Hacker News Reader

Comment by MiddleEndian | Hacker News Reader

In theory, maybe. In practice, how do we do that? There have been zero successful investigations as a result of "flaky" voting machines in America to my knowledge.

And what are the incentives for using voting machines? They are harder to verify than people from multiple parties counting lines drawn on a sheet of paper. They are susceptible to conflicts of interest, general incompetence, and security risks.

View on HN

I've read a few studies in the past. What you do to experiment is send people to use the equipment, then quiz them about how they voted, and see what the equipment says they voted. Problems like the confusing "butterfly ballot" in the famous Florida election pop right out.

You can also look at issues like undervotes and overvotes from the actual election. And ask people if they have any complaints about the process.

This is all somewhat studied already. We should study it a lot more.

This is in addition to doing systematic post-election recounts of a small % of ballots to verify that the counting machines are working.

Finally, everything is a voting "machine", including optical scan cards. Even if hand-counted. You should measure how well they work.

Even if we somehow verify that reliable machines exist, how do we verify that the machines haven't been tampered with? So far we've seen all sorts of problems with the machines in practice, usually in favor of the party in power.

What's our incentive to switch to machines? Speed isn't a concern here.

Here's one audit scheme to verify that the counting machines are reliable:

https://www.verifiedvoting.org/resources/post-election-audit...

One good way to implement this is with paper ballots and optical scan machines. You get speed/less labor with accuracy.

But you still need to study that the paper ballots are reliably usable by all types of voters, in addition to verifying that the machines count accurately.

Paper ballots aren’t immune to issues. Like the Al Franken election and ballots in the trunk; hanging chads in Florida; etc. I don’t know what the answer is here. I do not trust electronic voting machines. Paper isn’t immune to issues. Personally I’d love to see a dual system: votes are recorded electronically with a paper receipt.

Paper ballots with optical scan machines are popular in states that care about these issues. Compared to your suggestion there are fewer machines, easier recounts and audits, and it's very easy for the public to understand.

Paper ballots with optical scan machines are the reason we have electronic voting now.

It was paper ballots and optical scan machines that led to the whole "hanging chad" situation.

And filling in a bubble with a pen or pencil isn't going to help. Anyone who read BBS G-philes in the 80's knows how to defeat a Scan-Tron.

... I wasn't recommending using the worst possible paper ballot. The modern version, that tests well, is where you draw a line to complete an arrow. And this modern version is the one that's becoming popular.

Re: hanging chads

This is a problem with overtechnical solutions already. Random mechanical points of failure. Pencil drawing a line is unambiguous.

Re: ballots in a trunk

This was an auditing/ procedural issue.

Re: receipts

Why not just turn in the receipts and count them. Why have a black box do the counting? If a machine says candidate A had a thousand forty votes and B had a thousand thirty, what are you going to do? How do you know if your receipt lied? If people keep their receipts, you no longer have anonymous voting. And what are you going to do, track down every voter? If those running the election keep the receipts, they should just count them manually.

Its not unambiguous, look at this ballot and tell me who they voted for. It tied a race in Virginia just last year.

https://mobile.twitter.com/virginianpilot/status/94358996960...

This is why I like lines instead of bubbles and pencils instead of pens.

Regardless, this is actually the fault of the user rather than machines. As far as I know you can still get a replacement if you do this.

Pencils scare me. Mutable paper trails seem less than ideal. I think I prefer a pen along with an option to request a new form in the event of a mistake. Optical scan for counting along with a paper trail for random sampling. I don’t know the perfect system, but a dual recording mechanism for back checks seems better than just relying on paper or just relying on machines.

The receipt is not to be taken by the voter. The voter will check that the receipt is correct and it will be deposited into the voting urn. Then, manual counting may be done for all precincts or by sampling.

Right. And in the paper-plus-optical-scan system, the receipt is the thing the voter directly marked. No machine between voter and ballot, no different thing to check.

The counting is still not trustworthy if performed by a machine.

Hand counting is important, because it involves many eyes. If you want to manipulate the counting on a huge scale you cant, if you have to worry about idealistic volunteers.

If you let a machine count, you loose these checks.

Reply on news.ycombinator.com