On Mon, 27 Sep 2004 17:26:49 -0400, Erik Reuter <[EMAIL PROTECTED]> wrote: > On Mon, Sep 27, 2004 at 04:19:33PM -0500, Horn, John wrote: > > >From what I have read here, it would only be more consistent to the > extent that the umpires can't tell the path of the ball. It seems there > is still a judgement call about the vertical position and height of the > strike zone. Now, if you also had a prior agreement on the vertical > position and height of the strike zone FOR EACH PLAYER among umpires, > then consistency would be better. But this could be done even without > the technology being discussed. > > I think I am not expressing my point very well. Briefly, it seems to me > the real problem isn't technology or lack of it, but rather agreement on > defining the strike zone for each player.
As I see it, an umpire needs to judge both were the strike zone is for a player, and if the ball passed through it. Either or both of those judgements can be mistaken or biased by the ump's respect for the player or pitcher, etc. With the machine, the question of bias is removed, as is the ball's position, leaving only the strike zone question. My guess is the system uses the side cameras to analyze the batter's stance and pick out notable points like overall height, the elbow and shoulder heights and knee bend height. From these, it could calculate what the strike zone should be to some reasonable approximation. Given that (despite the rulebook) real umpires' strike zones are quite subjective, I don't think that exact precision or an exact match to some umpire's preference is required. Of course, since the machine is being used to grade umpire's calls, that would have the effect of encouraging umps to match the machine's strike zone. Whether that's good or bad depends on how good a job it does and how consistent the machine is in picking strike zones. Both of those criteria should be testable, though, by analyzing the film footage and the machine's choices. _______________________________________________ http://www.mccmedia.com/mailman/listinfo/brin-l
