Hacker News Reader

I feel you are not taking into account the importance of "unless something changes". I see the publication date of this as being 1997. That was around the time when globally it started to be conceivable that restaurants and public areas and even bars could be areas that were smoke-free. Changing the norm seems to have lowered the number of smokers and thus smoking deaths.

You scoff at these projections but I feel that turning your energy a bit and focusing it at how these projections indeed were used to good effect would get you closer to the truth than what appears to be your assumption that nothing was done and the projections were false. So much has changed in the intervening 22 years.

Maybe I wasn't clear enough because I'm seeing a lot of the same comment here.

At some point anybody who's producing statistics and projections has to decide what's more important between accurately predicting the future and changing the future. Is being truthful more or less important than being revolutionary? It's fine and noble to want a better future but it introduces a bias into projections and how they're reported.

Carl Sagan was more than smart enough to know that society was likely to learn the dangers of smoking and to act accordingly. I'm guessing he had his public prediction which he put in his book that deaths due to smoking will rise to ten million but if you asked him over coffee when he wasn't driving home a socio-political rant about Big Tobacco he would probably produce a lower number more in line with what has actually happened.

If this sort of convenient number selection happens in a lecture about baloney detection we should probably expect it to happen in other places, too.

You are missing something really important though. Those kind of claims usually start with the phrase, "If X continues at the current rate..."

Just in the last few years we (in the U.S. at least) reached the lowest rates of smoking ever recorded, which surely accounts for the lower rate of deaths at least in part.

The CDC in the 90's couldn't have known how effective the aggressive campaign against smoking would end up being over the next 20 years. They had to make a prediction based on the rates of smoking they had at the time, and that's really all you can ever do.

Are you suggesting that we shouldn't try to predict anything because what if we're wrong?

I have found one of the best ways to be a skeptic is to be clear what you are skeptical about. Example: are you skeptical that the oceans will rise at all or just that they won't rise to X level? Skepticism is often interpreted as dismissal which shuts down dialogue. I like to think the point of skepticism is to further dialogue not shut it down. The last thing charlatans want is dialogue. They thrive on monologue. That's the skeptic's power, to me at least.

Keeping with apples to apples comparison, the WHO says that more than 8 million a year die from smoking, for a growth of 5 million per year over 22 years, which is not that far below 7 million. https://www.who.int/news-room/fact-sheets/detail/tobacco

There’s always an implicit (or explicit) “if conditions X, Y, and Z hold” on these statements. That prediction wasn’t wrong, conditions just changed, largely because predictions like this managed to wake people up to the dangers.

Ignoring the unstated assumptions in these statements isn’t really skepticism.

>While we're on the subject of skepticism and 'baloney detection' whenever somebody says X will happen in Y years there's usually a bit of baloney in there.

Agreed.

One of the most popular New Year’s Day posts here was a popular tech personality with his 2019 predications that definitely had Trump removed from office.

I look forward to bringing that up next New Years, not because I’m a Trump fan, but because if all these Ms Cleo fortune tellers are so smart I wonder why they waste time with blogging.

Was it the one where he also assigns a probability rate to his prediction? i.e. "Trump will be removed from the presidency - 90%", and then at the end of the year he calculates all his predictions (which are True/False by nature) and weighs them based on how certain he was to determine a marker of how accurate he is.

Yes! That’s the one I think. Thanks, that’ll help me find it next year for the thread that will invariably have the same predictions :)

https://slatestarcodex.com/2019/01/25/predictions-for-2019/

I think he actually predicted that Trump would remain the president, but I've always found the predictions post as strange. It seems like he's making a bunch of safe bets and scoring them safely, and then declaring himself accurate as a result. "What everything thinks is most likely is probably most likely, so you should listen to what I have to say" would be the takeaway, which isn't that interesting.