That this is the narrative that survived the election is one of the greatest indictments for our society's ability to engage in critical thinking.
The day before the election, the Huffington Post published a hit piece criticising Nate for overrating Trump's odds and inspiring panic. Huffpost predicted a 98.2% chance of Clinton winning, NYT predicted 85%, and Nate's model dared to give her only a 65% chance of winning.
Then the election happens, Trump wins, and the credible figure who gave him the highest odds is now lambasted from the other direction. He was so wrong to give Trump a chance that the mainstream media were publishing articles about it, and he was so wrong to not give Trump a 100% chance that it ruined his reputation. The moral: you literally can't win, because people are too fucking stupid to comprehend probability, period.
538 made thousands of forecasts of events they predicted to happen 30% of the time, and those events happened 29% of the time in actuality. Does that mean they got it wrong every single one of those times a 30% event actually happened? For a forecaster to be 'correct', is it necessary for events forecasted at 30% to never happen?
Here's the last archived version of the page:
https://web.archive.org/web/20250306183754/https://projects....
The idea that Silver somehow 'hit' in 2012 when he correctly predicted all states and 'missed' in 2016 is so juvenile I get second hand embarrassment whenever I see it.
But yes, I agree with you that it's surprising to hear people on Hacker News having the 180 degrees wrong impression that the general population appears to have taken away from the one thing normal people care about polling for: during presidential elections.
After reading this book, The Party Decides https://press.uchicago.edu/ucp/books/book/chicago/P/bo592160... , he was a big advocate of the idea that the "endorsement race" of state officials and unelected party leaders.
There was a whole "Party Decides: Endorsement Tracker" graphic and everything, but Trump securing the Republican nomination and eventually the presidency pretty conclusively showed that theory to be a relic of the past.
So the 538 election coverage that year was: - Party endorsements matter more than early polling (they didn't) - Hillary's up so big there's no way Trump can win (he did, and yes I know they didn't actually say that but that's what the layman saw)
(ironically the Party Decides thesis seems to have correctly predicted events in the Democratic primary that year)
[0] https://web.archive.org/web/20250306183754/https://projects....
They had a good article about how their predictions were much better than you'd expect, but obviously I can't link it anymore because ABC removed it.
But in the end people pick on Nate because he really enjoys being an asshole on the internet. It's far more about when he acts as a pundit, not as an expert on statistics.
Once again, a failure to understand probabilities.
"They gave heads a 50% chance and it came up tails. They're claiming success for giving heads more of a chance than anyone else but they were still wrong". 65-35 probability is squarely in the neighborhood of a literal coin toss.
That's why people get all weird about probabilities flipping from 49% to 51%, and not at all interested in - say - a probability going from 60% to 90%.
¯\_(ツ)_/¯
So in an election, that happens all the time. It just doesn't always happen in the race for president.
--Nate Silver (538 founder)
ABC seem pretty petty here.
It was definitely sad to lose 538 how we did, but G Elliott Morris has really stepped up to continue the spirit on his Strength in Numbers blog. It's the best data-driven US politics reporting out there right now imo. He also contributes to Fifty Plus One with Mary Radcliffe, and that's excellent too, reminiscent of the old 538 polling roundup stuff that went beyond just core US politics. Recently he started a podcast with David Nir of The Downballot, which is another solid resource for lower level races.
If his comment is accurate, then ABC’s decision is a clear lose-lose decision driven entirely by personal spite.
I think Nate has become better at this sort of stuff in recent years. But still, he said that Disney "hardly ever" interfered in their editorial process, oblivious to the implication.
(What it means is that Disney/ABC were perfectly willing to interfere in their editorial process, but rarely needed to since Nate said what they wanted him to say anyway).
Parent says because they “criticized our management of the brand”.
Big corporations don’t give a crap. They are 100% about the bottom line.
This isn’t “big corporation” behavior.
Any exec who operates that way should be shown the door ASAP as they are likely doing similar emotional management of other aspects of the business.
It seems silly at best to take that as a literal statement.
Not that it's always the same company doing both at the same time, but it's crazy 538 was just left to die. It was a very recognizable brand among wonky professionals, a very desirable customer base. It's not as if politics and sports have gotten less relevant in the world over the past decade. ABC's decision to toss this aside is baffling.
Much of the 538 alumni seem to be doing well, either independently or as part of a major organization, so I don't think much was lost overall. But I sure empathize with the folks who lost their dream job and ABC looks pretty bad for frittering away a successful business for seemingly no reason. Taking down these articles is nonsensical.
Guess we better back up their GitHub repos before that gets taken down as well
when i lived in SF i found al pastor at Tacqueria Cancun messianic
Still chasing that dragon to this day.
Things got worse after Disney had their first round of layoffs. Their problem was they weren't profitable outside the presidential election years when interest peaked in the general public. 3 out of 4 years only diehard election polling wonks tuned in.
In practice, of course, corporate bean counting doesn't work on that latency scale. There's always a middle manager seeing fat that can be trimmed today, harvest be damned. But the financial argument was sound.
If Nate Silver buys it back (for pennies on the dollar) and then makes it successful, it's embarrassing and makes ABC look bad at business.
Maybe that was the logic on ABC's part but it's ridiculously wrong given how much clear market demand there is for the 538 people and content.
It's a world of difference to the political standing of the ABC Vice President between "Nate Silver launched something and made a gazillion dollars" vs "Nate Silver bought FiveThirtyEight back for a song and made a gazillion dollars" even if Nate Silver did the exact same thing. In the second case, the ABC Vice President gets fired because he signed off on the purchase.
This is why long copyright is such a terrible idea. With long copyright, there is every incentive to sit on IP and do nothing with it because of political losses. With short copyright, the incentive is to do something quick because the copyright will expire otherwise.
I expected it would be resurrected outside the Pushkin network, but hasn't happened yet.
What I _don't_ miss is listening to podcasts on Pushkin. I had nothing against Malcolm Gladwell, but something about having his voice on every one of the network's very numerous ads became incredibly grating.
Gladwell also annoys me, so that didn't help matters.
<https://web.archive.org/web/20250305183642/https://projects....>
NB: one of my gripes about current / contemporary content management / publishing systems is that almost all of them make it really hard to find either a specific article or a particular day's version of a site.
NB2: I realise writing the above that HN is an exceptionally welcome exception to that rule, with its "past" link (<https://news.ycombinator.com/front>), which not only exists but is prominently placed (top bar, 3rd link of 8 content-based links) on the site.
Edit: nm it was definitely the burrito battle royale bracket. Big burrito couldn’t handle the truth being revealed about their restaurants.
I occasionally read articles on Nate Silver's substack but I'm still missing the breadth of 538conbined with the trademark data-driven analysis.
Strength in Numbers is very US politics centric, but Elliott also works with Mary Radcliffe on Fifty Plus One. That's a new hub for raw polling data (and averages) but they also do some broader polling roundup style stories that have a 538 feel.
David Nir of the Downballot is pretty good too if you are looking for information on the smaller races.
Outside of the SiN extended universe, Marist have a podcast called Poll Hub which is very light and fun and reminiscent of the cuddliest 538 podcasts before they started getting that weird contrarian podcast bro energy.
I've trialed a bunch of other sites since 538 went away and also checked in on the other alumni, but none of it outside of SiN-and-friends or the Marist quite hit for me.
i was a casual reader of 538 back in the day. his substack feels pretty similar, if smaller in scope.
Fortunately the Github is still up: https://github.com/fivethirtyeight
I need to mirror everything to keep it accessible when they decide to shut this down, too?
I loved that site, and referred people to it frequently.
ABC officially sunset 538 over a year ago (and laid off most/all of the staff).
Sometimes companies acquire upcoming competitors specifically to shut them down so that existing cash cow product lines can continue.
All this proves is when the press was deregulated to allow one person to own all the media they can afford brought us were we are now.
I wouldn't touch that with a ten foot poll.
The first time I noticed this trend was during one of the W elections. The exit polls for the whole country were spot on except in some republican controlled districts in swing states. In all the districts with a discrepancy, the polls showed a much stronger democratic turnout than the vote tallies. All the districts in question had electronic voting without paper trails.
I guess some combination of those factors makes exit polls unreliable. /s
That pattern has repeated for most presidential elections since about 2004, and always indicate systematic tampering that caused official vote tallies to favor republicans more than the exit polls did. The effect is only seen in places where the officials were republican, where the difference was likely to matter and where recounts were impossible.
I'll miss 538. Here's an epitaph:
> Freedom is the freedom to say that two plus two make four. If that is granted, all else follows. -1984 by George Orwell.
As you suggest, it's good that the alumni seem to be doing fine, although Harry Enten's commentary on CNN is not as thougtful as he was on 538 podcasts.
... and to think I thought I was being clever. I see this has been suggested already.
Maps for elections generally don't come out till a few months beforehand because there isn't enough polling data to make a meaningful prediction.
There's not much extreme about Nate Silver.
IIRC, he got to keep all of the models and etc from 538 which is kinda all that mattered about it. Like anybody is really going to lookup the 2016 election in 2030 but they're definitely want the model's output in 2030 for 2030.
Management seems to be clearly inept and if somebody wants to give you a wheelbarrow of cash for something worthless you're generally foolish to not accept.
He’s a sellout at heart, including his recent association with Polymarket.
Someone cool would have never sold out a website as good as 538 used to be. Now he’s more interested in profiting off of gamblers.
now you can actually bet on your beliefs and dont need to debate with anyone on whether the koolaid colored wave of choice will actually happen, or whether you are a slave to an algorithm induced hall of mirrors. note, the senate has been 50-50 for over a decade, its probably the latter
so the only intervention necessary here is on yourself, focusing on things you cant control while delusionally thinking this time will be different, over and over and over again
extracting value from partisan gullibility and pathetic power struggles is unironically the move
A better way to assess accuracy would be to bin predictions (0-10% chance, 10-20% chance, etc.) and see if the observed frequency aligns with the predictions.
In parallel your signal could be a data source adding to individuals contribution to price discovery
This is frustrating as a consumer. Any further insight, on the solution side?
https://www.macrotrends.net/stocks/charts/DIS/disney/net-inc...
$12B profit is obscenely large.
"It's okay set millions of dollars on fire because we have billions in this pile over here!"
Who's shareholders are the public.
> The shareholders aren't going to care
This is not a valid defense in court. You can't let "attitude of investors" override "sound financial decisionmaking."
(Internally I'm sure they could probably phrase it some other less negative way such as chance of people confusing the brand as still owned by them, etc) association
Regardless, a single metric is almost never sufficient to provide a well rounded analysis of anything.
As far as I can tell the fiduciary duty to make money for the shareholders is something that Jack Welsh of GE said enough times that people remembered it. However, I’m always interested in additional details concerning the history of this meme, and happy to learn more.
(Not legal advice. I'm not licensed in either state.)
The duty of care is otherwise known as the duty to be informed. And the duty of loyalty is otherwise known as the duty not to usurp corporate opportunities. I stated the law in Delaware, which is consistent with the law on the rest of the United States on these points.
You and I simply use two different sets of words to describe the only two fiduciary duties of an officer.
I know it can be confusing that many places exist and that the corporation has multiple customers.