Why 538 is wrong polling Trump at 31.5% (Nov. 2016)

web.archive.org

3 points·fdeage··6 comments

"It’s great to build a complex model and load it with empirical data like polls, economic reports and presidential approval ratings. But if the output of that model is implausible, it’s time to go back to the drawing board.

As a financial analyst at an investment bank, or a research analyst at an economic consulting firm, your job would be in serious jeopardy if you produced 538’s model output without a clear explanation of how those fat tails that represent an inordinate number of close to impossible scenarios could actually occur. A model like that just isn’t client-ready. Time to re-think those assumptions!"

It’s important to remember that OP want alone. Virtually all pundits gave trump less than 25% chance of winning—most of them far less.

There are lots lessons to take from this. I’m not sure which is most important.

It's true that most polls were wrong, and the HuffPost isn't the only one to blame here. But the author seems to imply that 538's _process_ was wrong because they were getting different _results_. For the record, the HuffPost gave Clinton a 99% chance of winning.

Where does the headline come from? I don’t see it in the article.

I think you are confusing “chance of winning” (as I recall 538 was give Trump about a 1 in 3 chance of winning in the days up to that election) with “share of the vote” (I don’t think anyone was saying Trump would only get about one third of the votes cast).

My mistake, I got the meaning of "polling" wrong (not a native speaker). What would be the proper verb then? I know we're not suppose to editorialize, but the original title ("What’s Wrong With 538?") seemed less informative.

“Why 538 is wrong to give Trump a 31.5% chance”?

Reply on news.ycombinator.com