Hacker News Reader

Comment by aerovistae | Hacker News Reader

I sort of lost faith in percentage guesses the night of the 2016 US elections when I watched the NYTimes prediction flip from >95% Clinton wins to >99% Trump wins in a matter of seconds.

It really opened my eyes to how illusory the utility of the percentage was - if it can flip that hard that fast, clearly it was never really only a <5% chance of happening, and the guesser really didn't know anything meaningful. It's astrology for the educated.

View on HN

If bad predictions make you lose faith in percentages as a whole, why don't lies make you lose faith in words as a whole? In both cases I would assign the fault with the errant speaker.

While the change in who was favored to win the 2016 election was indeed large, it didn't change that dramatically and didn't change in seconds. The NYT 2016 prediction chart is still available on their site: https://www.nytimes.com/elections/2016/forecast/president

The labeling on the horizontal axis on the chart is terrible, but roughly this is a change from 20% to 95% in about 2 or 3 hours. It doesn't look like they ever had it at less than 15% (again, give or take a few points due to bad infographic-esque picture).

Because their predictions were based on how correct polling was, and once they found gross errors in the polling they adjusted.

> I watched the NYTimes prediction flip from >95% Clinton wins to >99% Trump wins in a matter of seconds.

Did that really happen? The numbers I remember before the election were about 67% chance of Clinton, though I don't recall the NYT. On election night, I recall Clinton not doing well from the start.

Reply on news.ycombinator.com