from Amazon
Thursday, January 25, 2007
Batted Balls at Fenway Park
A researcher can be quite content to spend half his time analyzing the effects of Fenway Park. It has all the parameters that makes it interesting to study. Plus, it’s a fantastic place to see a game. A great marriage of analyzing cold hard numbers, and then appreciating baseball in all its glory.
But, we need help:
1. I am shocked to see that a picture of Fenway did not convert well to an image. That’s according to John Walsh, with respect to MLB.com.
So, I went looking at Fox Sports. Opening up the image in MS Paint, I get these coordinates for the Wall:
116,319 to 233,191
The RF line is:
353, 431 to 470, 321
What does that mean? The Wall’s x-coordinate moves 117 pixels, while the y-coordinate moves 128 pixels.
The RF line’s x-coordinate moves 117 pixels, while the y-coordinate moves 110 pixels. So, they are definitely not completely parallel. In order for them to have been completely parallel, the Wall’s second xy-coorinate should have been 233, 209. The Wall on the Fox Sports site is positioned 18 pixels farther away (about 14 feet).
This is in direct contrast from the Walsh image which shows the Wall much CLOSER than parallel.
Looking at the Google image, if the Wall is not directly parallel to the RF line, it’s pretty darn close.
So, in both the cases of Fox Sports (who I presume get their images from STATS, Inc) and MLB.com, they are off, and in opposite directions.
2. There are millions of people who watch and love baseball. Presumably, everyone who created the baseball PBP recording systems also love baseball. Why they track batted ball the way they do is beyond me. It’s as if they see it from the viewpoint of some lackey who knows nothing about the business.
If you are going to record the batted ball location, any baseball-loving fan knows that the ball does not go in a straight-line, and in one direction. How should a batted ball be recorded? With reality. The way to do it is to *always* record the absolute most you can, and *then* decide how you can cut corners, while trying to still record something that approximates reality.
So, work with me. A ball is hit, what happens? It lands 50 feet from home plate, at which point it then lands 20 feet farther, at which point, it is touched by the SS on a downward descent. So, you have *three* xy-coordinates to capture, and possibly an extra trajectory and z-coordinate (3 feet from the ground, on a downward trajectory). The whole time (the fourth dimension) that took was 2.2 seconds. This represents reality. And, if you convert a graphical image into a textual representation, you’ll end up with something like I’m describing. And yet, all we get from PBP is:
x,y=75,90
type=GB
Doesn’t it matter if it was a 1-hopper, or 4-hopper? Doesn’t it matter if it took 0.5 seconds, or 3.0 seconds?
It gets worse when the ground ball goes through the infield. Some PBP records the location of the ball as it is picked up by the left fielder! Isn’t that insane? Others record if the ball hits something (like a wall), and records that position instead of the fielder pickup. So, you’ll never really be able to draw a line back to home plate to figure out through which IF zone the ball went through.
Then comes Fenway’s Green Monster. Don’t you really care where on the wall the ball hit? Isn’t that a critical piece of information to have? Did the ball hit 5 feet from the ground, or 25 feet?
Again, simply represent reality. It hits the wall on a certain x,y,z point in space, and then lands on an x2, y2 spot on the field, at which point it’s picked up at spot x3, y3, z3 by a fielder. Time to wall is 4.5 seconds, and time to fielder is 2.2 seconds.
Is that so hard?
Greg Rybarczyk at Hit Tracker has plans to do something similar on all BIP, speed off bat, horizontal and vertical trajectories based on his HBT article. I read somewhere he was going to start with 1 park for 2007.
No wonder Manny is always rated as the worst OF. Look at all those balls hitting 10 ft or more off the wall that he didnt catch which a normal OF in any other park would have caught if the ball landed at that x-y location.