The Baseball Graphs Blog

Saturday, March 19, 2005

Foulke’s Appearances

There has been an ongoing discussion at Baseball Think Factory about Keith Foulke’s usage last year.  Foulke did have a strange record, in that many of his appearances were in low-intensity situations.  This is unusual for a closer.

Two of the issues raised were whether Foulke’s low-intensity appearances occurred primarily in the beginning of the season, and do high-intensity appearances tend to be “clumped” together.

The best way I know to show that is with a graph.  So here ‘tis:


The lines/bars represent each of Foulke’s individual appearances, and the superimposed line is a ten-game rolling average.  P is a measure of the importance of a reliever’s appearance.

As you can see, there is some natural clumping of high-intensity situations, and he was used a lot in low-intensity situations from mid-April toward the end of May.

For reference, the top closers averaged around 0.09 in average P Value over the full season, and Foulke came in at 0.067.

You can find a full definition of the “P Value” in my Hardball Times article.  More bullpen information like this is available in The Hardball Times Bullpen Book.

Posted by Studes on 03/19 at 11:13 AM
Friday, March 18, 2005

States I’ve Been To

I’ve made two cross-country tips by car in my life, and one by train.  I’ve also traveled for business a lot.  Still I haven’t made it to every state in the country:

create your own personalized map of the USA.

Thanks to the Ball and Stick Guy for pointing this site out.

Posted by Studes on 03/18 at 04:03 PM
Wednesday, March 16, 2005

Good Graph Design

It’s nice when a baseball publication manages to include graphics, but it’s also nice if they do it well.  I received the latest issue of Baseball America yesterday, and I was pleased to see that they had a bar graph in the Major League Preview page on page 17.  However, the bars in the bar graphs are a dark blue color, on top of a moderately blue background.  You can barely see the data!  I think the graph has something to do with average age at which contracts are signed, but I’m not completely sure.

Color contrast is a critical element of graph design.  Lately, I’ve been lucky enough to assist David Pinto in the further development of his fielding graphs.  Let me share with you one of the graphs I mocked up for David, focusing on Derek Jeter’s prowess at fielding ground balls (the gray areas represent the minimum and maximum number of outs achieved among all shortstops in each fielding vector):


This graph reinforces what many observers have said; that Jeter is fine going to his left, but among the worst when moving to his right.

But I really want to talk about two things I think are critical for a good graph: color contrast and positioning.  There is a lot of information in this graph, but I like to think it’s relatively easy to read.  Color contrast and positioning help make it so.

Regarding the color, the graph starts with a white background, and next adds gray bars to represent the range of potential performance in each vector.  Next comes the yellow line to represent the predicted performance in each vector (based on the last three years and adjusted for ballpark).  Yellow stands out on top of the gray boxes.  We also made the yellow line relatively wide.

The final line is a black line that represents Jeter’s actual performance.  Black stands out against the white, gray and yellow in the background, and also sits on top of all three.  We also added symbols for the vector data points in the black line.  By making the yellow line wider, you can see those instances when the black line is exactly on top of it.  In other words, you can clearly see where Jeter’s performance was exactly equal to predicted performance.

Finally, we added a couple of gridlines that relate to a typical baseball field: the thirdbase line and centerfield.  This way, you can relate the vectors and graph data to an actual baseball field.

I know I’m biased, but I think this is much better than a bargraph of blue bars on top of a blue background.  The contrasts of color, size and positioning are critical for making good graphs.  I know David is going to include some of these thoughts in the final graphs he produces—and I hope this example can serve future baseball graph producers too.

Posted by Studes on 03/16 at 02:16 PM
Thursday, March 10, 2005

Even More Timelines and Graphs

Yesterday, David Pinto rolled out a new set of graphics that show the relative fielding prowess of shortstops David Eckstein and Cristian Guzman, and he has just posted new ones that compare Mike Cameron and Carlos Beltran.

This is a superb step forward in the field of representing baseball information in a way that makes intuitive sense.  Congratulations, David.

A couple of weeks ago, I talked about baseball timelines,  which are particularly pertinent to baseball due to its rich history.  Well, I just came across a wonderful new baseball timeline, which is available from Parthenon Graphics.  This is truly baseball history at a glance!

Posted by Studes on 03/10 at 11:53 AM
Wednesday, March 09, 2005

Dave Likes Graphs

As David Letterman said last night, charts and graphs are great, because they communicate so well.  You can understand numbers and trends much better when you use a chart or graph.  Paul agreed.

In fact, Dave and Paul like graphs so much that they presented a bunch of them during last night’s show.

Posted by Studes on 03/09 at 09:54 AM
Tuesday, March 08, 2005

Change Partners?

The recent proposed buyout of the NHL by a partnership/corporation has baseball fans wondering if the same ownership structure could be applied to Major League Baseball.  It’s a good question; one I’ve wondered about many times.  You might say that a partnership structure, in which individual partners are relatively autonomous but split the “firm’s” profits according to some formula, makes a lot of sense for baseball.  Many professional organizations, like law firms, are run this way.

A partnership structure is appealing to some, because owners would theoretically be concerned about the overall financial health of baseball.  They’d be partners.  Imagine how George Steinbrenner would act if his profits came out of the total Major League Baseball bottom line—not just the Yankees’.

But partnerships have their faults (sounds like a marriage column, doesn’t it?).  For one thing, a partnership would reduce player salaries and freedoms significantly.  The team owners/partners would essentially collude to keep salaries low, except it wouldn’t be collusion.  It would be legal.  The New Yorker has an interesting article that suggests that the old Marxist conflict of capital and labor is obsolete; that the relevant conflict today is capital vs. talent.  And talent, in today’s knowledge-based world, provides a greater share of value and takes a greater share of the profit.

There are some flaws in the article—when did baseball and hockey players become “knowledge workers?” —but it is true that the baseball profit pendulum has swung away from the capital and toward the talent over the last thirty years.  And this is a good thing.  I believe baseball has reached a relatively equitable balance between the two and I’d hate to go back.

And a partnership structure would likely have other unintended consequences, such as contraction.  Come to think of it, with revenue sharing and friends of Selig ruling the roost, baseball might be closer to a partnership than true competition today more than at any other time in its history.  In my opinon, this is not a good thing.

Posted by Studes on 03/08 at 08:09 AM
Saturday, March 05, 2005

Don’t Blink or You’ll Miss the Ball

Blink O Rama is a site that captures celebrities on TV mid-blink.  For example, heres’s one of the Commander in Chief.  Baseball and sports are captured too, and baseball seems to have a knack for dual  mid-blinks.  When I first started playing baseball as a kid, I closed my eyes every time I swung.  Sure glad no one caught that on camera.

Also, the best baseball site on the web has added something new called Age-based Leaderboards.  Pretty neat stuff, and just waiting for trendline graphs.

Posted by Studes on 03/05 at 07:45 AM
Wednesday, March 02, 2005

Japanese Baseball has marketing woes too

Here’s an article from the Christian Science Monitor that talks about the marketing woes hitting baseball in Japan, and some of the steps professional baseball is taking to address them.

For the first time since 1954, a new team is entering the Japanese league, and the new owner is talking about letting fans vote over the Internet to take out a pitcher, or click on a screen to catch views from the dugout or locker room (!).

It does sound like Japanese baseball could use an infusion of marketing savvy, particularly with recent bribery and attendance padding scandals, as well as the migration of key stars to the US.

Does this sound familiar?  “Young people are leaving baseball because the actual games played at stadiums have lost excitement,” he says. “The tempo of games somehow needs to be sped up.”

Sounds like the USA and Japanese leagues have more in common than I thought.

Posted by Studes on 03/02 at 03:17 PM
Tuesday, March 01, 2005


I’ve got an e-book for sale at The Hardball Times.  It’s called The Bullpen Book of 2002-2004, and it consists of several background articles and a bunch of stats regarding bullpens over the last three years.

I’ve spent some time this offseason writing about bullpens.  To my mind, there are two big issues regarding the bullpen these days.  First, saves and holds are just not adequate to the task of tracking relievers.  We have such great stats for hitters and starting pitchers— if you just stick with ERA and OPS, you’ll be fine—but saves and holds don’t really tell us who the best relievers have been.

Second, managers do not appear to be using their relievers in the most optimal manner.  They show a high willingness to bring in their aces in the ninth with a three-run lead, but not when the score is tied.  This is just backwards thinking.

So I’ve posted a series of articles about this:

The One About Win Probability
Team Bullpens
Ranking the Relievers

And if you’re interested in some more history about the subject, you can try A Graphical History of Relief Pitching

So the Bullpen Book is about 110 pages long, mostly taken up with statistics —every reliever, every year and every team for the past three years— that make far more sense than saves or holds.  Want to see who the best reliever was for a given year or given team the last three years?  The book will tell you, and you can read more about it here.

You can order it by clicking on this button and paying with a credit card.

And if you’d rather not pay by credit card, drop me an email and we’ll work something else out.

Posted by Studes on 03/01 at 08:55 AM