Hacker News Reader

Comment by nverno | Hacker News Reader

“We have this fiction in our minds that voting is perfectly accurate: that even a differential of a few votes is meaningful. In reality, elections have error rates every step of the way, and a very close election is a statistical tie."

I'm curious about the error rates he mentions. What causes them in practice?

This reminded me of another article on hacking voting machines - https://www.politico.com/magazine/story/2016/08/2016-electio...

View on HN

People underestimate how accurate voting is. We can know how millions of people voted with an error rate in the hundreds or even dozens. That's almost six nines accuracy.

A bold statement, when we're discussing a country where many states have no paper trail whatsoever.

Where exit polls can be off by over 12% in those states.

And where the result of those discrepancies was to stop doing exit polls.

How does one even begin to approximate the accuracy of a process like anonymous voting, let alone claim 6 nines of accuracy?

I would guess things like - theoretically perfect election where every eligible voter votes instantly and magically in an instant, and then determine all the things that detract from that: late registration, sickness, distraction, and then add up all the tallying errors and such (none of which even considers if people are making the “true” decision they’d make if they had perfect knowledge and foreknowledge of the outcomes).

Elections are basically an emergency valve and a way to get people to not complain too much.

> all the tallying errors and such

These are the types of errors I assumed he was considering there- I guess calling a close election a statistical tie makes even more sense if you're accounting for people not voting/having voted without 'perfect knowledge' as well.

Even with perfect knowledge and mandatory voting, if you are trying to determine the will of the people then giving a choice of (usually) two quite similar parties doesn't really capture that.

I suspect you end up with two or three or ten similar parties because of the vast majority of the people don’t basically agree on most things, you can’t have a democracy and instead have a civil war instead.

> Elections are basically an emergency valve and a way to get people to not complain too much.

No, they're a collective agreement that we compete with votes instead of on a battlefield. Instead of getting recruited to be in some politician's army, you get recruited to be their voter.

It's a fallacy to think (or deflection to claim) that errors could corrupt the system when even without errors, the system is inherently corrupt and riddled with antidemocratic rules. Think gerrymandering, the fact that election day in the US isn't a national holiday, Citizens United, the joke that has become the Supreme Court, the list goes on and on. US democracy is a farce.

I don't think any of those things you mentioned, aside from the Supreme Court, are inherently part of the 'system', and I don't think the design of the US democratic system is inherently corrupt - in fact, most of the rules in place were designed with the intent of discouraging the inevitable corruption.

Individual problems can be attacked without dismantling the whole thing. Gerrymandering, especially, seems like it could have an outsized effect on the error rates of some elections.

I guess my point is, I don't think it's helpful to frame antidemocratic practices as errors when they are kept in place very deliberately by the powers that be.

I agree, I was thinking the error rates he alluded to were probably logistical or sources of human error.

To me democracy is like asking a bunch of 6 year olds what they want for dinner. It’s based in the assumption that people know the candidates and vote for the policies they agree with, a bit of contemporary history can tell us how well that works.

I'll bite. Vote harvesting and being able to vote without ID.

Reply on news.ycombinator.com