Nate Silver is partisan

I won’t go as far as saying that he is wrong but I agree that he is definitely partisan.

Another commentator who doesn’t understand basic math, statistics and probability. Even if Romney wins, Nate Silver is not “wrong.”

Oh boo hoo hoo. Cry moar.

The only thing I can tell from that article is that the author lacks even a basic understanding of statistics. He’s partisan because he might be wrong?

The comment regarding poll weighting is the only point that speaks to even potential partisanship sneaking into his models.

Hell, the whole piece is a general commentary that statistical models have vulnerabilities. I don’t get how you go from “its statistics” to “he’s partisan [in his methodology]”.

OP: care to add your reasons for agreeing with the article?

You forgot to mention that he might be gay.

Don’t like the polls showing your candidate losing? Shoot the messenger!

I mean, Tim Stanley shows his own partisanship by pretending like the Republicans don’t point to the polls when it suits their purpose, and he also pretends like the Democrats entire strategy is to point to the polls as a fait accompli. So right out of the gate the he is trying to deceive me. He either doesn’t understand the math or doesn’t want to understand it. Then he derides Silver’s extreme high degree of accuracy in the previous presidential election because he had access to polls many didn’t, which kind of passes over entirely that he was indeed, accurate. He also derides Silver for picking and choosing which polls he finds more accurate. The bottom line is accuracy. Any moron can run a poll, that doesn’t mean it is going to be as accurate as another. It’s just a partisan hatchet job, and a transparently incompetent one at that.

This guy must not follow politics too closely. Everyone does this. It’s how campaigns decide everything from which ads to run in which states to where to spend time campaigning, all in order to spend money in the most efficient way possible and maximize the number of electoral votes achieved.

Oh yeah, horrible article.

“The voters will decide Romney v Obama, not The New York Times” should have been enough to discredit this article.

My thoughts on Nate Silver

  1. Statistically he’s objective.
  2. He’s subjective in creating his statistical models.
  3. He’s clearly dem leaning.
  4. I believe that when he misses, his misses tend to cluster around under-estimating R’s chances. IE, it’s not on both sides of the confidence interval.

This pretty much tells you all you need to know about his understanding of “polls” and statistics. Only Rasmussen has PBO within statistical spitting distance in Wisconsin… and even they have PBO up by 5 in Minnesota. For the latter he had to find a single poll that just came out that still has the President up by three points, but since it was “within the margin of error” and closer than last month, he presents it as evidence for why all polling should be thrown out the door… and especially if it favors those cold, calculating Democrats.

This seems like something that could be proven one way or another, though ideally we’d have more than 1 presidential election year and 1 mid-term year to use as a sample.

A biographer of Pat Buchanan writing for the Torygraph and bashing the NYT is calling someone biased and wrong?

Wow, a person who has become deeply involved in politics has opinions about politics! Alert the media! Note to Tim Stanley: just because you are incapable of separating your ideology from your job (a historian whose most recent book is about Pat Buchanan? Surely he’s non-partisan!) doesn’t mean everyone is as incompetent and unprofessional as you are.

Edit: I just remembered something quite good that I read on the subject of Nate Silver, but I don’t remember where it was. It was pointing out that election forecasting is similar to weather forecasting. If you’ve got a 75% chance of rain, the accuracy of the model isn’t disproved by rain. Rather, you have to look at a long sequence of data to find out whether the model’s % chance matches the proportion of rain to sun over the whole set. In other words, Romney winning with only a 25% chance on the 538 model doesn’t invalidate the model, it just falls into that 25%.

Of course, the problem here is that elections happen way, way more infrequently than weather.

Plus more bias against faith-based anti-pollsters:


Is your empty chair lawn display racist?

Why are right-wingers so intellectually retarded?

Now now, not all of them are, just…several of them?

I wonder what Tim Stanley makes of the the 7 or 8 other meta-pollsters that are calling the election in much the same way? Are they all Democrat stooges? How deep does the rabbit hole go?!

I can’t even imagine what exactly would even cause someone to conceive of lynching a chair.

Also, what is with the one chair having a gun and a golf club taped to it?

I mean, maybe the gun is there because… well, they like guns I guess? But what is the golf club all about?