And Sandy Koufax

A few years ago, I wrote an article called The Best Pitchers of All Time. It was an admittedly grandiose title, one intended to get lots of hits from search engines and that sort of thing.

The article itself was a vehicle for rolling out a variation of Bill James’ Win Shares system I had created, called Win Shares Above Bench. Previously, I had rolled out a couple of articles ranking the best hitters of all time, but the titles of those pieces were pretty boring.

Evidently, my titling strategy worked because I still get emails about that pitching piece (I never get emails about the hitter rankings). But, mostly I still get emails about the article because I left Sandy Koufax off my list of the 40 greatest pitchers ever.

The response is understandable. From 1962 through 1966, Koufax was 111-34 with a 1.95 ERA. He finished first in ERA all five years and won three Cy Young awards. He won one MVP and finished second in MVP voting twice. He was also terrific in the postseason, with a 0.95 ERA. He was dominant.

Still, I left the guy off my top 40 list, and I feel I owe people an explanation. Instead of responding to emails individually, I’ve decided to just post my reasoning here and refer to it in the future. It will save me a lot of typing.

So…where to begin? Well, one commenter told me I never would have left him off the list if I had actually seen him pitch. Let me assure you that I am an old man. I saw Koufax pitch. He was indeed a great pitcher in the early 1960’s—when I was at my most impressionable, to boot—but my article wasn’t about impressions. It was about a new metric that I was pretty proud of (still am, in fact). And in Win Shares Above Bench, Koufax not only doesn’t rank in the top 40, he’s 91st.

Crazy, you say? Well, maybe. But other, comparable systems also rate Koufax below the top 40. Sean Smith’s WAR rates him 61st all time. In the 2007 THT Annual, David Gassko rates him 64th. If you want to argue that WSAB is wrong and Koufax should be 30 places higher, I can accept that. But 50 places higher? Nope. Here’s why.

One. Sandy Koufax pitched in a very pitcher-friendly environment. Here is a graph of the average number of runs scored per game throughout baseball history, with Koufax’s five prime years highlighted:

image

Teams scored an average of 4.09 runs per game in Koufax’s prime. The overall major league average has been 4.39. Plus, he pitched in Dodger Stadium, a notorious pitcher’s park at the time, with a pitching park factor of .91. So the effective run environment in which Koufax pitched was .91 times 4.09, or 3.72 runs per game—more than half a run less than the historical major league average.

To hammer home the point, Koufax’s ERA in Dodger Stadium was 1.37. Not for one year, for his career. You could even go so far as to say that the ballpark was primarily responsible for the first of his five superb years (1962).

So the baseline, the essential comparison point, for Koufax’s fantastic five years is different than that for most pitchers. As a final consideration, look at Koufax’s ranking in ERA+, a measure that accounts for run environments. Koufax is 37th in all-time ERA+. Yes, that’s in the top 40. But I’m just getting started.

One(A). ERA+ is flawed in a way that helps pitchers in low run environments. In fact, this is one of my pet statistical peeves. However, it’s not a big deal and that’s all I’ll say about it.

Two. Koufax pitched in the aftermath of major league baseball’s first big expansion. On a percentage basis, the hitting talent may have been more diluted from the previous years than in any other time in history. Koufax posted a 1.90 ERA against the Astros and a phenomenal 1.44 ERA against the Mets. The next lowest team-specific ERA he posted was 2.44 (Cubs). I don’t think it’s entirely a coincidence that Koufax’s five best years occurred immediately after the league expanded and the Dodgers moved into Dodger Stadium.

Three. Koufax only had five great years. Now, five is a lot and what’s more, he had them consecutively. That is extremely impressive. In fact, if you rank pitchers by something I called All-Star Win Shares (which only considers years in which players were above average and doesn’t discount the below-average years), Koufax is 31st all time. Koufax is justifiably revered for those five fantastic years.

But he had very few “pretty good” years before that and, of course, his career ended abruptly due to arthritis. Outside of the big five, he wasn’t an above-average pitcher overall. If you don’t want to count those years against him, perhaps he should rank in the top 40. But when you do include them, it is very difficult to say that he was a top 40 pitcher.

Now, don’t get me wrong. I love Sandy Koufax. He was an amazing pitcher and, from all accounts, he’s a terrific person. But I’m the type of person who ranks people according to stats and not according to impressions. I also believe in ranking players fairly, taking care to incorporate as much context as possible and including their entire careers—not just their peaks. And that’s why Sandy Koufax didn’t make my list of the greatest pitchers of all time.

May I Have Your Autograph, Please?
The payoff of being polite.

Print This Post
Dave Studeman was called a "national treasure" by Rob Neyer. Seriously. Follow his sporadic tweets @dastudes.

Leave a Reply

30 Comments on "And Sandy Koufax"

Notify of
avatar
Sort by:   newest | oldest | most voted
Lou
Guest
Lou

Good stuff Dave.  Also, I think a lot depends upon on how much each person defines “best” in terms of career length.  I think most fans tend to ignore the bookends of each career.

John
Guest
John

Just nitpicking here, but if Dodger Stadium’s park factor was 0.91, and we assume Koufax pitched half his games there, then shouldn’t you correct by 0.955 (halfway between 0.91 and 1) to take road games into account? 

The correction may be slightly more since Dodger Stadium’s 0.91 should not be included in the mean value for the road parks.

Regardless of where Koufax ranks all time, he’s at the top of my list of pitchers I wish I had seen.

Dave Studeman
Guest
Dave Studeman

I thought Baseball Reference’s Park Factors were already calculated so that you could apply them in a straightforward manner, making the ratio of RS+RA in Dodger Stadium to RS+RA on the road roughly equal to 0.8.

Dave Studeman
Guest
Dave Studeman

A quick look at Retrosheet indicates I applied the BRef park factors correctly.

http://www.retrosheet.org/boxesetc/L/PK_LOS03.htm

Tim
Guest
Tim
I have always loved baseball because of the ability to measure different eras with stats. Stats are great but if I needed one pitcher to pitch one game, the 7th of the World Series, I’m not sure who I would pick over Koufax. If you saw him pitch, you measure his won-lost, you can argue it, but it’s a tough argument. A minor point is that Koufax pitched for notoriously weak hitting teams. He won every way you needed to win. Goes to my first point. Stats people love to talk about bookends but to me, when you say the… Read more »
Silver King
Guest
Silver King

Really interesting details about the home ERA and team-specific ERA.

Dave Studeman
Guest
Dave Studeman

@Tim, understood.  That falls into the “impression” category.

BTW, the #1 pitcher I wish I had seen is Walter Johnson.

Chris Wright
Guest
Chris Wright

I have to agree with Dave. For the first 7 years of his career, Koufax was basically a .500 pitcher with a walk rate of 5-6 per nine innings. He was certainly dominant the last 5 years when he found the strike zone but a 5 year string of dominance does not completely wipe out 7 years of mediocrity.

Now, if he’d been able to maintain the dominance on his last 5 years for another 5 years, we might not be having this conversation…

Tim
Guest
Tim
Two things. First, I wasn’t offering an impression, I was trying to make the point that some 5 year careers transcend longer careers because of their sheer uniqueness and dominance. I know this is extremely arguable but I’d take those 7 years of mediocrity to get those 5 years. And I would take that over most more consistent pitchers. Secondly and perhaps more importantly was Koufax under pressure. That is a different statistical measurement and to me, if I was a manager, more germane to the question of who is the greatest. He rarely worked with much of a lead… Read more »
studes
Guest
studes

Tim, regarding the tradeoff between his mediocre years and prime years—unless you’re willing to quantify the difference, it’s an impression, or perhaps a judgment call, but not a quantifiable difference.

I don’t know what you’re saying about pitching under pressure.  Are you saying that his won/loss record is even better than his runs allowed would indicate?  Do you know something about his run support?

He DID pitch very well in high-leverage situations, which is no surprise, really.  Is that what you’re referring to?

studes
Guest
studes
So I did a little more digging at Baseball Reference.  Based on the run support Koufax received during his five prime years and the number of runs he allowed, he has a pythagorean winning percentage of .770.  His actual winning percentage was .766.  Seems pretty much spot on, which implies that his ERA fully reflects his pitching value. BTW, the Dodgers scored 4.2 runs per each one of his starts during those five years, which was higher than the overall major league average at the time. Having said that, it does appear that he pitched to his run support to… Read more »
Eddie
Guest
Eddie
As to this debate about mediocre/dominant seasons vs. more consistency, I remember reading an article contrasting the careers of Don Drysdale and Don Sutton – the conclusion was that their careers were of similar value, if Sutton’s wasn’t more valuable overall, yet the way they distributed that value made a difference – Sutton had more consistent value over a long period of time, while Drysdale concentrated most of his value in a few years (like Koufax).  If the overall goal is to win the World Series, then, the article concluded, having a few years with really high value are more… Read more »
studes
Guest
studes

Eddie, what you’re talking about is exactly what WSAB does.  It gives more value to peak years than average years.  However, it doesn’t ignore the non-peak years.

By setting replacement level at, say, 12 Win Shares, an 18 Win Share year (about average) gets six WSAB, while an outstanding year (say, 30 Win Shares gets four times as many WSAB.

Numbers are estimates only.

studes
Guest
studes

Bad math.  30 WS minus 12 WS is 18 WSAB, three times as much as an average player.

All these systems, by the way, such as WAR give more relative value to peak years.

John
Guest
John

Thanks Dave.  The park factors explanation I read obviously lead me astray.

Keith
Guest
Keith
Good article, but I feel that ERA+ actually favors pitchers in higher run environments.  Unlike OPS, which has no upper limit, an earned run average can only get so low.  In 1966, for Sandy Koufax to have an ERA+ that matched what Pedro Martinez had in 2000 (291), he would have needed an ERA of 1.13.  While it is possible to manage an ERA that low in a 300-inning season, it is a pretty unrealistic expectation.  It’s just too rare for a team not to be able to manufacture a couple of runs, no matter how depressed the run environment… Read more »
Dave Studeman
Guest
Dave Studeman

@Keith, did you read the article I linked to?  It directly compares Koufax and Pedro.

Tim
Guest
Tim

Whoops, I just realized I got involved in the wrong discussion. I checked your original story about the Best Pitchers of all time and saw Bert Blyleven. While there were a few other names, quite a few, that stuck out I almost fell out of my chair when I saw that. Over Koufax? Right.

Craig Tyle
Guest
Craig Tyle

This illustrates, for me at least, that something like Wins Above All Star corresponds better to our intuition of who are the best players of all time.

I’d love to see a list that compared (for both hitters and pitchers) Wins Above All Star and Wins Above Bench (or Replacement) and which players see the biggest differences.

Keith
Guest
Keith
Well yes, and I don’t agree with your conclusion.  The fact that it was more difficult to post an ERA+ of 130 in the 60s than it is today makes a 130 ERA+ in the 60s more impressive, not less so.  Also, when comparing pitchers with an equal ERA+, I would favor the pitcher with the lower ERA.  We chalk up era differences to dynamic playing conditions, and that is true to a large extent.  But it is also possible that there were just a larger percentage of talented pitchers in the 60s than there are today, and we shouldn’t… Read more »
Northern Rebel
Guest
Northern Rebel

Bill James had two categories:

Lifetime, and peak.

There is no doubt at his peak, Koufax may have been the best. Certainly Lefty Grove may have topped both categories.

It is highly subjective, but there are ways to compare pitcher’s stats with the league they played in, etc.

I was too young to see Koufax, but I did see the greatest peak pitcher of all time:

Pedro.

Dave Studeman
Guest
Dave Studeman
There is no doubt at his peak, Koufax may have been the best. I’d say there is a lot of doubt about that, for all the reasons listed above. It is highly subjective, but there are ways to compare pitcher’s stats with the league they played in, etc. I don’t believe the way I’ve approached it is subjective at all.  I’m sure the approach can be improved upon, but I wouldn’t call it subjective. But it is also possible that there were just a larger percentage of talented pitchers in the 60s than there are today, and we shouldn’t penalize… Read more »
Keith
Guest
Keith
Of course we should.  Who knows what drives offense/defense up or down, but in low-offensive eras, good offensive players are more valuable than good defensive players.  If you focus on value instead of how “good” a player actually is, you can come up with real answers to your questions.  Otherwise, you tie yourself up with Gordian Knots. That’s the same reason a high ERA+ is less valuable in a low-run environment era.  The fact that you can’t allow less than zero runs means that there is a limit on your value in that type of environment. Bottom line in all… Read more »
Dave Studeman
Guest
Dave Studeman

I’d love to see a list that compared (for both hitters and pitchers) Wins Above All Star and Wins Above Bench (or Replacement) and which players see the biggest differences.

In the hitters article, I talk a lot about the relative differences.  Not so much in the pitching article.  One of the biggest jumps from WSAB to All Star WS is Smokey Joe Wood, who goes from 134th to 22nd. Dizzy Dean also has a big jump, slightly more than Koufax’s.

Dave Studeman
Guest
Dave Studeman
Good post, Keith.  This point… Our big objective in comparing across eras is normalizing variables.  We want to know how Koufax etc. would have performed under neutral conditions so we can compare them on an even keel. …is where we disagree. To me, it is impossible to neutralize conditions. How do we account for steroids, blacks not playing ball, etc.? I don’t think we can or ever will. When we “normalize” a player’s environment, we get a better handle on his value.  A 1.73 ERA is worth less in a low-run environment.  That is a fact and stats like ERA+… Read more »
Ned Lips
Guest
Ned Lips
Other issues that should be addressed is the height of the mound Koufax and his peers pitched from versus those after 1968 and the number of innings they pitched relative to today. No one from that era achieved 300 wins. Also the practice of having valuable starters pitch in relief stopped in 1960 for some reason. Ford, Spahn, etc all had 8 – 12 relief appearances a year. Once Koufax bacame a full time starter in 1958 he has 14, 12, 11 and 7 relief appearances each year, which was not uncommon. From 1961 on, his great years, he had… Read more »
Ned Lips
Guest
Ned Lips

As to wins, over a period of time there seems to be some value. After all that is what they played for. Comparing the greats of that era in terms of wins over their 6 best consecutive years:
Koufax   129
Gibson   119
Jenkins 127
Marichal 133
Spahn   126 (who obviously won over 300 in this era)
Alot of National Leaguers on different teams, teams up and down, with similar 6 year win totals.

Ned Lips
Guest
Ned Lips

In my mind, Dave has the rating correct. Koufax was comparable to his contemporary greats for 6 years, worse for six years, and they had much longer and more successful careers. They all won games in the environment they were faced with. Top 100 all time is still pretty darn good.

Dave Studeman
Guest
Dave Studeman
@Keith, I was thinking about my reply and I don’t know that I addressed your concern directly enough.  Let me give it one more try, for old time’s sake. Think of ERA as a value statement.  A 2.43 (or whatever) ERA has an implied value to a casual observer. ERA+ is also a value statement—one that takes into account the player’s run environment.  It doesn’t “normalize” a variable (at least not to my way of thinking).  It’s a separate value statement that has a consistent continuum across different eras of baseball. Really, it’s just like linear weights.  Linear weights vary… Read more »
Keith
Guest
Keith
I think the corollary you are looking for with regards to linear weights is not the multiplier for the value of an out – that is basically the same thing that ERA+ does by setting the league average to 100 (with batting or pitching runs, zero).  Instead, the corollary to using mERA over ERA+ I believe is using batting wins in favor of batting runs.  I do realize I am in the minority of preferring batting runs to batting wins; maybe I am looking at player comparisons across eras differently than most.  At the same time, I think if you… Read more »
wpDiscuz