Did I miss it, or do the authors not bother to explain how the original methodology is wrong? Do we just take their word for it?

▲

amluto 2 hours ago | parent | next [-]

Not that I can see. They claim to have reverse engineered the methodology, but they have not obviously published what they came up with. They also have no explanation as to why they can find positive correlations with all kinds of locations but no negative correlations with anything.

▲

barney54 4 hours ago | parent | prev | next [-]

Do you think Costco really is associated with 2.2 million deaths?

▲

hatthew 3 hours ago | parent | next [-]

No, but how do I know that the authors used the same methodology? The most they say about it is "we developed our own methodology by guessing what they did, and we changed stuff around until the results from our methodology matched their results", which is very unconvincing to me.

▲

IMTDb 3 hours ago | parent [-]

The fact that they had to guess what the original author did is already an issue in itself. What we see here is that: 1. the original author did not sufficiently document their methodology and 2. when a best effort at reproducing the results is applied and we look deeper, we then discover that the results are wacky.

▲

kbenson 3 hours ago | parent [-]

When someone comes along and proclaims 2+2=5, just because someone is right that they're incorrect doesn't mean their own assertion that 2+2=3 is any more correct, or that their reasoning for believing the original person was wrong has any validity.

Being accidentally correct is only barely better than being wrong, and for many uses (such as when the reasoning is extrapolated and used elsewhere) is no better at all.

	▲	cwillu 2 hours ago \| parent [-]
		They're not claiming 2+2=3, they're showing that the same methodology that proclaimed 2+2=5 also proclaims (as best as they can determine give the flawed documentation of the methodology) that 3+3=7, 1+1=3 and 17+17=137

▲

shoobiedoo 3 hours ago | parent | prev [-]

Does choking to death on their hotdogs count? It's on my todo list

▲

neolefty 3 hours ago | parent | prev [-]

They didn't! They mentioned receiving "eight lines of code" and, at the end of the article, write:

> It is especially telling that no matter what landmark we applied to the methodology, we have yet to get a negative result. There is, in fact, a real chance that you simply cannot get a negative result from this method.

My guess: The code is not shareable. Maybe its method is laughably disprovable?