Locklin on science

My favorite photo of this wacky election

Posted in stats jackass of the month, Uncategorized by Scott Locklin on November 9, 2016

This dope got lucky in 2012, essentially using “take the mean” and was hailed as a prophet. He was wrong about virtually everything, and if someone were to make a table of his predictions over time and calculate Brier scores, I’m pretty sure he’ll get a higher score than Magic-8 ball (Brier scores, lower is better). Prediction is difficult, as the sage said, especially regarding the future. Claiming you can do prediction when you can’t is irresponsible and can actually be dangerous.
silver.jpg

While he richly deserves to find his proper station in life as an opinionated taxi driver, this clown is unfortunately likely to be with us for years to come, bringing shame on the profession of quantitative analysis of data. We’ll be watching, Nate.

 

 

Advertisements

12 Responses

Subscribe to comments with RSS.

  1. Huntington said, on November 9, 2016 at 9:22 pm

    Can you point to any write-ups that dive into the issues with his methods, and/or how he managed to get so lucky the last time around? Genuinely curious.

    • Scott Locklin said, on November 9, 2016 at 11:46 pm

      His model is inherently bad; the data feeding it was garbaged up (plagued by sample bias), and it didn’t account for hidden variables like trend, the fact that Clinton was running against a Putin conspiracy theory and a cartoon frog, or the “trump closet.” The way he used monte-carlo also didn’t make a hell of a lot of sense. They also assumed the population of people going to the polls was distributed the same as ever, which it obviously wasn’t. The more I’ve looked at his “modeling” efforts, the more contempt I have for the guy. Taleb rightly pointed out that Silver doesn’t even understand what probability means.

      As for why he got lucky: I perfectly predicted 4 toss up states and the election outcome, literally using a magic-8 ball for the entertainment of my friends. It happens (I got witnesses). If there were 100 people with magic 8 balls predicting this election, chances are, one of them would have done quite well. That would be the guy we remember. Silver was that guy in 2012. This is an old scam in the markets; you send out a bunch of different predictions to a bunch of people; some of them will hail you as an investment guru and give you money. The rest will forget about you. It’s also the basis for White’s Reality Check; making sure you didn’t over fit your model on limited data.

      • Huntington said, on November 10, 2016 at 12:53 am

        Sample bias in this case being systematic (weighted to Trump supporter) non-response to polls? How does trend affect election forecasting, and was Silver ignoring it or simply not incorporating it properly? Have you come across anyone who does the election forecasting in a more rigorous way?

        • Scott Locklin said, on November 10, 2016 at 2:29 am

          I bet Parabellum Labs and places like it do a better job. Certainly there was a reason Trump was campaigning in Michigan, even though nobody thought that was a sane thing to do.
          My magic 8 ball does a better job.

  2. William O. B'Livion said, on November 9, 2016 at 9:33 pm

    Pretty much everyone got this one wrong, and got it wrong for the same reason–no one sane wanted either the Clown or the Criminal in office, and those at the top of the food chain preferred The Evil that they knew (Criminal) to the insanity they couldn’t know (The Clown). And thus everyone from Paul Ryan to Nate Silver wanted the evidence to prove that The Criminal would beat The Clown (note that those labels are relative. The Clown is more clown than criminal, but is certainly criminal. The Criminal is mostly criminal, and no clown, but some other C words–Corrupt for one).

    Given the inputs as of June, this was probably the least-worst outcome. but that’s like saying basal cell cancer is better than having a melanoma.

  3. Bob Van Wagner said, on November 9, 2016 at 10:51 pm

    Rabbi Mendel Kessin called it months ago. https://www.youtube.com/watch?v=GbwrLkxzoaY&ab_channel=torahthinking

    • Scott Locklin said, on November 10, 2016 at 2:28 am

      Yes.

  4. brian said, on November 10, 2016 at 4:07 pm

    This guy got it right: http://michaelmoore.com/trumpwillwin/

  5. SerbSuperb said, on November 12, 2016 at 7:23 pm

    The woman in the back looks like she’s having war flashbacks while the other one is staring open-mouthed at her computer. Its so perfect it could seem arranged.

  6. George Okromchedlishvili said, on November 12, 2016 at 9:12 pm

    This dope still had better odds than most US pollsters/mainstream political analysts AFAIK.

    His success in 2012 obviously did not come from being some super statistician (he clearly isn’t) but just looking well compared to lots of “pundits” who used to get away with empty talk and little actual research.

    • Scott Locklin said, on November 12, 2016 at 9:15 pm

      There should be a pundit prediction app which keeps track of their statements.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: