advantage for someone to get his career off to a dominant start or (apart from the rookie year) is a pretty-good rookie season just as impressive in the long run? At first glance, you'd think that the guys who are very young and very successful immediately would have an advantage, but so many young pitchers (Bunker, Fidrych, Candelaria, McDowell, Kerry Wood etc.) who get off to hot starts don't have much of a career, while some who struggle for a few years do have long successful careers, I wonder if we need to be skeptical of the long-term virtues of a rookie season such as Harvey seems to be having, or Strassberg had.
Asked by: sgoldleaf
Answered: 5/10/2013
Well, thank you for the question, and, not to blow too much sunshine up your skirt, but to be able to take on questions of this nature is really the goal and purpose of this feature.   I appreciate questions that
a)  require research, and
b)  lead to gains in understanding once the research has been done.  
That's the definition of a good question.
Anyway. .. .I took all rookie pitchers in the years 1960 to 1990, minimum of 20 starts, maximum of 40 starts at the conclusion of the season and maximum of 50 game appearances at the conclusion of the season.   There were 327 pitchers in the study.   I sorted these into 10 groups, 33 pitchers in Group 1, 33 in Group 2, etc. . ..some of the groups had to have 32, obviously.    Then I looked at the "rest of career" and "end of career" results the ten groups.      Fidrych and Bunker were in Group 1. ..Kerry Wood is too late for the study.  Candelaria doesn't qualify for the study because he made 18 starts as a rookie, and I don't know which "McDowell" you mean, in that both Jack McDowell and Sam McDowell had lousy rookie seasons.   Group 1 includes Fernando Valenzuela, Doc Gooden, and Ron Guidry, but also Wayne Simpson, Dave Rozema and Pat Zachary.
Anyway, the conlusion is that ONLY group 1 has a meaningful "rest of career" advantage over the other pitchers.    Group 1 had a rookie-season average won-lost record of 15-9, 216 innings with a 2.81 ERA.    In the rest of their careers they had an average won-lost record of 103-90, 3.74 ERA, which is quite a bit better than any of the other groups.
However, interestingly enough, once you get out of Group 1, then one group is pretty much the same as the next in terms of rest-of-career performance.    Group 2, as rookies, was on average 13-9, 198 innings with a 3.20 ERA.   In the rest of their careers they were 61-60, 4.08 ERA.       Group 6, as rookies, was just 9-10, 158 innings with a 3.90 ERA, but in the rest of their careers they were 57-62, 4.18 ERA--not really very much different than Group 2.       So. . .not only do the "superstar" rookies have a long-term advantage, but they are actually the ONLY group of rookie pitchers who have any long-term advantage.




Off the charts neat!
What it doesn't mean is that you can roll any MiLB thrower out there for 30 games and then he becomes, voila, a decent major leage pitcher. You anticipated that point, but we need to be careful here.
But it does mean that a guy like Maurer, clearly with MLB starter stuff, is just as likely as Joe Saunders to have a Joe Saunders type career.
Greg Maddux had a 5.61 ERA, a 1.64 WHIP, walked 4.3 while only striking out 5.8 p/9 and a 6-14 record in his rookie year.
He got better.
Worth remembering.
M. Pineda's rookie season would certainly be in Tier 1. Is he going to have a prettier career than Maurer? A whole bunch of Pineda's would become, on the average, better than a whole bunch of Maurers. But that is about the only guarantee.
Maybe I like this data because I'm a believer in an accelerated process for young talent. Get 'em in the bigs and let 'em figure it out.
Paxton next.


James lumped these pitchers by ERA, right Doc? Because Michael Pineda had a 3.70ish ERA in 2011. That's tier 5...which means he does NOT have an increased chance at success.
I think there's also something missing from this study. How many innings did each group get to pitch for the rest of their career? I'll bet the high ERA rookie group has a worse attrition rate.


But ya, Pineda would have been OUTside Tier 1, even adjusting for era.
The first Q that comes in here:  are there differences in talent that are there, but which are NOT showing up in rookie results (because the various tiers have a random mixture of talented and non-talented pitchers)?  Or are differences in talent (tier 3 = talent, teir 8 = no talent) being overwhelmed and swamped by the experience factor?
Is it that the "good" rookie pitchers were NOT YET more EFFECTIVE than their bad rookie counterparts (and so finding themselves sprinkled across layers as rookies)?  Or is it that they ARE more effective (and are bunching in layers 2 and 3), but that they lose this effectiveness gap as time goes on?
James' next follow-on is quite interesting - 

Add comment

Filtered HTML

  • Web page addresses and e-mail addresses turn into links automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <blockquote> <code> <ul> <ol> <li> <dl> <dt> <dd><p><br>
  • Lines and paragraphs break automatically.

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.


  • Allowed HTML tags: <a> <em> <strong> <cite> <blockquote> <code> <ul> <ol> <li> <dl> <dt> <dd>
  • Lines and paragraphs break automatically.
  • Web page addresses and e-mail addresses turn into links automatically.