Barring mail-in ballots (the solicited ones) and from an in-person voting perspective only...

How is it that each ballot does not have a unique identifier in a database... separate from any unique voter information? It has nothing to do with who voted or where, nor who they are voting for... preserving anonymity. The voter is merely the next off the stack based on how they arrived to the polls (not including midwest "you go first" courtesy).

There would be a pre-determined number of identifiers printed and sent for each precinct based on the registered voter count + some padding within reason. Provisional ballots if we have historic turnout...

There would be a separate count for the number of voters in person, with no unique information attached. If this voter count is not reasonably consistent with the number of unique ballots used at a given precinct, does that not warrant further investigation?

Dealing with any excess provisional ballots beyond the unique number would be easier than the shit-storm of ballots appearing randomly now, wouldn't it?

Run the same ballots several times? Tough shit, it's been counted and ends up a key collision, resulting in said ballot counted once. Flag it and figure out what happened. If someone votes and the vote with an identifier has been hijacked (or "voted") already... shit's fishy and warrants legal intervention.

As a programmer... this boggles my mind. These duplicate ballot flags could be added to a database separate from the vote database to be reconciled one way or another and could hopefully turn up rampant fraud.

Maybe I'm crazy, maybe I'm drunk... why not both? Thoughts?