From: Jeff Bone (jbone@jump.net)
Date: Tue Nov 28 2000 - 15:40:08 PST
cdale@silly.techmonkeys.net wrote:
> If the original counts are wrong, then how are they the only valid votes?
Because they're presumeably wrong in a uniform way across all the counties involved. If you accept
that *any* count is going to be wrong, then the goal is to minimize the wrongness and normalize it
across all participants so that the result is representative of the actual situation --- that's
called "fair," it's a technical term, perhaps you've heard it before? ;-)
> Why are there differences in the counts, if the machines are accurate
> enough to count them?
Because *any* measurement of a large enough dataset, of any sort, has built-in inaccuracy. This is
more than just an observation, it's a statement of physical law at the most fundamental level ---
Heisenberg and all that. Now, clearly, the problem here isn't quantum mechanics, but it *is* a
measurement problem. Anybody with a minimal science background can tell you that you don't go and
massage your datapoints in your problem areas, using different instruments, just to make the data
fit your curve and say what you'd like it to say.
> If they were 2 or 3 or even 50 off, no biggie, but
> that's not the case.
The only way you're going to convince me that *any* deviation is a problem is if you can demonstrate
that the difference between machine and hand counts in the disputed counties differ substantially
from the difference between machine and hand counts statewide. But we can't do that, because we're
not examining ballots statewide.
Look, I'm not in principle opposed to recounts. Every vote *should* count --- and count equally.
My issue hinges on the latter; there's no way to conduct a fair recount in Florida unless it's
conducted statewide, using the same measurement instruments / standards.
> If the machines are bork, count em by hand.
Fine. Nobody's proven, though, that the (i.e.) Miami-Dade machines are any more bork than the
machines up in the panhandle. If that's the case, fine --- recount *everything.* If all the
machines are roughly equally bork, who cares? The deviation's going to be the same statewide.
jb
This archive was generated by hypermail 2b29 : Tue Nov 28 2000 - 15:46:54 PST