Skip to content

How soon can we get rid of the human element?

Apr 7, 2010, 7:30 PM EDT

Anyone who followed along during our postseason in-game chats last year knows I’m strongly in favor of using technology to aid umpires. Someday in the not-so-distant future, it’s going to be feasible to do balls and strikes by computer, and who knows, maybe 20 or 30 years after that, MLB might actually come around to making a change.
In the meantime, we get this. Using Pitch F/X data, John Walsh has determined that the size of the strike zone is nearly 50 percent larger in a 3-0 count than it is with 0-2.
Go check out the full article with its graphs and pics. It’s very impressive stuff.

  1. willmose - Apr 7, 2010 at 11:44 PM

    I know it is too much to ask that sportswriter have a basic grasp of mathematics and science, but there is old computer saying garbage in garbage out, GIGO for short. My, my. Look at the pretty graph and colors, it must be right. Pitch F/X strike zone are as phony and subjective as those strike zone graphics FOX shows on the game of the week. The problem is two fold. First, tracking the ball from various camera angles does not product an accurate track of the ball. Small differences produce larger difference in tracking. Secondly, the strike zone is determine by the batter’s stance when the ball leaves the pitcher’s hand, not when the ball crosses the plate. The top and bottom of the strike zone changes with every pitch.
    There are, of course, ways to get an accurate picture. A sensor in the middle of the ball, sensors surgerical inplanted in the knees and letters of every batter, and a localized GPS system fixing the plate relative to multiple receivers.
    Of course, while you are at, I make the pitching rubber weight sensitive so we can tell in the pitcher has his foot on the rubber when he release the ball. Otherwise, it is balk or no pitch.

  2. Patrick - Apr 8, 2010 at 12:26 AM


  3. Patrick - Apr 8, 2010 at 12:29 AM

    It’s pretty clear you haven’t read much about Pitch Fx, which is a shame.
    There are a lot of good articles on understanding Pitch FX out there, I recommend googling around a bit.
    Moving on:
    To your statement about the strike zone, here’s the official rule:
    The STRIKE ZONE is that area over home plate the upper limit of which is a
    horizontal line at the midpoint between the top of the shoulders and the top of the uniform
    pants, and the lower level is a line at the hollow beneath the kneecap. The Strike Zone shall
    be determined from the batter’s stance as the batter is prepared to swing at a pitched ball.

    You can find the rules on the offical MLB site. That’s copied directly.
    As a batter is prepared to swing at a pitched ball. So, no, it is NOT when the pitcher is ready – It’s when the batter is ready to swing. If a batter (think Kevin Youkilis) changes his stance significantly during the flight of the pitch – BUT BEFORE HE STARTS TO SWING – it is from THIS ready stance – not that weird pre-pitch stuff where he waggles the bat around over his head – that his strike zone is judged.
    Going on to another point:
    Lets talk errors and type of errors. Lets assume for a moment that Pitch FX is severely inaccurate, that it has a random error such that it puts the pitch location as off by up to 6 inches (so, a circle around the actual location, 6 inches in radius, with the reported location random distributed in there. Maybe you think it’s REALLY bad and it’s 12 inches, and while I’m sure MLB would’ve long ago fixed an error of that size, it doesn’t matter. Go ahead and say 12).
    Now the error must be at least mostly randomly distributed, if it was always (or mostly) wrong in one direction, it would be obvious and would be corrected. We would see a ton more balls called on one side of the plate or the other by Pitch FX, regardless of the handedness of the batter-pitcher, etc. We don’t see that.
    So then:
    The nature of random error is such that in a large enough sample, it cancels out. If you go look at a bunch of Pitch FX graphs, it’s obvious it has no major systematic errors, it’s quite internally consistent and doesn’t have a strong bias in any PARTICULAR DIRECTION. So, the error is random.
    So if Pitch FX is getting the location of every single pitched ball wrong (but again, the error is random), it would be a horrible tool for a pitch-by-pitch study of a game or a study of an individual pitcher…
    And it wouldn’t matter in the SLIGHEST for the study above. It would be essentially as valid as if Pitch FX were perfect. After a few thousand pitches, we still get a perfectly accurate idea where, on average, the boundary between a ball and a strike lies. The errors all cancel out when you average them… For every 6 inches to the right, there’s a 6 inches to the left, or close enough, and taking an average cancels it out and gets you the right answers. That’s the nature of random error.
    In addition, MLB uses Pitch FX data to grade umpires. You may just say they’re badly wrong about using it, but you’ll forgive me if I take their opinion over yours.

  4. IsleOwnage - Apr 8, 2010 at 1:05 AM

    Lets say that the Link is 100% perfect (which we can agree that it is not 100% acurate). We also must remember that the game of Major League Baseball is a business, and in this business most people who watch would enjoy a ball put into play rather than a strikeout on 0-2 or a walk on 3-0.

  5. willmose - Apr 8, 2010 at 9:36 AM

    If Pitch FX is so good why are not margins of error, real scientific standard given? You, of course, assume the errors are random? What basis do you have for that without any error analysis? You arguement that the errors all even out in the long run is just as bogus even if they are random. Gee, sorry the Red Sox lost the game to the Yankees because the last batter was called out on a ball that Pitch FX called a strike. If bad calls by the umpires are random don’t they also even out in the long run.
    Besides not have having any margins of errors about the placement of the ball as it crosses the plate, there is NO margin of error about whether the pitch was a strike or not. Pitch FX has no way to compare to tell whether or not the pitch was exactly a strike or not.
    I am glad you agree with me about the strike zone changes with every pitch and even during the pitch. The batter is, of course, prepared to swing at the ball when the ball leaves the pitcher’s hand, otherwise batters would simply fall flat and every pitch would be a ball.
    MLB uses Pitch FX to grade umpires. That is a good use for it because in spite of all the comments above it is consistent, calling two similar pitches the same way. Pitch FX may call them both balls strikes when they were not, but they will be consistent about it. That is what MLB and the players would like to see a consistent strike zone from the umpires.

  6. Dan W - Apr 8, 2010 at 10:05 AM

    The day the strike zone is a computer, is the day I stop watching. Heck, at that point lets just make it a giant video game since the computer could be perfect. We could just sim the seasons and crown the champion without a single umpire mistake.
    Taking this out of preportion…sort of…just makes my blood boil when we talk about removing the human element of a game.

  7. MrMagoo - Apr 8, 2010 at 1:14 PM

    There are rules about calling balls and strikes. How many umpires follow these rules. The reality is, the strike zone is wherever the umpire says it is. And it can change from batter to batter, pitcher to pitcher, umpire to umpire. It can even change from one pitch to the next. Each umpire has a preconceived idea about how he is going to call the game. And if that ump has a headache, a hangover, is tired or sore, doesn’t like a certain batter or pitcher, wants to get the game over because he has someplace to go, has had a pregame discussion with the commissioner, or simply can’t see straight, then his calling of balls and strikes might not be right. That’s the human element.

  8. Steven Ellingson - Apr 9, 2010 at 3:37 AM

    Dan, your comment is ridiculous. We aren’t talking about replacing the players with computers. we’re talking about making the game as fair as possible for them.

Leave Comment

You must be logged in to leave a comment. Not a member? Register now!

Top 10 MLB Player Searches
  1. G. Stanton (2756)
  2. B. Crawford (2618)
  3. Y. Puig (2442)
  4. G. Springer (2344)
  5. C. Correa (2243)
  1. H. Ramirez (2226)
  2. J. Fernandez (2121)
  3. J. Hamilton (2110)
  4. D. Wright (2107)
  5. H. Pence (2087)