Base Runs and Carl Hubbell

Base Runs

Some years ago, the big focus in baseball analysis and sabermetrics was the development of “run estimation” calculations that estimated the number of runs produced by individual players for their teams. In the 1980’s, two competing approaches emerged in the general mindset of people who cared about such things: Bill James’ Runs Created and Pete Palmer’s Linear Weights, also known as Batting Runs (scroll down the linked page for the Batting Runs definition). Today, Baseball Reference lists both stats in the “Special Batting” section of their player pages. Here’s an example, partway down the page.

Since then, many people have derived their own run estimation formulas. Dan Fox covered most of them in this space a few years ago and John Jarvis has also posted a detailed study of various formulas. The Internet’s resident expert on run estimation formulas may well be the analyst known as US Patriot. Here’s a recent example of his work.

In my humble opinion (“IMHO” to you text messengers), the best overview of run estimation formulas was a three-part series posted by Tangotiger at BBTF (and now available on his site):
Part One
Part Two
Part Three

One of the main lessons I took from Tango’s articles was that context truly matters. In a world in which every batter hit a homer, each one would be worth exactly one run because no one would ever be on base. Conversely, in a world in which every batter walked, each walk would also be worth exactly one run.

Of course, we don’t know of any worlds like that, at least not on the professional level. But analyses such as Dan’s and John’s posted above, which depend on comparisons with team statistics, can be misleading. Why? Because the difference between the best and worst offensive teams is much less than the difference between the best and worst offensive players. Albert Pujols may not be as extremely different from Tony Pena, for instance, as the world in which everyone hits a home run all the time. But they are very, very different. And the results of team-based analyses may not apply to them at all.

So Tango makes a pretty compelling argument that Base Runs (invented by David Smyth) is the best run estimator available, based on two things:
{exp:list_maker}It is flexible in its architecture, such as the way it always values a home run as one run. That makes the Pujols/Pena comparison more apt.
It can be adjusted for context by simply recalculating one part of the formula. {/exp:list_maker}Why do I mention all of this right now? Because Tango recently suggested to Bill James (at Bill James Online, subscription required) that he create a new version of Runs Created that is based on the Base Runs formula. Bill’s response was, “If you can’t convince anybody, how can I?”

Now, Bill likes to be flip in his answers, so I’m not sure we should interpret his response too specifically. But in case there is any question, I’m convinced. In fact, I venture to say that most THT writers who’ve spent much time thinking about the issue are convinced.

Up to now, we have reported Runs Created in our stats section for three basic reasons:
{exp:list_maker}It’s a well-known run estimator.
It’s pretty good.
It’s used in Win Shares, which we also publish in our stats. {/exp:list_maker}That’s about to change. The Hardball Times stat section will soon be publishing Base Runs instead of Runs Created, and I will start using it as the batting estimator in our Win Shares stats. Why? Because I’m convinced it’s the best available run estimation formula.

Here’s one reason. If you add up all the Base Runs of all players this so far this year, you get 9,686 runs. If you add up all of their Runs Created (before the final step, in which you force the RC totals to equal team totals), you get 9,906 runs. The actual major league total for runs scored as of Wednesday? 9,693. Base Runs comes within seven runs of the total; Runs Created is more than 200 runs greater.

In the Win Shares system, James forces the RC totals for all players to equal the team totals by applying a proportionate ratio. In other words, the system applies an arbitrary approach to make the numbers “foot.” Once I apply Base Runs to Win Shares, that step won’t really be necessary. By definition (and through a more logical approach), the Base Runs figures will virtually foot with all the team totals.

To give you a sense of the difference between the two estimators, here is a comparison of the top ten batters in Base Runs and their Runs Created totals (through Tuesday’s games. See my note below regarding this version of Base Runs).

Player               Team    BR     RC    Diff
Berkman, Lance       HOU     64     67      -3
Utley, Chase         PHI     62     61       1
Burrell, Pat         PHI     59     59       0
Hamilton, Josh H     TEX     58     56       2
McLouth, Nate        PIT     58     59      -1
Kinsler, Ian M       TEX     57     55       2
Jones, Chipper       ATL     57     65      -8
Sizemore, Grady      CLE     56     56       0
Gonzalez, Adrian     SD      55     60      -5
Bradley, Milton      TEX     54     53       1

As you can see, there usually isn’t much of a difference between RC and BR for most batters, but there are some significant differences for a few individual batters, particularly Chipper Jones and Milton Bradley.

So Base Runs is better, and it can make a difference for some individual players. It’s time to switch. If you’re not convinced, post a comment at Ballhype and I’ll follow up with some more analysis of the two systems. In the meantime, I’ll post a note to THT Live once we have Base Runs in production.

Carl Hubbell

THT reader Warren Corbett is working on a book about the life and times of Paul Richards, which promises to be an extremely interesting read. Richards was one of the most innovative figures in baseball history.

While conducting his research, Warren came across something that has never been published before: a 1982 oral interview with Carl Hubbell, the great Giant screwballer. (I know that sounds funny. Let’s see what my editors think.) The interview had been conducted by SABR’s Oral History committee, which loaned it to the Baseball Hall of Fame and only recently got it back.

Here’s the interview, with a little introduction by Warren. I thought you’d find this interesting, particularly after reading Mike Fast’s recent article about Daniel Herrera’s scroogie.

A Hardball Times Update
Goodbye for now.

Carl Hubbell was retired after 253 major-league victories and many years as the Giants’ farm director when author Walter Langford interviewed him in 1982, a week before his seventy-ninth birthday. The tape is in the collection of SABR’s Oral History Committee.

After Hubbell graduated from high school in Meeker, Oklahoma, he went to work in the oil fields and pitched for company teams. He began his professional career in 1923 at age 19, with Cushing in the Oklahoma State League. When the Class D league folded the next year, he moved to Ardmore of the Western Association and Oklahoma City of the Class A Western League. That’s when his life changed:

I think probably one of the most fortunate things in my life was, there was an old lefthanded pitcher on the ball club named Lefty Thomas (short-term major leaguer Claude Thomas)…Really knew how to pitch, and for some reason or other he really attracted me. He was the only one I ever looked at. He made pitching look easy, and he had a kind of a sinking fastball. Of course the righthanded hitters like to tee off on the lefthander, especially when they get him in a hole. And he’d go out there and got one or two on, and he’d just throw that sinker on the outside corner to get a three-hopper to shortstop, boom-boom. He’d walk off, wouldn’t break a sweat or anything. And, boy, it just enthralled me. I guess I had the instincts of knowing – I was real young – and to know that was the one, right there.

[In spring training the next year] I was going to work on turning that ball over…I had a pretty good sinker, but they wanted me to try to get a change of pace that goes off my fastball. Fortunately I had pretty flexible joints and wrists.

I found out that the more I turned it over, the more I come up and over [overhand] I could get a much better break on it, you see. Of course, the more spin you get on the ball, the more break, and it slows it up. When I threw the screwball I came right over the top, and I turned my arm clear over and let the ball come out of the back of my hand. That’s exactly the way I threw it.

And, you know, everybody said, ‘Oh, you’re gonna tear your arm up.’ The amazing thing, in 30 days I could get that over the plate as good as my fastball. I just came right up and turned it over, and it came out of the back of my hand. And that slowed it up tremendously. The real effectiveness of the screwball was not the break at all. It’s the speed of the ball. I threw it just like my fastball…I did have a fairly good fastball. And that was the effectiveness of the screwball. I threw it exactly with the same motion.

…The only thing it bothered in my arm was here [elbow]. I never had an ache or a pain in my shoulder, a ligament or a muscle or nothing. Just a calcium deposit [in my elbow], just from pounding. When you throw, you take a hell of a pounding. And nature starts putting in there something [calcium deposits] to protect it…My arm started crooking up in 1934. I was just half a pitcher. (Hubbell won more than 20 games each year from 1933-1937, but his career quickly tailed off after that. It was said that he had to have his coat sleeve shortened because he could not straighten his left arm.)

[What do you think of other pitchers who throw screwballs?]

Nobody ever mastered it. They throw ‘em, I mean [Warren Spahn and Mike Marshall]. They don’t turn their arm over. [Fernando Valenzuela] comes right up over the top, but unless you get slow-motion you can’t tell.

[After Detroit managers Ty Cobb and George Moriarty rejected him, Hubbell joined the New York Giants in 1928. His manager was John McGraw.]

He played that ‘Little Napoleon’ bit to the hilt. I mean, that was his life. His life. He lived it. He must have dreamed it. The Little Napoleon.

I just couldn’t understand it. I’d had such a bad experience in Detroit. He gave me a chance to pitch in the big leagues, and I appreciate it a hell of a lot.

But he called every pitch, and every pitch was a breaking ball, be a curveball or a screwball. He didn’t think you had sense enough to throw a fastball. He must have been the worst curveball hitter in the world.

The only pitcher he would allow to pitch his own game was Christy Mathewson. And the only reason for that was that he was the only college graduate pitching in the major leagues at that time…

He give me a chance to pitch. I didn’t want to say anything [bad about him]. Four years with him, it’s a wonder my arm lasted as long as it did, it really is…

[Roger] Bresnahan, the old catcher, was a coach. [Bresnahan played for McGraw from 1902 through 1908.] I started asking him about John McGraw, how he could be like he is. He said, ‘You don’t understand how it was.” [McGraw] started managing before the turn of the century, with Baltimore…

[Bresnahan] said in those days a regular player made about twenty-five hundred dollars a year. A star like Mathewson might make five thousand. He said, hell, they couldn’t stay in a decent hotel. They had to stay in rooming houses down by the railroad tracks, you see….

Then Ruth came along, and the salaries went up, the college graduates came in. They could afford to stay anywhere…But McGraw did not change. He was the Little Napoleon.

He had the news media so conned, this Little Napoleon did. Nobody interviewed McGraw. After every game Bozeman Bulger [of the New York Times] would come in [to McGraw’s office]. Nobody said a word. McGraw would write out everything that happened in the ball game. If we won, it was something he had done, and if we lost the ball game, it was something some damn ballplayer did…

And Bulger would take it out and say, ‘Fellows, here’s McGraw’s report.’ And they would write it down.

Great stuff. I particularly like the comments about John McGraw and Hubbell’s observation that “the real effectiveness of the screwball was not the break at all. It’s the speed of the ball.” Many thanks to Warren Corbett, and be sure to watch for his book on Paul Richards.

References & Resources
Base Runs is a very flexible formula and can be implemented in a number of ways. In order to facilitate a comparison with the most recent version of Runs Created, I calculated it in the followings ways:
{exp:list_maker}I used a league-average scoring rate to adjust the “B” component.
I calculated Base Runs by subtracting each player’s stats from a league-average team. The difference was that player’s Base Runs.
I added a “clutch” component to Base Runs, giving each player credit for hitting well with runners in scoring position. I used the exact same methodology employed in Runs Created.
I applied a park factor to both figures. {/exp:list_maker}

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

Comments are closed.