Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

All models are wrong, but some are completely wrong

rssdss.design.blog

Post reply on HN
Print view View original

161–170 of 216 posts

Re: All models are wrong, but some are completely wrong

#161

Post by strenholme »

justanotherc wrote: post #150
strenholme wrote: post #136

I am working on trying to make an accurate model for predicting COVID-19 growth based on the per-county figures we have with COVID-19 (courtesy The New York Times). To say the data is noisy would be an understatement. What I have found, so far, is that if we look at current daily growth (averaged over seven days) and use exponentiation to predict future growth based on the previous week’s figures, the numbers are too…

It could only be exponential if the population was infinite. It has to be more of a bell curve, because as infection grows, there are fewer hosts to infect, and therefore growth would start to decline.

Even in places where 0.05% of the population has confirmed cases of the virus, I’m seeing the curve flatten. Whether that’s from quarantine of from the virus hitting its limit, I can not say. But the curve is flattening.

Re: All models are wrong, but some are completely wrong

#162

Post by deltron3030 »

kerkeslager wrote: post #109
sah2ed wrote: post #90

Earlier quoted context omitted.

> If there is one less to be learned from this whole Covid-19 debacle (and I'm sure there are several), it is that our entire news ecosystem, public and private, is fundamentally structured wrong for doing what is supposed to be its purpose, which is to make people better informed. It's not bad at it by mistake, it's bad at it as an inevitable consequence of its design. You speak of a failure in design as the root ca…

If you're smart enough to come up with a point like this, why not use that intelligence to make a less inane point? You're not wrong, you're just correct in a way that responds to a very narrow reading of what your parent comment was saying, and doesn't make any attempt to figure out why they'd be saying it. Clearly the person you're responding to wants a system that makes people better informed, and I think there's…

>So the question is obviously, how do we change the news ecosystem from what it is, to a system that makes people better informed?

By bettering basic education. People could learn the relationships and interdependecies of the system and society itself, including the role of information. Why not learn everything through RPGs?

This would create a demand for better news sources in the long term.

Re: All models are wrong, but some are completely wrong

#163

Post by AnthonyMouse »

amitdeshwar wrote: post #119
renewiltord wrote: post #74

Earlier quoted context omitted.

The problem is on the demand side, not solely on the supply side. People desire reinforcement not challenge. That is natural. I suspect it is true of us all. Those who are capable direct their potential to be challenged in specific ways, allowing themselves to have biases reinforced in the unimportant places. And they have access to the knowledge in their specific areas. This is the reason for the enduring power of t…

What about 538?

Does 538 have broad-based popular support?

How does its viewership compare to that of gossip-based media like E! or anger-based media like Fox or CNN?

Re: All models are wrong, but some are completely wrong

#164

Post by antpls »

cryptonector wrote: post #155
6510 wrote: post #154

I think making medical claims is quite punishable if one doesn't have a license. This could easily be extended to include journalists and politicians as well as other areas of science (if the stakes are high enough) It would work like this: For each field there has to be a committee similar to the ones medical practitioners already have. If the behavior is not up to standards they do an inquiry and if need be the cre…

> I think making medical claims is quite punishable if one doesn't have a license. Well, it's tricky. If you say "I don't believe covid-19 is a serious problem", that can't be punishable as there is a large spectrum of legitimate thinking as to its severity and what trade-offs we should want to make. Such a statement is not remotely like practicing medicine without a license, but some will argue that it is if it coul…

> While saying that "chloroquine phosphate is a good prophylactic for covid-19" to someone who would believe it certainly should be punishable (as attempted murder perhaps! and regardless of whether the speaker is a licensed physician!).

I would disagree. As an example in the history, saying "The Earth is a sphere" was once punishable by death. Fake news and pseudo science has always existed, but censorship is not a solution. Only research, education and communication can help, and even then, each generation will always have his own set of questions without answers.

I don't agree with the parent' solution either, for the same reasons.

Re: All models are wrong, but some are completely wrong

#165

Post by cryptonector »

antpls wrote: post #164
cryptonector wrote: post #155

Earlier quoted context omitted.

> I think making medical claims is quite punishable if one doesn't have a license. Well, it's tricky. If you say "I don't believe covid-19 is a serious problem", that can't be punishable as there is a large spectrum of legitimate thinking as to its severity and what trade-offs we should want to make. Such a statement is not remotely like practicing medicine without a license, but some will argue that it is if it coul…

> While saying that "chloroquine phosphate is a good prophylactic for covid-19" to someone who would believe it certainly should be punishable (as attempted murder perhaps! and regardless of whether the speaker is a licensed physician!). I would disagree. As an example in the history, saying "The Earth is a sphere" was once punishable by death. Fake news and pseudo science has always existed, but censorship is not a…

Chloroquine phosphate is fishtank cleaner, and -to humans- poison.

Granted, if you're just dumb and ignorant and choose to take it yourself, that's no crime. But telling others to take poison, without telling them it's poison, should be a crime. Of course, in the case of chloroquine phosphate, if you look at the packaging, you'll see it's poison.

Re: All models are wrong, but some are completely wrong

#166

Post by Proziam »

deltron3030 wrote: post #162
kerkeslager wrote: post #109

Earlier quoted context omitted.

If you're smart enough to come up with a point like this, why not use that intelligence to make a less inane point? You're not wrong, you're just correct in a way that responds to a very narrow reading of what your parent comment was saying, and doesn't make any attempt to figure out why they'd be saying it. Clearly the person you're responding to wants a system that makes people better informed, and I think there's…

>So the question is obviously, how do we change the news ecosystem from what it is, to a system that makes people better informed? By bettering basic education. People could learn the relationships and interdependecies of the system and society itself, including the role of information. Why not learn everything through RPGs? This would create a demand for better news sources in the long term.

The most educated people I know all have at least a handful of opinions that are provably untrue. We're talking about people who correctly modeled economic crashes, developed and implemented SAP modules for a lot of companies you know, and (multiple on some of them) PHDs. Among them, a flat-earther, a guy who believes communism is straight-up awesome, and one who honestly believes that most "corruption" is a made-up attack on the corrupt.

Each time I encounter this, I am stunned. But it still happens.

Re: All models are wrong, but some are completely wrong

#167

Post by sah2ed »

perl4ever wrote: post #149
sah2ed wrote: post #114

Earlier quoted context omitted.

> I and perhaps many other people think the media should inform people; that is a statement of its purpose, and it appears that the collective actions of people in society have redesigned it to eliminate most of the informing. Your expectation that the media should inform people is reasonable but unrealistic. The original designers had fairly specific design goals in mind and it did not explicitly include “to inform”…

"Your expectation" I wrote "should"; I am not predicting in my comment that that the media will inform people [more] in the future; therefore calling my "expectation" unrealistic is uncalled for. Referring to "the original designers" is confusing to me. You seem to be implying that "the media" was designed by some person or entity as a monolith, which is a theory that is unfamiliar to me and not mainstream (that I'm…

Let us leave aside my use of designers for a moment and turn to the actual businesses that constitute the media.

On the one hand, media businesses are, in most cases, for-profit enterprises similar to other enterprises.

On the other hand, “to inform” or to have better informed citizens is in the public interest. “To inform” essentially an altruistic goal. Something done for the greater good is hard to monetize unless you ask for donations. Altruism as a business model will not pay the salaries of journalists.

My argument is that the two goals: “for profit” and “to inform” are incompatible goals — either you found a non-profit media entity with the altruistic goal of better informing people, or you found a for-profit without the pretension of altruism.

By arguing that “I and perhaps many other people think the media should inform people”, that the media’s collective purpose is to better inform, essentially what you are really saying is that you want for-profit enterprises to behave in the manner expected of non-profit enterprises? As I said, this is a noble but unrealistic ask. This is the crux of our disagreement.

Re: All models are wrong, but some are completely wrong

#168

Post by bad_user »

com2kid wrote: post #158
kick wrote: post #141

Earlier quoted context omitted.

https://en.wikipedia.org/wiki/List_of_fallacies#Red_herring_... Fallacy of relative privation (also known as "appeal to worse problems" or "not as bad as") – dismissing an argument or complaint due to what are perceived to be more important problems. First World problems are a subset of this fallacy. Nestle is evil. Microsoft is evil. Nestle kills more than Microsoft. Microsoft is still evil.

My bar for evil is a lot higher than "signed exclusive distribution agreements with some OEMs in the 90s." Illegal? Sure. Evil? So some other millionaires didn't get to become billionaires. Meanwhile the world got a single API to write software against for almost two decades, how many ISVs existed because Microsoft provided an insanely stable (in regards to API churn) platform to develop against? Hell right now with…

While I agree with the general feelings about the evil threshold, Microsoft in the 90-ties built and took advantage of a natural monopoly in order to crush their competition and expand in new markets.

Let's not go to the other extreme and downplay what they did. If they had their way, the Internet wouldn't be an open platform and open source would be illegal.

Personally I don't care about exclusive OEM deals. What I care about were the dirty campaigns against open source and them trying to leverage their huge patents portfolio against Linux and later Android. Or them planting shills inside Nokia, weakening it enough for a takeover of their mobile division and then running it into the ground, thus destroying one of Europe's top tech companies. A conspiracy theorist would say that this was economic warfare, maybe sponsored by the US, but for me incompetence is enough for blame.

Or how about their long battle against open standards, like ODF and their use by public institutions?

And the Microsoft of today, in spite of popular belief, isn't very different. Their priorities may have shifted, but even post Balmer they continued to use their patents portfolio against Android and they continued to fight against the adoption of ODF. Their predatory culture is still there and as seen by the ads and aggressive telemetry in Windows 10, as soon as they can gain some economic advantage, they'll take it, regardless of cost to society.

Their "Microsoft changed" marketing campaign has been genius in its execution. On the other hand I'm glad that they are doing well, because otherwise they have the potential to become the biggest patent troll. Just like when their Windows Phone failed to gain traction, "if you can't innovate, litigate" seems to work well.

Re: All models are wrong, but some are completely wrong

#169

Post by noobermin »

nitrogen wrote: post #88

Would we encourage an epidemiologist to apply ‘fresh thinking’ to the design of an electrical substation? Yes, absolutely. If an epidemiologist identifies and models a trend in human disease around substations, or a trend in failures of substations, or a new way of modeling the ways electrical demand can change over time and influence demand in other times and places, etc., then their input should absolutely be consi…

The easy medium is talking to other scientists, may be getting them on as a co-author.

Re: All models are wrong, but some are completely wrong

#170

Post by noobermin »

YZF wrote: post #14
martingoodson wrote: post #3

Author here: happy to take comments or criticism

I'm finding myself in disagreement with rule #6. Using a model effectively is about a lot more than just the domain knowledge. I'd value analysis from a mathematician/statistician more highly than from an infectious disease physician. There's the stuff that informs models, i.e. the observations, the experimentation etc. and then there's the science of modelling itself which isn't really in the same domain.

Domain knowledge is very, very important in modeling because it helps you figure out what to put in the models. Math is important but certain odds and ends that are of great interest to mathematicians are not helpful for the usual simulation, and the simulations that are most helpful in informing policy.
Post reply on HN