I don't understand. How do they know how 100 random ballots were tabulated?
I mean, I assume there is a pile of ballots and a total for that pile but the only way they can know how each one was tallied is if it is a very small pile, or if the software keeps an image of each ballot and also the software interpretation of how it was tallied.
So I am guessing the software has feature where you can see each scanned image and "audit" that ballot.
If this is the case, then they are testing the accuracy of the optical recognition system. But they are not testing the accuracy of the tallying, or any algorithms, or ballot stuffing, or dead voters, or manufactured ballots, or mismatched signatures, or harvested ballots.
This would be only the tip of the iceberg. Am I reading this right?
More important, this is a very small sample and if this is all they verify, there is no guarantee it will reverse things - They need a hand recount and a forensic audit of all ballots and envelopes.
I think you are saying the barcode links to a scan, and the software tells how that ballot was tallied. Otherwise there is no way to know which votes in the total come from which ballot.
Even if they go through the pile and get the physical ballot, it should be no different from the scan. All they are checking is the scan, and how it is read. Not the algorithm, etc. They are still relying on the software to self report.
I don't understand. How do they know how 100 random ballots were tabulated?
I mean, I assume there is a pile of ballots and a total for that pile but the only way they can know how each one was tallied is if it is a very small pile, or if the software keeps an image of each ballot and also the software interpretation of how it was tallied.
So I am guessing the software has feature where you can see each scanned image and "audit" that ballot.
If this is the case, then they are testing the accuracy of the optical recognition system. But they are not testing the accuracy of the tallying, or any algorithms, or ballot stuffing, or dead voters, or manufactured ballots, or mismatched signatures, or harvested ballots.
This would be only the tip of the iceberg. Am I reading this right?
More important, this is a very small sample and if this is all they verify, there is no guarantee it will reverse things - They need a hand recount and a forensic audit of all ballots and envelopes.
Not sure if you are agreeing with me.
I think you are saying the barcode links to a scan, and the software tells how that ballot was tallied. Otherwise there is no way to know which votes in the total come from which ballot.
Even if they go through the pile and get the physical ballot, it should be no different from the scan. All they are checking is the scan, and how it is read. Not the algorithm, etc. They are still relying on the software to self report.