I get that Silver likes to apply stats analysis to sports and politics, and that he personally enjoys both.
I don’t. I’m just interested in the politics, not the sports. The sports stats articles just take up space that could be devoted to politics, in my view.
Nor do I find the bletherings of him and his fellow 538ers very interesting. Their podcasts and round tables are just so much wasted band space as far as I’m concerned.
A few months back I was checking 538 daily. Then it started dropping off to every other day. Now, I can’t remember the last time I looked. Towards the end, all I did was look at his map and polls-plus and then leave.
Now I tend to go to electoral-vote, check out the map there and then surf the summaries of political stories it mentions. From there, I check out various other news sites.
Electoral-vote is particularly appealing to me because it’s politics and nothing but, unlike Silver’s weird mish-mash of “anything that can be measured by stats must be equally interesting to everyone, right?”
Anyone else finding Silver’s web-page less than interesting?
Neither of these are same. It’s more akin to when the History Channel started having non-history shows. The channel was originally for one thing, but then went into other topics. FiveThirtyEight’s very name shows that it’s supposed to be a politics website, and now that actually is a minority of the content (albeit the largest plurality).
If the SDMB added a blog forum for teens, and that started overwhelming the current usual content, it wouldn’t be unusual for people to leave.
Though, yeah, I do recommend just always opening the Politics tab. No need to read the other stuff if you’re not interested. If there’s not as much content that you like, well, you can vote with your ad dollars by going there less often.
Personally, I find anything non-sports related to be fine. I just wish they’d put the Sports stuff directly on ESPN.com. Unlike the politics tab, there is no easy “not sports” filter.
I like the whole website and follow him on Facebook too. The only part that’s pissing me off is how they now advertise segments then you have to listen to the Podcast or watch some video. Dude, I want an article with charts and graphs! If I wanted to watch or listen to something I’d turn on the TV or radio.
I disagree. A TV channel has a finite capacity and adding off-topic shows inevitably reduces the core output. A website is essentially unbounded so adding sports in no way diminishes the politics content. It’s much more like adding a completely new channel within the same brand, a channel many people may appreciate but nobody is forcing you to watch.
His lost a huge amount of my faith by being so completely off regarding Trump in the primaries. Why should his predictions for the general be of any use at all? The range of uncertainty is “huuuge” this year and their methodology doesn’t handle this.
So the political data part is “meh”.
As to the rest, The Puzzler should be right up my alley, but it is incredibly inconsistent. Some require fairly hard Math and others (like the camel-banana one) utterly trivial.
Almost all of the rest is filler. Filler doesn’t pay the bills on the web. The ratio of ad clicks to content is crappy. A web site with less filler is more profitable. The cable network analogy doesn’t work.
“Uncertainty” might refer to two things: the number of undecided (and third-party voters) and the extent of variance in the polling numbers over time. The first kind of uncertainty is integrated directly into the model, since it is a real empirical reason for why the polls at this time are less reliable. See “Step 4” of the methodology explanation on the 538 website.
It is unclear whether the second kind of uncertainty affects the predictive power of the polls. Whether the election is swingy or stable, whoever is winning in the polls post-conventions wins almost every time. Because there is no clear data on whether past variance implies future uncertainty, they have not explicitly put it into the model. That said, this election isn’t especially “uncertain” on that metric. If you look at the polls-plus model which predicts conventions bounces and factors them out, this looks like a remarkably stable race, with Clinton have a 3-5% advantage since the model started.