RWCG


Linkdump
January 5, 2013, 5:37 pm
Filed under: Uncategorized
  • Hipsters and coffee. If better-tasting coffee were really the result of hipster coffee hobbyists, I’d be the biggest hipster cheerleader on the planet. Mangan is right about coffee & cream however. I can’t/won’t drink it with anything less than half & half. The added benefit is you get some fat in the process. (I actively seek out fat in my diet.)
  • Arnold Kling writes something about probability that sounds like it must be an echo of the Nate Silver Wars of October/November 2012. Which at this point means reading it is like reading Civil War letters. He really needs to go back to blogging rather than these ‘essays’.
  • Subsidy gone wrong. There’s no doubt of course that the subsidies in question were written by Smart People who think of themselves as oh so very Smart.
  • Another study suggesting that not CO2 but the Sun might, in fact, have something to do with how warm or cool the Earth gets.
About these ads

5 Comments so far
Leave a comment

Dr. Kling expresses my objection, more-or-less, back in the Nate Silver days. He writes more clearly than myself.

Using his language, Nate’s model was a combination of problastat and problatherm. I don’t object to someone publishing or using such a thing, but I do object to its implied precision – suggesting that there was no subjectivity to the input assumptions.

Comment by Dave

As I wrote on Kling’s comments, his three different types of probabilities all boil down to the same thing IMHO.

That aside, what the Sonic Charmer Electoral Model™ showed was that even if Silver had made a bunch of ‘problatherm’ type assumptions, they (evidently) didn’t have a material effect on the model’s output.

You can quibble about the ‘precision’ and say that 75.3% should have been stated as ‘about 75′ or whatever, but there’s just less to that objection than meets the eye. You’re ultimately arguing about a few percent here and there. A reasonable answer might be: so what? It’s not like there were any reasonable assumptions you could cook into a model and come out with Romney instead of Obama a 75+% favorite. Right?

Comment by The Crimson Reach

Right, I agree, his outcomes were correct (modulo the cheating).

My objection is not about the outcomes of the model but the way in which they were portrayed.

Let’s say that someone purports to invent a machine to predict the price of wheat in tomorrow’s market. You obtain results from the machine, and find that for at least a few day’s worth of performance, they were close to being correct.

If the creator of the machine had to manually tweak some settings on the machine before it produced a result, would he be lying if he presented that information as a product of mechanical reasoning?

Comment by Dave

I guess. But remember how I blindly made a ‘dumb’ model with no tweaks to speak of, and got pretty much the same result? Didn’t that tell you something? There was a point to that exercise of mine (besides my Aspergery love of building spreadsheets ;)

Comment by The Crimson Reach

Yeah your spreadsheet tended to show that his model was a pretty obvious Monte Carlo.
I guess I am excercising my Aspergery tendency to label stuff correctly. If you are publishing an opinion in the form of a number, that is a different thing than publishing an approximation of pi.

Comment by Dave




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



Follow

Get every new post delivered to your Inbox.

Join 454 other followers

%d bloggers like this: