Most predictions markets and election forecast models predict Trump will lose. One might be inclined to say, “Ah ha! See, democracy works!”
Two problems with that:
1. HRC isn’t that great. Using the enlightened preference method, we see that HRC diverges significantly from the policies a fully-informed public would want. Yes, she’s closer to those policy preferences than Trump is. But, still, if the American voting public were fully-informed about basic information, we’d either have someone much better than HRC, or HRC herself would be pushing a better set of ideas.
2. Suppose a dad was choosing a nanny for his children. He narrows it down to two candidates, whom he’s still interviewing. One of them is mediocre. The other appears to be grossly incompetent, frequently says horrible things, has a terrible record of dishonesty, appears to be abusive, and has terrible ideas for raising children. Right now, the dad admits he’s got about a 55% chance of picking the better nanny over the worse one. You probably wouldn’t conclude that the dad is doing a good job and “works”. You’d conclude that even though he’s more likely than not to pick the better of the two finalists, the facts that A) the second finalist is obviously terrible, and B) he’s still giving serious consideration to this finalist, shows that there’s something seriously wrong with the dad. Even if he ends up picking the better candidate, there’s still something wrong with him. “In the end, I decided to pick the mediocre nanny over the child-abuser,” isn’t a display of competence.