January 9, 2013

Is Nate Silver "concerned that during future elections, the accuracy of your predictions will lull readers into a mindset of 'it has been foretold, therefore I needn’t bother to vote'"?

Yes, "a bit."

Did Nate Silver "enjoy getting the ire of pundits (not the few who actually critiqued your method, models, or assumptions, but those who just dismissed your work wholesale)"?
At some point in the last few weeks of the election, I guess I decided to lean into the upside outcome a little bit in terms of pushing back at the pundits in my public appearances — as opposed to emphasizing the uncertainty in the model, as I had for most of the year....

Stupid poker analogy: part of playing well is in maximizing the amount of value you get from a hand in the event that things go well, in addition to mitigating your losses if they don't.

18 comments:

campy said...

Not a problem. If people don't go to the polls, the fraud machine can cast their votes for them.

Unknown said...

Isaac Asimov noted something like this 50 years ago:

http://en.wikipedia.org/wiki/Franchise_%28short_story%29

Shouting Thomas said...

Voting is really, really overrated.

Like it makes a fucking difference!

edutcher said...

Campy's right, the Demos will warp the bogus voters wherever they want them.

machine said...

How many stayed home based on what Dick "the Genius" Morris claimed?

hmmmmm....

edutcher said...

Probably none.

How many stayed home because they knew the fix was in?

At least 4 million.

Nonapod said...

Obviously we should just let brilliant seers like Nate Silver choose our leaders from now own since he's never been wrong before.

Triangle Man said...

Silver's model was right, but it was merely based on the polls. Other simpler models had the same prediction. If Silver wanted to, I suppose he could lie about what the model says and call an election for someone else, but it would be pretty clear to other analysts that his call diverged from the underlying polling data.

Anyone who was surprised that Silver got it right wasn't paying attention to the polls, or was kidding themselves about the potential degree of bias in the polls.

machine said...

"How many stayed home because they knew the fix was in? At least 4 million."

What?!?!?

Anonymous said...

Silver has done well predicting two out of three elections.

Anyone who believes this gives him a lock on predicting future ones needs to read more.

Strelnikov said...

With great power comes great responsibility.

Unknown said...

Reminded me of a great Asimov short story. Then I got beat to the punch.

Still a great story.

SomeoneHasToSayIt said...

Nate Silver should be a bit more humble.

After all, all he really did is realize that, yes, the majority of voters are that dumb.

SomeoneHasToSayIt said...

Nate Silver should be a bit more humble.

After all, all he really did is realize that, yes, the majority of voters are that dumb.

Anonymous said...

ALL quants should be a bit more humble after the Crash of 2008. But it's generally not in their nature.

Anonymous said...

ALL quants should be a bit more humble after the Crash of 2008. But it's generally not in their nature.

Paul Z: Good point. Models work well until they don't.

Alex said...

I can see conservatives are still very butthurt over the election.

John Althouse Cohen said...

Of course, you've used an ellipsis to omit the part where he clarified: "(Nothing about the model design itself changed — just how I tended to talk about it.)"

So, for instance, he wouldn't change the fact that the model said Obama has an 80% chance of winning. He'd just realize there are different ways to talk about this, e.g. "Obama is very likely to win," or "It's quite possible that either candidate could win," and he'd make a self-interested decision about which of those true statements to make about the same probability. This is in contrast with most pundits, who can wildly vary their fundamental predictions for whatever reasons they want. Silver binds himself to a model that yields precise predictions, and yes, he has some choice about how to talk about them.