Search THT:

Or you can search by:

March 09, 2010

A note on minor league ground ball rates

While I continue to pour through mounds of Gameday data from Minor League games, I thought I'd stop and share some numbers. I haven't found a "story" that's compelling enough to write about yet. Nor have I finished running through the numbers.

In a nutshell, I compared "year 1" and "year 2" performances for all pitchers who faced at least 200 batters in one league in 2007 or 2008 to the same pitcher in another league in 2008 or 2009. For example, Joe Schmoe if pitched in the Southern League in 2007 and the International League in 2008, he goes into the pool. There's all sorts of problems with this approach, namely I'm not filtering pitchers who may have had substantial experience at a higher level during "year 1", but it does give some idea of how guys progress.

A pitcher with a 50% ground ball rate in the Southern League (Double-A) would project to a 45% rate in the following season in the Pacific Coast League (Triple-A). A 40% rate projects to 37% and a 60% rate to 53% (y2 = y1*.7702 + .0656; r²=.4693). This ignores any information other statistics may provide in projecting ground ball rates. It does, however, let me split players into some cohorts (out- or under-performed) to carve out some case studies and start digging up those key individual differences that muddy the statistical waters. In this particular group, outliers included Max Scherzer, Esmerling Vasquez and Greg Miller, for whatever that's worth.
Posted by: Harry Pavlidis

November 23, 2009

Greinke and fly balls

In God we trust. All others bring data. -W. Edwards Deming

Along with Zack Greinke's 2009 American League Cy Young Award came a widely-circulated article by Tyler Kepner in the New York Times. In the article, Brian Bannister is quoted, "So a lot of times, Zack would pitch for a fly ball at our park instead of a ground ball, just because the zone rating was better in our outfield and it was a big park."

There's quite a bit to digest in Kepner's article besides that one excerpt from Bannister, but it was one piece that I thought would be easy to check. What did I find?

           Road    Home
Bunt         5%      2%
Fly         30%     33%
Ground      36%     41%
Liner       21%     16%
Popup        8%      8%

The batted ball type comes from the MLB Gameday data. The results don't really line up with what Bannister what saying. It may be that Greinke was pitching for more fly balls at home but not necessarily getting them, but that's a question for another day.

Update: Here's the batted ball log by game.

 Date   Park   Fly  Ground  LD   Pop   Bunt
   4/8  cha      0     8     2     1     1
  4/13  kca      3     4     5     0     0
  4/18  tex      6     9     7     1     1
  4/24  kca      4     9     3     1     2
  4/29  kca      9     6     2     1     0
   5/4  kca      7     8     2     4     0
   5/9  ana     10     9     2     0     2
  5/15  kca      8     7     5     1     0
  5/21  kca      5     8     2     1     1
  5/26  kca      3    13     3     3     0
  5/31  kca      7    11     2     1     0
   6/5  tor     11     3     6     2     0
  6/11  cle      3     8     5     2     3
  6/17  kca     11     7     1     0     1
  6/23  hou      6    14     3     0     3
  6/28  pit      4     4     8     4     1
   7/3  kca      5     9     3     2     0
   7/8  det      7     6     2     0     1
  7/14  sln      0     0     0     1     0
  7/18  kca      4    11     2     3     0
  7/24  kca      3    10     1     1     0
  7/29  bal      6     3     4     3     0
   8/3  tba      1     8     4     0     2
   8/8  kca      6     8     5     3     0
  8/14  det      6     4     5     3     0
  8/19  cha      7     5     4     3     0
  8/25  kca      7     4     1     1     0
  8/30  sea      9     9     2     3     0
   9/5  kca      9     4     8     1     1
  9/11  cle      6     9     2     1     0
  9/17  det      1     3     5     0     0
  9/22  kca      8     3     2     2     0
  9/27  kca      6     8     5     0     0
  10/3  min      7     7     3     0     1

Posted by: Mike Fast

July 10, 2009

From PITCHf/x to HITf/x…now GAMEf/x?

Sportvision is taking leaps and bounds with expanding their ability to track the game of baseball. The Second Annual PITCHf/x Summit convenes in San Francisco this upcoming weekend, and if I could be anywhere other than with my newly expanded family, it would be at the summit. One of the main topics of the summit is sure to be the recently released HITf/x data, but potentially more exciting is the news that Sportvision and MLBAM are undertaking a project to install cameras at every stadium to track not just every pitch and the flight of the ball leaving the bat but the entirety of the game!

If this project works out as early data promises it will, defensive positioning, routes of fielders, baserunning plays, and perhaps even umpire positioning will all be tracked and digitized. The potential to revolutionize analysis of the game and even the game itself looms large!
Posted by: Mike Fast

July 09, 2009

Trackman Baseball

The Trackman Baseball system makes its public debut for the All-Star home run derby in St. Louis. ESPN is calling its implementation of the Trackman feature "Ball Track". Trackman is a portable Doppler radar system already in use for tracking golf ball trajectories and now being deployed in baseball.

The Doppler feature of the radar measures the radial velocity of the ball relative to the radar detector by looking at the frequency shift of the reflected signal relative to the emitted signal. An array of three detectors arranged orthogonally is used to gather phase information from the reflected signal. The difference in the phase of the reflected signal between a pair of detectors can be used to determine the angle to the ball in that dimension. Thus, with two such pairs of detectors, the angle to the ball in two dimensions can be measured. This can be accomplished most optimally with three detectors, with one pair in a line at a right angle to the other pair. If the initial distance between the ball and the radar is known (for example, by measuring the distance from the radar to home plate by some other means), by combining the two angles and the radial velocity information, the exact position of the ball is established in three dimensions.

In addition, a unique feature of this system is that radar reflections from the spinning stitches on the ball can be used to determine the spin rate by measuring the separation of the sideband peaks produced by the signals off the stitches.

That's probably more than most of you want to know about how the system works, but I'm guessing a few physics nerds out there are as jazzed by the possibilities of this system as I am. Many thanks to Dr. Alan Nathan and a few other folks who will remain nameless for their explanations of the Trackman system. You can see a cool graphic of the Doppler returns from a batted ball in a presentation (7 MB Powerpoint, see slide 37) by Alan Nathan on the subject of the physics of baseball.

The bottom line is that this system can be used to reconstruct the full trajectory of a batted ball. It gives us a direct measurement of spin magnitude that we can't get from HITf/x or PITCHf/x. It's nice to see this system getting some public air time.

Hat tip to JBrew at Beyond the Box Score. Thanks also to Alan Nathan for some corrections to my initial post.
Posted by: Mike Fast

June 09, 2009

Which batters hit the hardest balls in April?

With the release of Sportvision's HITf/x data for April 2009, we can look at who hit the ball the hardest in April. Since batted ball speed does not decrease much even when the batter hits the ball off center, it is more instructive to look at batted ball speed in the horizontal plane. In other words, hitting the ball very hard straight up is likely to turn into an out, while hitting the ball very hard straight toward the outfield is likely to turn into a hit. In fact, we find that batting average is maximized at an elevation of 11 degrees, so we'll look at the average batted ball speed for batters in that plane. We should also remove bunt attempts since they purposefully attempt a low batted ball speed.

Here are the results for the 303 batters with 20 or more non-bunt batted balls recorded by HITf/x. Average batted ball speed for a non-pitcher was 75.1 mph. When pitchers were batting, their average batted ball speed was 66.1 mph.
Click for more...

Posted by: Mike Fast


Click here for more THT Notes.