Looking just at one particular mechanism: how often will weather counterfactually impact the outcome of an election? Presumably only if the election is very close, but seems very plausible for e.g. the 2000 election, so it’s not that rare, though definitely not the norm. Conservatively call it one election in 50 sensitive to the weather? (I expect it should be much higher than that given how recent 2000 was, but I’m trying to get a lower bound here.)
… and I’d be very surprised if weather isn’t mostly-quantum-randomized on a four year timescale.
So I’d expect that at least one in fifty elections is counterfactually sensitive to quantum noise.
On the other hand, most elections are not that close, and I’d expect that even a moderately-stronger-than-human AI could usually guess pretty confidently in advance which elections will and won’t be that close.
For elections which aren’t unusually close, are the tail probabilities sensitive to quantum noise? I’d be surprised if they were; central limit theorem bites hard at the scale of elections. With a hundred million voters we’re talking about standard error on the scale of ten thousand, or 0.01% of the electorate… and e.g. the 2020 election had ~4.5% advantage for Biden according to wikipedia… so that’s hundreds of standard deviations. (Though admittedly that’s ignoring the electoral college, which could definitely have weird effects.) So quantum noise would have to have, not just local effects in a few places, but huge correlated effects over a large chunk of the electorate in order have appreciable effects on the election counts themselves.
Impact on individual candidates is maybe more plausible? Not sure what specific mechanisms would be relevant there. E.g. interstice’s suggestion of a candidate randomly dying seems probably less than 0.1% probable in a typical election? Based on a quick google someone 65-75 years old has roughly a 0.8% chance of dying over four years, it will be an OOM lower for younger candidates, and even for old candidates most of that probability will be packed in health problems which can be seen coming well in advance if one is looking for them.
Looking just at one particular mechanism: how often will weather counterfactually impact the outcome of an election? Presumably only if the election is very close, but seems very plausible for e.g. the 2000 election, so it’s not that rare, though definitely not the norm. Conservatively call it one election in 50 sensitive to the weather? (I expect it should be much higher than that given how recent 2000 was, but I’m trying to get a lower bound here.)
… and I’d be very surprised if weather isn’t mostly-quantum-randomized on a four year timescale.
So I’d expect that at least one in fifty elections is counterfactually sensitive to quantum noise.
On the other hand, most elections are not that close, and I’d expect that even a moderately-stronger-than-human AI could usually guess pretty confidently in advance which elections will and won’t be that close.
For elections which aren’t unusually close, are the tail probabilities sensitive to quantum noise? I’d be surprised if they were; central limit theorem bites hard at the scale of elections. With a hundred million voters we’re talking about standard error on the scale of ten thousand, or 0.01% of the electorate… and e.g. the 2020 election had ~4.5% advantage for Biden according to wikipedia… so that’s hundreds of standard deviations. (Though admittedly that’s ignoring the electoral college, which could definitely have weird effects.) So quantum noise would have to have, not just local effects in a few places, but huge correlated effects over a large chunk of the electorate in order have appreciable effects on the election counts themselves.
Impact on individual candidates is maybe more plausible? Not sure what specific mechanisms would be relevant there. E.g. interstice’s suggestion of a candidate randomly dying seems probably less than 0.1% probable in a typical election? Based on a quick google someone 65-75 years old has roughly a 0.8% chance of dying over four years, it will be an OOM lower for younger candidates, and even for old candidates most of that probability will be packed in health problems which can be seen coming well in advance if one is looking for them.