The home plate umpire wears a headset which reports to him one of five values: strike, borderline strike, borderline ball, ball, no data (system knows it has malfunctioned). The "borderline" is a buffer zone, probably an inch wide but maybe 1.5", centered around the edges of the agreed-upon strike zone.
The same data is reported in real time to both dugouts (and ultimately to fans) but in three buckets: strike, ball, borderline.
He is urged to called borderline pitches the way the system sees them but has discretion to do otherwise. The intent of the discretion is to maintain consistency with certain types of borderline pitches that the system sees one way but umpires have historically seen the opposite.
If the intent is just to maintain historical consistency, then I don't really see the point of the "borderline strike" and "borderline ball" calls. Rip the band-aid off and go all-in (choose your metaphor). If you think that there really is a "borderline" strike call that could fairly go either way (even though it does or does not meet the strict definition of "strike"), then I would just report "borderline" to the ump and let him make his call. There's also the virtue of keeping the system simple.
Also, why would you report the pitches to the dugout? All it's going to do is inflame a team one way or another based on the ump's calls when they won't be able to do anything about it anyways (what, protest the game?). What you really need is credible enforcement based on the accumulated data. But if MLB were willing to do that, then you might not have this problem from the outset.
Finally, an alternative to piping it into the dugout: Let's make this a true Survivor-style Umpire Efficiency Program. No headset for the umpire. He's allowed to make his call however he sees it. But the computerized call is displayed on the stadium scoreboard. If the umpire gets it wrong, the call is reversed. Now THAT is a program I can get with.
Edited by BroodsSexton, 06 October 2008 - 03:53 PM.