Hand counting of ballots works well when you're only voting for a single elected office. In the US, most elections involve voting for many different offices all on the same ballot.
Really, we should just be switching to using scantron style ballots everywhere. They're paper ballots which can be efficiently counted by a machine, and are super easy to manually recount later
> Really, we should just be switching to using scantron style ballots everywhere. They're paper ballots which can be efficiently counted by a machine, and are super easy to manually recount later.
Better yet, switch to Scantron style ballots with some clever mathematics in how the ballots are made, and some clever chemistry in how they are marked, and you can then make it so that all the ballots can be published afterwards, any voter can verify that their vote was counted toward the correct candidate, anyone can verify that the totals for each candidate match the vote, a voter cannot prove to a coercer that they voted the way the coercer wanted, and any voter can do a check before voting to try to catch shenanigans with being given tampered with ballots, all without requiring any changes to the Scantron counting machines so that there is very little addition cost for all of these protections--and for voters that don't care about verifying that their vote was correctly counted it works just like plain Scantron voting--fill in the bubbles next to the candidates you want to vote for.
Unfortunately such systems assume that all of the parties in the election are working in good faith towards finding out the true result. They don't deal with the threat model of one (or both) major parties deciding to deliberately cast doubt on the outcome of an election by lying about or losing one of the secret keys needed to check the results.
Even if all the parties behave correctly, it's still not clear that such a system is really solving the underlying problem. If someone claims there were thousands of fake votes sent in by dead people, are you really going to reassure the public by saying "Don't worry, we've re-calculated the modulo-n equations using the cryptographic key's bitstream acting as a stream cipher"? That's just asking for the backfire effect to kick in.
As a coercer, I'd expect the voter to provide me with their verification data before the results are released. Then, I can verify that they voted like I told them to
I think the "verification data" is a string of characters that the voter can read off the ballot and either commit to memory or note down on a piece of paper they bring with them.
Importantly, though, there are as many strings printed on the ballot as there are candidates, so there is nothing stopping the coercee from noting down the string that corresponds to the vote they were told to cast, without actually casting it that way.
I'm not sure how this still gives rise to the claimed properties of the system, but the inventors did at least take the coercion problem seriously, from what I remember.
The UK uses hand-counted paper ballots and often has multiple elections on the same day. You just have a separate ballot paper for each election.
It's not unusual to have votes for some combination of Member of Parliament, Regional Parliament, City Council, Local Mayor and Police Commissioner all on the same day. Members of the European Parliament too, until recently.
It really doesn't take that long. Polls close at 10pm and first results can come out around midnight. Depending on how close the result is, the winner is often known around 6am the next day and the new Prime Minister can be in office by the afternoon.
Really, we should just be switching to using scantron style ballots everywhere. They're paper ballots which can be efficiently counted by a machine, and are super easy to manually recount later