The Red Sox and Analytics

[Kielty's Last Pitch]

Off topic, but wouldn't the more appropriate screen name be Kielty's Last Swing?

I agree. But that would leave open the possibility of a BB, HBP, or CI after the HR.

I just think it's kinda cool that it was the very last pitch he ever saw in the majors.

 

[Kielty's Last Pitch]

The Red Sox signed Ortiz on January 22, 2003. PITCHf/x was first used in the 2006 play-offs, then in most of the parks the next season but it wasn't until 2007 that it was operational in all 30 parks. This meant there were only crude ways to evaluate pitches (from the point of view of batters and the point of view of pitchers) and through new metrics (people began to realize that batting average and fielding average were not the best methods of evaluation). Sabermetrics is named that because the vanguard of new thought was coming from members of the Society for American Baseball Research (SABR...thank you, Bill James). As sabermetrics began to move into MLB, so did SABR members. It was slow at first but as more teams began to adopt the theories, it exploded and techniques such as PITCHf/x and TrackMan and StatCast followed.

Actually, PITCHf/x was tested in a few MLB games in 2005. Somewhere among my things I still have a media pass from the Cardinals when I visited a test at a game during SABR's National Convention (thanks Dr. Alan Nathan). I started going to the Statistical Analysis meetings at SABR conventions back in 1983 along with people like Bill James, Pete Palmer, Sean Forman, and many more. I don't recall Mike Gimbel being the first person every listened to.

Great post, appreciate it.

Gimbel got into Sabermetrics because of James, and then published his first book in 1990. Duquette hired him in 1991 and immediately began utilizing his work and advice on player personnel. I'm not aware of others who were hired by an MLB team to give their advice prior to Gimbel. James didn't get hired by an MLB team until 2003 with the Sox. I'm not aware of Forman doing anything baseball-related prior to launching BR.com in 2000. Palmer was more of a football guy, wouldn't you say?

 

[charlieoscar]

While Pete Palmer did some work with football stats, I think he is best known for his work with baseball. The Hidden Game of Baseball, written with John Thorn (MLB's official historian) covered the history of the development of baseball stats. He also was involved with producing yearly baseball encyclopedias Basically pre-internet). The is a written interview with him on teh SABR site: http://sabr.org/research/history-total-baseball-and-pete-palmers-baseball-database

Seems to me that Duquette (or someone around that time) hired a person to do some statistical work, who once based something on the standard deviation of two numbers. When you had clubs hopping aboard the sabermetric revolution in the early days, you might have had people who really did not understand mathematics doing the hiring...Oh, that sounds like a great idea. You are hired. Only to later find out that the person they hired really didn't understand the math well, either.

 

[joe dokes]

While Pete Palmer did some work with football stats, I think he is best known for his work with baseball. The Hidden Game of Baseball, written with John Thorn (MLB's official historian) covered the history of the development of baseball stats. He also was involved with producing yearly baseball encyclopedias Basically pre-internet). The is a written interview with him on teh SABR site: http://sabr.org/research/history-total-baseball-and-pete-palmers-baseball-database

Seems to me that Duquette (or someone around that time) hired a person to do some statistical work, who once based something on the standard deviation of two numbers. When you had clubs hopping aboard the sabermetric revolution in the early days, you might have had people who really did not understand mathematics doing the hiring...Oh, that sounds like a great idea. You are hired. Only to later find out that the person they hired really didn't understand the math well, either.

As Palmer and Thorn have recognized, it started with Branch Rickey hiring Allan Roth.

 

[charlieoscar]

As Palmer and Thorn have recognized, it started with Branch Rickey hiring Allan Roth.

That might have been the first baseball club but there were people developing new stats before that...as Palmer and Thorn also say.

 

[Kielty's Last Pitch]

As Palmer and Thorn have recognized, it started with Branch Rickey hiring Allan Roth.

That's not the same thing though. Roth was the first to be hired as a statistician. Gimbel was the first to be hired as a consultant, the first who placed value on aspects of the game (ie: OBP) that were undervalued up to that point in time. He was the first to advise an MLB team on which players to acquire and remove. Duquette was skewered by the media for the innovative approach to player personnel decisionmaking, even after Gimbel was let go. Remember Duquette publicly touting Offerman's OBP and the backlash he got for it? Mo Vaughn's infamous "You can't evaluate this game through a computer or over the Internet."?

As is often the case, pioneers are generally not treated well by those who are incapable of thinking outside the box and who are reluctant to embrace change.

 

[joe dokes]

That might have been the first baseball club but there were people developing new stats before that...as Palmer and Thorn also say.

Of course. Even Roth himself first turned his analysis to hockey before he hooked up with Rickey.

That's not the same thing though. Roth was the first to be hired as a statistician. Gimbel was the first to be hired as a consultant, the first who placed value on aspects of the game (ie: OBP) that were undervalued up to that point in time.

Whatever Gimbel's status, duties and contributions, he was not the first, either generally or specifically. Roth, too, used OBP in his work::

https://sabr.org/bioproj/person/350155ef

Rickey asked that Roth send Ed Staples, his assistant, a detailed outline of Roth’s ideas. The four-page letter contained proposals to track a wide range of statistics. Some of these were standard, but others, such as where the ball was hit and the count it was hit on, hadn’t been compiled regularly. Roth also proposed to break the statistics down into various categories that would reveal tendencies which the front office and the manager could use to win ballgames. Breakdowns such as performance against left-handers and right-handers, in day games versus night games, in the various ballparks, in situations with runners in scoring position, are all mundane to us now. But in Roth’s time, they were rarely compiled or used, and never part of the public discussion. The letter was intriguing enough to get a meeting with a still-skeptical Rickey.
The conversation turned positive, Roth said, when Rickey asked him about runs batted in. Roth said he didn’t think much of runs batted in unless they were correlated with the chances to drive them in, and differentiated again by which base they’d been driven in from. This meshed with Rickey’s own beliefs and the conversation flowered. Roth was offered the job.

.........

In the offseason, he would refine the numbers further, seeking longer term trends and finding the outliers. Everyone knew right-handed hitters generally performed more poorly against right-handed pitchers and vice versa. Roth would look for, and find, the left-handed hitter who broke the mold and could provide a manager with an unexpected platoon advantage. He tracked bases advanced, a metric that encompassed baserunning statistics as well as the ability to move runners along with outs. He recorded what happened at each point in the count, what happened in bunting situations and differences between night and day games, home and away games, and in individual stadiums. No other team had access to such analysis at that time.

.........................

In his first season, for example, Roth used another of his innovations—spray charts showing the location of all of a player’s batted balls—to show that Dixie Walker’s hits were going to the opposite field more and more frequently. Rickey, following his own dictum that it was better to trade a player a year too early, sent Walker to the Pirates. “The People’s Cherce” hit .316 in 1948, but was down to .282 the next year, and became a player-manager in the minors. A year after his Walker revelation, Roth’s numbers showed that in 1948, Jackie Robinson drove in a higher percentage of baserunners than any other hitter in the lineup. Manager Burt Shotton moved Robinson, who had barely broken into double digits with 12 homers, into the cleanup spot. He hit only four more home runs in 1949, but drove in 124 and won the National League Most Valuable Player Award.

..............

In 1954, Roth’s work hit the big time—with a heavy coating of Branch Rickey. Life magazine, one of the largest-circulation magazines in the country, ran an article titled “Goodby [sic] to Some Old Baseball Ideas.” The article said it had been written by Branch Rickey, whose picture graced the first page. Roth’s back is visible in the background of that photo, and he is pictured on the article’s third page, along with a multipart equation. That equation was clearly Roth’s work; Rickey called the equation “the most disconcerting and at the same time the most constructive thing to come into baseball in my memory.”
In many ways, “The Equation” was years ahead of its time. Its first two terms were what we today call on-base percentage and isolated power. It would take the book Moneyball half a century later to cement the importance of on-base percentage. The equation, which contained eight different terms, including pieces devoted to run-scoring efficiency, pitching, and fielding, was vastly complicated for contemporary baseball organizations.

......................

In looking for meaning in the numbers, Roth’s methodology was much like that of Bill James and later members of the Society for American Baseball Research: take a piece of accepted baseball wisdom and analyze whether it was true. “Some fellows have mentioned that batting average increases of 10 or 12 points would result from the sacrifice fly rule,” Roth said during a 1953 discussion of scoring rules, “The figures on the Dodgers for the last two years don’t come anywhere near such figures.”

 

[Kielty's Last Pitch]

Just to reiterate, Gimbel was the first to be hired by an MLB team as a consultant and Duquette was the first to hire an analytics guy to assist in player personnel decisions. I never claimed Gimbel was the first at anything else, and have already stated that Bill James' Baseball Abstracts was Gimbel's inspiration..

 

[Yo La Tengo]

Thinking about the Sox putting Owens on waivers and the team's lack of success in assessing/drafting pitchers. I'm guessing this has been discussed on the site at some point, but, is there an analysis anywhere that compares successful hitter v. pitcher picks in the early rounds? My general take is that a higher percentage of early round hitters become successful MLB players, but that is a perception based on a shallow review and could reflect my personal bias.

Since 1965, Sox have had 75 first round picks, 34 pitchers, with three excellent to good picks (Clemens, Hurst, and Buchholz) and a few other useful pieces (trade or otherwise... Barnes, Kelly, Bard, Fossum... others?). As for hitters, 41 picks, with a number of very good picks (Conigliaro, Rice, Vaughn, Ellsbury) and a number of useful pieces (Nixon, Hatteburg, Murphy, Lowrie) plus Bradley and Benintendi. These numbers don't really move the needle one way or the other.

Obviously, the picks should be weighted with a bigger sample size to have a meaningful impact. And, as an aside, it is pretty amazing that the Astros just won the World Series after drafting Mark Appel and trying to draft Brady Aiken number 1 in back to back years (talk about dodging a bullet... I believe they ended up with Bregman the next year as the consolation prize).

 

[charlieoscar]

Some questions to be asked about the drafting of pitchers are (in no specific order):

1. Teams have roughly the same number of pitchers as they do position players, so are the numbers of each drafted approximately equal over all?
2. Some pitchers also played positions. Do they always get drafted as pitchers?
3. Is there any way of telling if teams tend to go for the best player available at their pick or whether they take teh best available player at a position they are trying to fill?

The draft changed drastically a few years back and it no longer goes into rounds 50 or beyond, so strategies must have changed. Did all teams adapt equally well to that change?

 

[shepard50]

Some questions to be asked about the drafting of pitchers are (in no specific order):

1. Teams have roughly the same number of pitchers as they do position players, so are the numbers of each drafted approximately equal over all?
2. Some pitchers also played positions. Do they always get drafted as pitchers?
3. Is there any way of telling if teams tend to go for the best player available at their pick or whether they take teh best available player at a position they are trying to fill?

The draft changed drastically a few years back and it no longer goes into rounds 50 or beyond, so strategies must have changed. Did all teams adapt equally well to that change?

This doesn't answer all your questions, but it has insight into some of them. It's a Verducci article in SI from June of this year and he talks about success rates in the draft, specifically the first round), the risk of high school players, and the higher surety of position players. Good read.

 

[charlieoscar]

This doesn't answer all your questions, but it has insight into some of them. It's a Verducci article in SI from June of this year and he talks about success rates in the draft, specifically the first round), the risk of high school players, and the higher surety of position players. Good read.

That was an interesting article and it does get to some of what I thought needed to be answered better than I phrased it.

 

[Yo La Tengo]

Thanks for sharing that article.

I'm with Theo: The Cubs are the only franchise not to draft a pitcher within the first 30 picks in the past five drafts.