Reddit reviews Dilbert and the Way of the Weasel: A Guide to Outwitting Your Boss, Your Coworkers, and the Other Pants-Wearing Ferrets in Your Life

I probably break from the conservative majority here but I don't think HuffPo or NYT were deliberately lying when they made that 95% confidence interval about Hillary winning. Nor do I think Nate Silver's analysis was fundamentally wrong.

Nate's prediction was "incorrect" in the same way as me saying "I'm 75% sure you won't draw a spade from that deck" could be proven "wrong" on any one draw. The uncertainty of not knowing the deck is the stand in for a polling error here -- you can make an educated guess, but you can't know. Nate Silver pretty much said "Trump is one normal polling error away from winning" and then that happened. Crowing about landslide victory aside, the victories within the battleground states were not particularly large.

Idiots who assumed 70% was as good as a victory deserved to get the disappointment they got.

NYT and HuffPo read the same polls, but they assumed that the errors in each state would be independent -- sure Wisconsin might be an overestimate for Hillary but then Michigan might be an underestimate for Hillary, and what are the odds that the polls are all wrong in the same direction? Nate Silver assumed the errors would correlate. Nate's assumption was correct -- the errors correlated.

Having seen the analysis, I find it higher quality than people who confidently asserted Trump would win based on whatever intuitive feeling they had about the situation. It knew the limitations of its own analysis. Someone like Scott Adams who made weasel predictions like "Trump will win unless something changes" -- hilarious for someone who wrote this book https://www.amazon.com/Dilbert-Way-Weasel-Outwitting-Pants-Wearing/dp/006052149X -- comes across as looking better, but he had a narrative for if he was wrong that would have played almost as well.

So yeah, the people saying Hillary would win fucked up. No malice, just basic mathematical mistakes. Shockingly the guy who used to put money on the line with his odds-reading is better at it.

The people who "knew" Hillary would win were overconfident about their ability to predict in a world of uncertainty, and deserve to get their ass kicked.

Disclaimer: I read Nate's book "The Signal and the Noise" after the 2012 election and found it grounded my assumptions of everything he said after that, so I probably didn't read as much certainty or hubris into anything he said in 2016.

There is a Scott Adams book that teach you exactly this, Dilbert and the Way of the Weasel.

No problem and happy cakeday!

Weasel Book