6 ms·
I don't buy 538's new election model
- hnpolicestate 2y ago538 without Nate Silver is like Apple without Steve Jobs. Probably worse.
- threeseed 2y agoPretty tortured analogy made worse by the fact that Apple is on every metric a more successful company after Steve Jobs.
- SoftTalker 2y agoThey are also a company that exists outside of history books thanks to Steve Jobs.
- xnx 2y agoThey might be referring to the early 1990's Apple, before Steve Jobs came back.
- ggm 2y agoThe one experientially almost nobody in the cohort today experienced.
- jvanderbot 2y agoI wasn't a consumer back then but I absolutely know the analogy and followed the logic.
- ggm 2y agoYes it's a compelling analogy. I just find it striking this event is closer to the end of the Vietnam war 22 years before Job's return to Apple in 1997, than the 27 years from 97 to today. There are almost certainly HN readers who weren't alive in 97, and yet the Jobs story works much as "It became necessary to destroy the town to save it" vietnam era quotes work: they're part of the culture now. I doubt we have many 13 year old HN people but it won't be long before there's a cohort here who lived entirely in a post Jobs world.
- kaonwarb 2y agoSome things that matter can't be measured.
- danielmarkbruce 2y agoSeriously. A very wise man once said as much.
- cududa 2y agoYeah but also Nate went to work for a Thiel company so take whatever he says with that context
- safety1st 2y agoWhatever else may be the case, it was certainly interesting to learn the following: * That the model which 538 used up until 2022 puts Biden's chances of winning at 28% * That the new model (which puts the candidates at "about even") is owned by Disney Yeah, that tracks.
- threeseed 2y agoYou do know that the 538 model in 2022 was inaccurate as well. https://en.wikipedia.org/wiki/FiveThirtyEight#2020_U.S._elections https://en.wikipedia.org/wiki/FiveThirtyEight#2020_U.S._elec...
- itsdrewmiller 2y agoThere is nothing about the 2022 election at that link.
- uptown 2y agoWhy does anyone concern themselves with the popular vote, when it comes down to the electoral college? The popular vote seems like a pointless metric.
- readthenotes1 2y agoI think he's somewhat addressed that in the article by pointing out that state by State polling is more inaccurate then the national polls in part because it's done so rarely. So I guess using the national pulling as a proxy for state is something they feel is more accurate when they try to figure out the electoral college results. I would find it interesting if pulling in these days is actually able to figure out anything. In addition to the polling problem itself, we have two frontrunners who have probably garnered more "not that one" votes than any previous race in history.
- DemocracyFTW2 2y agoAu contraire, the popular vote should be the only metric that counts. The electoral college and the per-state weighing of votes are pointless and detrimental, as is the first-past-the-post system that has brought about the U.S. two-party system, a system that is almost as bad as a single-party one. Instead I suggest one should move to abolish the electoral college and instead count votes directly. The candidates should all be on the ballot. Then use Approval Voting where each voter gets to mark all candidates that they like. There should be a French-style second round in case the top winners are too close together, say inside a 1% or 5% margin—in other words you could never win by a single vote, you'd always need hundreds of thousands of votes.
- Ekaros 2y agoI would also go same way and remove states. Just move everything under centralized federal government. Or maybe keep them for some regional offices, but all deciding power should rest on congress and senate. And even there, getting rid of voting districts and just electing top x number of candidates seems most reasonable solution.
- 2y ago
- bitshiftfaced 2y agoIt isn't clear to me why we should weigh fundamentals less in relation to polling on the basis that it has a higher variation at this point of time in the cycle. Couldn't it be that it has higher uncertainty and it's more accurate (in terms of expectation) x days before the election, compared to polling?