Of the people I asked, the older more politically conservative people opposed it because of the perception that it would make it more difficult for their party of choice to win (I heard this argument on both sides of the political spectrum) and the younger, more liberal voters opposed it because it "wasn't good enough". Less than half of the population even bothered to participate. So we're still stuck with FPTP.
The only chance we've had to completely overhaul our election system and people voted overwhelmingly against it. Ho hum...
Here are a couple of small tweaks that will help a bit:
1. Ditch the electoral college, elect the president by popular vote. There's a concrete proposal for doing that, the National Popular Vote Interstate Compact: https://en.wikipedia.org/wiki/National_Popular_Vote_Intersta...
2. Fix gerrymandering in congressional districts and state senates. Again, there's a concrete proposal, which recently won an important legal battle in Wisconsin (Whitford v Gill): http://www.fairelectionsproject.org
Neither of those helps bust open the two-party system, but they'll at least help level the playing field so that the big two parties don't have artificial incentives biased towards special interests.
Fixing voter suppression is very important. Anyone know of any good efforts afoot there?
As for third parties, I think that's probably best done first at the state level. If one or more states adopt a more proportional system internally, and show it can work well, that would provide a blueprint for implementing it nationally.
I guess Trump could get away with some really radical proposal, like using PR for the Senate. Both parties would almost certainly resist that though. He would need substantial popular backing, whereas every bungling decision he makes right now is making him less popular.
Since the latter is what is empirically shown in real democracies to be strongly associated with both satisfaction with government and ability of a system to support multiple viable parties and multidimensional debate on policy in the public sphere (see, e.g., Lijphart's Patterns of Democracy), I'm as disinterested in wasting time with the former as I am with wasting time with any other method of rearranging deck chairs on the Titanic of American democracy, like fiddling with how district lines are drawn for single-member districts.
(It also shows that IRV is only marginally better; while "worse is better" sets my teeth on edge, if IRV is the doable option I'll take it rather than wait for people to come around on score methods).
Most simulations assume consistent mapping from preferences to ballot markings, which is reasonable for ranked, but not score, voting.
For example, if the scoring range is 1-10, it's better for me to use the full range even if none of the options on offer is really a 1 or 10. Using the full range gives my opinion more weight in the result.
I strongly suspect score voting tends towards ranked voting when people vote tactically, which they always will, just with extra noise and randomness. (I don't know of any studies on this, though.)
There are many versions of it, but let's go with Maine for now, since it got acceptance and will soon have a track record.
[1] https://ballotpedia.org/Maine_Ranked_Choice_Voting_Initiativ...