Me = not an expert on elections
Yet, at almost every turn along the path of the 2016 election, I was far more accurate than Nate Silver.
Is Nate Silver's expert opinion still more valuable than mine? Is expert opinion valuable even if it is right no more often than coin flipping?
Nate Silver is a statistical analysis expert, not particularly an election expert.
> Yet, at almost every turn along the path of the 2016 election, I was far more accurate than Nate Silver.
Really, for all the elections that year? What about 2008, 2010, 2012, and 2014?
> Is Nate Silver's expert opinion still more valuable than mine?
Where can we see your historical set of outcome predictions to validate their superiority?
> Is expert opinion valuable even if it is right no more often than coin flipping?
Nate Silver has been right more often than coin flipping.
In math, if you only write the correct answer, you'll get a zero for not showing your work. In debate and discussions, there can zero, one, or more than one correct solution to a topic, and "showing your work" is often more valuable than having an answer, and answers can be right or wrong explicitly depending on your work.
You got lucky. If you can consistently beat his methodology then you'd be just as famous and wealthy and well regarded as he is instead of an internet nobody.
This is like saying "college is for the birds, I just plan to win the lottery."
>Is Nate Silver's expert opinion still more valuable than mine?
Absolutely. Missing the boat on what turned out to be something of a freak-show election where 77,000 votes out of than 136 million ballots cast is tough for anyone to get right. And most of who, did so by dumb luck, not some wonderful new methodology that is useful against past election data or current election data. The very same people beating their chests over Trump's win predicted Moore would win Alabama or Saccone would beat Lamb.
In other words its just dumb partisanship, not some new polling or statistical technique at work here. Patting yourself on the back because you 'called it' for Trump is a meaningless statement unless you have the math to prove your methods are better than Nate Silvers.
Silver didn't call a close election, though. That would have been impressive. All he said was that it would be closer than a landslide.
Silver predicted a fairly high probability of a very close election, including a 10.8% chance Clinton would win the popular vote and lose the electoral vote.
> All he said was that it would be closer than a landslide.
No, Silver's model provided a lot more detail than that.
Most coins are 50/50, not 75/25.
Also, the answer entirely depends upon whether there's anyone out there who can do better than 50/50.