Hacker News Reader

Comment by UncleOxidant | Hacker News Reader

The polls were right in 2016 in that Clinton did win ~3Million more votes than Trump. The problem was with polls in some states that ended up being crucial for an electoral college win. Hopefully 538 (and other similar organizations) have learned something from that and have made some adjustments.

View on HN

That seems outrageously obtuse. The electoral college is fundamental to predicting the US presidential election. The idea that 538 got it right but just didn't factor in the electoral college is a real stretch. If they screwed up polls in states that were crucial for an electoral college win then they screwed up full stop.

538 et al don't, as a general rule, run their own polls. In 2016, some swing states were poorly polled, so they were flying kind of blind on those.

And, of course, a few tens of thousands of votes different on the day, below the margin of error of any poll, and no-one would be bothered about them. In a way, 538 can't win here; they said "there's a 30% chance of this thing happening", it happened, based on differences below the resolution that they could even see at, and everyone's annoyed with them.

If differences below the resolution that you can see can cause the opposite outcome to happen, then perhaps an estimate of "30% chance of this thing happening" should be paired with a "1% chance of that estimate being correct".

Especially given the fact that "538" (electoral votes) is right there in their name.

How can you possibly say they got it wrong? What evidence do you have that Trump didn't have a 30% chance of winning? Your sample size of 1?

Reply on news.ycombinator.com