3 ms·
Disclosure: I am a Nate Silver fan. Nate's expert opinion is only valuable if you accept the premise, which is that prediction science involves a range of poss
by jonbarker 8y ago
Disclosure: I am a Nate Silver fan. Nate's expert opinion is only valuable if you accept the premise, which is that prediction science involves a range of possible outcomes. If you accept that premise, which I do, his opinion on the 2016 election, in which he stated several times that the odds of a Trump presidency were actually higher than most of the polls suggested, was arguably a more valuable opinion than either a coin toss, or most of the individual polls.
- null0pointer 8y agoYeah this seems to be the disconnect. A lot of people will intuitively think (myself included) something like this. "You said Hillary had 90% chance to win, but Trump won. Therefore your prediction was WRONG". It actually takes a little more thought to understand that a Trump victory was a possibility in the model all along. And that you cannot say whether or not a prediction model is accurate based on a SINGLE result.
- aninhumer 8y agoMoreover, if someone makes a lot of predictions of people having "90% chance to win" and all of them win, that is a less accurate prediction than if roughly 10% of them lose.