Hacker News Reader

Comment by SilverBirch | Hacker News Reader

I think the funny thing is that Nate can be a flippant twit on Twitter but has totally fallen for it in critiques of him. Essentially this article is “a bunch of people baselessly speculated about an assertion I made, so I’m going to spend lots of time proving something they don’t care about”

View on HN

"Bunch of people" doesn't seem like a good characterization considering he's at least partially responding to Martin Kulldorff, a professor of medicine and biostatistician at Harvard and a co-author of the GBD.

A biostatistician who told this professional stats guy to go consult a statistician after taking a guess at a confounding variable. He could have run those numbers himself, but apparently didn't bother, because Silver took him up on it and Kulldorff was completely wrong.

It would be less egregious if it has been a bunch of random people. It was someone who really should have known better. Instead, it provides more evidence that Kulldorff is seeking to confirm an ideological position and tossing his medical and biostatistical accomplishments aside to do it.

> Instead, it provides more evidence that Kulldorff is seeking to confirm an ideological position and tossing his medical and biostatistical accomplishments aside to do it.

Nobody is immune to this. It’s just more frustrating when their ideological position is opposed to yours.

Nobody is immune to this. But he really did fail to do an easy test. Someone with his level of experience should have been able to make a challenge that wasn't trivial to refute, and being insulting in the process.

Especially since he happens to be in the thick of a position that is easily accused of being anti science. One would think that makes it doubly incumbent not to make such trivial errors.

As a bystander this seems accurate, but at the end of the article Nate himself states that he has little hope that it will convince people on Twitter/X. It's more about credibility to his own readership, I suppose.

Getting in an argument (vs disagreement) in real life is worthless. Getting into an argument on the Internet has even less value.

Personally I think this was an easy article for him to write because he already had the data and he wants more subscribers (please take this take as neutral opinion and not a dis to Silver)

It's worthless if your goal is to influence your direct opposition. It's not worthless if your goal is to influence the curious on-lookers.

Yep, always good to keep in mind that the audience for public discussion is not the small number of people you're discussing with, but the large number of people reading the discussion.

That depends on what the goal is. It's worthless if you're trying to convince the other side. But in many cases, you're actually trying to convince the audience, and certainly for something like a blogpost here, I'd argue that it is not only the case, but successful, since at minimum we know it got posted to hackernews and has traction.

I think he's writing to convince himself more than his readers, if he has any.

Look him up, he has readers.

FWIW, I think Twitter is a rabid cesspool that will be a big piece in the downfall of humanity, but I didn't really share your assessment. For whatever reason he decided to write this blog post, I thought it was really interesting, insightful, and I appreciated the detail and clarity of thought that went into it (e.g. the analysis at the bottom of the article that showed vaccination rates were the real factor that resulted in differences in death rates, and that partisanship was really just a proxy for that).

So, Twitter folk may not care about what he showed in this post, but FWIW I certainly appreciated it.

Typical narcissist. I'll always think of Nate Silver as the bedraggled man with a mid-range stare on election night 2016. Silver's utility is limited to knowing what the establishment talking points are.

Considering the incredibly close vote margins that carried 2016, Nate had one of the most accurate models of interpreting poll data available.

Or he got lucky, as next time he wasn’t even close in his predictions

By "not even close" do you mean he called 48 states correctly?

What should he have done differently to improve his analysis?

Reply on news.ycombinator.com