Hello. My name is Ben Welsh. I'm an Iowan living in New York City.
I am a reporter, an editor and a computer programmer. My job is to use those skills, together, to find and tell stories.
I work at Reuters, the world's largest multimedia news provider, where I founded the organization's News Applications Desk. In that role, I lead the development of dashboards, databases and automated systems that benefit clients, inform readers, empower reporters and serve the public interest.
[...]
- https://github.com/palewire/first-python-notebook
TBH I enjoyed looking up Ben and finding out what he's about and done in the past far more than I did just knowing there's a 538 archive on IA.
> Thousands of FiveThirtyEight articles seemingly vanish from the internet
https://www.editorandpublisher.com/stories/thousands-of-five...
And discussions here on hn:
ABC News has taken all FiveThirtyEight articles offline https://news.ycombinator.com/item?id=48152553
Disney erased FiveThirtyEight (article by Nate himself) https://news.ycombinator.com/item?id=48197703
https://web.archive.org/web/20230205124354/https://fivethirt...
It's kinda sad to know no one else will get to experience those interactive visualizations. Though its nice to see the approval comparison page still works
https://web.archive.org/web/20241031232233/https://projects....
After that election, a certain group would tirelessly work to discredit him any time his election predictions were not entirely one-sided.
The downside is this furthers the divide between folks who pay for subscriptions and masses who get shoveled ad-powered slop.
https://en.wikipedia.org/wiki/Electoral-vote.com
> Electoral-vote.com is a website created by computer scientist Andrew S. Tanenbaum. The site's primary content was originally poll analysis to project election outcomes. Since the 2016 elections, the site also has featured daily commentary on political news stories.
https://blog.archive.org/2017/04/17/robots-txt-meant-for-sea...
https://blog.archive.org/2018/04/24/addressing-recent-claims... which is a year later mentions that they have an automated process which is still following robots.txt for displaying old pages where the robots.txt was added later.
https://help.archive.org/help/using-the-wayback-machine/ does say they follow it for scraping, but this is phrased in such a way that would still be true for past sites whether or not they changed the policy. There is a page https://www.sysjolt.com/2021/archive-org-no-longer-honors-ro... which claims they don't follow it, but the site owner misspelled "robots" as "robot".
https://web.archive.org/web/20140701122958/http://fivethirty...
This is just some Disney suits being extraordinarily petty.
"I did approach Disney a year or two ago, through my agent, about acquiring the remaining IP. ...
We were told to basically get lost: ABC was annoyed with my critical public comments about their management of FiveThirtyEight. It apparently wasn’t a long conversation, so I don’t have a lot more color to report than that."
Here are some numbers roughly in the right ballpark: during the Disney era, which lasted about 10 years, FiveThirtyEight published about 20 stories a week. Let’s say that each story took about 20 hours to produce between research, writing, graphics and editing.3 Do the math, and that works out to about 200,000 person-hours of work that ABC News just deleted.In another sense, it's a journalistic source with information and commentary on past elections. Even aside from the political context that muddies the waters around or outright denies results, matters of public discourse on the web should not be ephemeral or subject to the decisions of the publication - they should be archived.
I remember thinking they were the best data journalists out there, and they had some nice visualizations but did their other predictions actually hold up?
This is why people don't really buy the "but he had Trump at 30%, you just don't understand statistics" apologist line. Sure he hedged in the dying days of the campaign (a cynic might think to try to protect his credibility), but the tone overall was of a person who comprehensively failed to understand the mood of the country from beginning to end.
Which is a problem because these election predictions are not just pure "mathematical models" and "data driven" like 538 would have had you believe. What mathematical model should be used? What data should and should not be used? At some point those things are based on the modeller's understanding of reality.
But predicting an election requires a lot more than polling datasets and statistics textbooks. That's the problem that he made himself out to be an election prediction wizard, but really that was off the back of his good prediction in quite a bland and conventional election.
When things got slightly more spicy and reality diverged from his vaunted "models", his "data science" predictably fell in a heap. The worst thing is almost not even that he got it wrong, it's that he seemed incapable of recognizing that present reality was quite significantly different from the past data he had used to build his models. Even after being wrong in so many of these predictions. He just kept churning out these pieces about how Trump was probably finished this time.
> But predicting an election requires a lot more than polling datasets and statistics textbooks. That's the problem that he made himself out to be an election prediction wizard, but really that was off the back of his good prediction in quite a bland and conventional election.
> When things got slightly more spicy and reality diverged from his vaunted "models", his "data science" predictably fell in a heap
The models were correct in two elections - arguably three because a 30% chance means that an outcome will occur in thirty times out of hundred. That is not zero.
To the person who is running this LLM, please find better things to do with yourself.
I still think that’s about accurate. Maybe it should’ve been 40%.
Everyone forgets that it was a pretty close election. Clinton could’ve won without the Comey announcement.
> I still think that’s about accurate. Maybe it should’ve been 40%.
It wasn't accurate. This is something people misunderstand about these predictions. If the 2016 election was held 100 times, Trump would have won 100 times. It's not the same as rolling dice.
These election predictions don't say that. They say something like "the observations I have agree with scenarios that have Clinton winning, 70% of the time". Which is fine and correct as far as his data and model goes, but none of those scenarios were the reality he was trying to predict. They are all just figments of the model though. Getting down to the brass tacks, he predicted Clinton would win, and he was wrong.
Which is fine, we just can't know anything about his process from that failure. Certainly we can't conclude that it was "accurate", since it was not. If we had a good sample of elections where he used the same process and built up a good record then sure.
The 70% figure is saying “we know most of the information needed to determine what the outcome of the election will be but we don’t know everything so can’t be certain”. There is no process where you can know every factor that determines the result in advance with absolutely accuracy and I don’t know why people expect there would be.
[1] https://www.sciencedirect.com/science/article/pii/S026137942...
40k voters, that's really not very many. So it's hard to say whether Trump had a 30% chance of winning or 40% or whatever, but the election at most was a toss-up.
Many random events could have resulted in a different outcome.
