Forensic analysis of overvotes to determine prepunching for B*?

In this link, a Cuyahoga voter says she was given a ballot prepunched for *:

http://www.freepress.org/departments/display/19/2004/1032

If the ballots can be examined by a forensic scientist versed in the signatures of mechanical devices, it might be possible to establish whether these ballots were all punched by the same punch machine. Why?

Let's suppose I wanted to prepare a bunch of prepunched ballots: I'm very likely to send them all through one particular machine, or maybe only a few. As mechanical devices, each machine may well have its own particular signature, just as typewriters had signatures based on how the letters where printed. Now, on election day, there are lots of punch machines, and B* overvotes should show a range of signatures, as should the Kerry overvotes. But if the B* votes on overvotes represent only one or a few machines, and the Kerry votes on the same overvotes show a larger number of machines, we have evidence of fraud, because it would prove that the ballots must have been prepunched.

Of course, there's the minor problem of getting access to the ballots. . .

In most (if not all) precincts in Cuyahoga County, "punch card" voting is done with a handheld stylus. There's no way it would leave any "signature".

It was a good idea - sigh.

but in practice this would be a nightmare to try to campare over large numbers.

The only way to efficiently punch large numbers of ballots in a short period of time is with, say, an older IBM punch card punch machine with a stylus closely resembling the normally-used hand-held punch.

Creating overvotes is actually the quickest way to ruin the other candidates votes and have some small measure of plausible denial in the case of a recount--much like Florida--i.e., "look at all these spoiled ballots," but time constraints would require the use of a machine for the effect to be significant enough.

One would only need to take a random sampling of just the overvotes and examine them for signs of machine punching (experts say that such is identifiable). That might encourage another random sample of the remainder. If the results were consistent, that would show voting fraud.

Cheers.

http://cardamation.com

Click on the link for card punchs

either a FULL hand count, (but how do you re-count ballots not there?) or how about a Nation wide RE-Vote for President ONLY?
Sounds right to me. . . . .