Hacker News Reader

Comment by tremon | Hacker News Reader

These don't appear to be general-public myths, but misapprehensions from fellow scientists working in the field. Most of them are too specific ("Firmicutes:Bacteroidetes ratio", really?) to be prevalent among non-scientists.

Also, this one is curious:

- There are 10^12 bacterial cells per gram of human faeces

They first say "it is incorrect", then continue with "the real figure is typically between 10^10 and 10^11" -- that's just one order of magnitude off, on a 12-digit number. The pedanticism is strong on that one. It only makes sense if they're addressing their fellow colleagues, who are supposed to know better.

View on HN

> It only makes sense if they're addressing their fellow colleagues, who are supposed to know better.

Well, this is an article written by a medical journal asking research paper authors to stop making these common errors.

And the first reference is from a professor named Dr. Hacker, which makes it more relevant here regardless.

Here’s what may be a more charitable interpretation:

A prevalent layman’s myth is that the bacteria in our gut outnumber our human cells by a wide margin (the article quotes the myth as 10:1). That’s a striking figure to a layman and “only” an order of magnitude off from what they claim is the true 1:1 ratio.

I can't say I've ever considered the ratio of gut bacteria to human cells in my life. I find it difficult to imagine this consideration is itself very prevalent outside of people who are particularly neurotic about gut bacteria.

I’m certainly a layman on this topic and have heard this quoted often in lay forums like podcasts, and have even said similar myself.

I think it's unfair to characterize this as a layman's myth given that it was a widely quoted figure in the literature at least until 2016.

https://www.cell.com/cell/pdf/S0092-8674(16)00053-2.pdf (Are We Really Vastly Outnumbered? Revisiting the Ratio of Bacterial to Host Cells in Humans)

I guess I could have made it clearer with better wording. I didn’t mean layman’s myth as in “originated by layman” or “only quoted by laymen” but rather “commonly known to layman”.

The intent was to distinguish it from other myths in the article that are probably only known to experts. E.g., I don’t think you’ll find many layman quoting a myth about firmicutes:Bacteroidetes ratios.

> These don't appear to be general-public myths, but misapprehensions from fellow scientists working in the field.

I am a biochemist and read a few relevant journals regularly, and that’s precisely what I’m talking about.

> just one order of magnitude off, on a 12-digit number

Doesn't matter how many digits are involved; a factor of 10 is pretty significant.

Depends on the scaling factors. 10x the sound energy is only twice as loud.

Only sounds twice as loud to a human observer.

I don't know. The difference between 10^101 and 10^102 seems insignificant.

Depends on context. The difference between 10^34 and 10^35 is the difference between a baby and a full-grown adult in planck sizes. For a more relevant example, a terrabyte is an unit we regularly work with that is ~10^18 bytes and an order of magnitude difference is still quite impactful (this is fewer orders of magnitude than your example, to be fair, but more than the original correction).

> I don't know. The difference between 10^101 and 10^102 seems insignificant.

Pour yourself two cups of coffee.

In one, add 10g of sugar.

In the other add 100g of sugar.

Do you think both are the same?

Compare the ratios of the exponents in your example and in mine.

I see another consideration, an implication of the other response - the domain of the exponents. For example, consider 10^101 and 10^102:

* The ratio of exponents is 102:101; not much apparent difference.

* But that doesn't specify a domain, so maybe we implicitly assume the domain is (0, 1, 2, ... >= 102). However, if the domain is (100, 101, 102) then the difference between 10^101 and 10^102 seems more likely to be as significant as the difference between 10^1 and 10^2.

> * The ratio of exponents is 102:101; not much apparent difference.

I'm not sure if you understand why this is nonsense. A factor of 10 is not small by any scale, and the relationship between 1 ton and 100kg does not suddenly become insignificant if you decide to express weight in nanograms.

Maybe instead of calling each other's ideas nonsense, we could try to learn something. Maybe next time!

Imagine you received a wage that was 1/10 of what it should have been, you notified the appropriate department in the company, and they wondered why is your pedanticism so strong when it is just one order of magnitude.

If my official wage was 10^12 dollars/euros/pounds sterling, I don't think I'd be worried, really. Even yuan or yen.

It's the difference between rich and poor in Venezuelan Bolivares (290k USD vs 29k)

29K USD is still more than I make now.

In biology things mostly only get interesting on a logarithmic scale. It has to do with the fact that on/off is decided by concentrations.

So I was a tad surprised for example that I had my drinking water tested for lead. It came back “clean” at 1.9 ug/L, a friend’s came back as “be careful, flush in the morning at 2.7. That does not make sense to me. Although of course a wage difference of 1.6x is pretty significant.

Imagine learning that the estimated age of the universe was off by 1 order of magnitude. Astronomers will freak out; 99.9% of the world will shrug and get on with their lives, since the number is unimaginably huge and irrelevant to their lives.

Scientists should be 100% pedantic. It's baked into the discipline.

> that's just one order of magnitude off

"Just one order of magnitude off"?

Just?

Do you understand that this means the real world value is between 1%-10% of the initial claim?

The one that follows shows what they call "incorrect" is just 2X larger.

When I read the article I determined it was basically some folks on a hobby horse trying to promote their view and saying other people are wrong.

I don't have a problem with debunking absolutely made-up (no original source) numbers, but if you're sying people are wrong with a factor of 2 or 10 difference, you're just being overly pedantic.

it's impossible to know where and how an assumption will be built upon common intuition in existing science.

What if the estimate for an LD50 for a new drug going into a trial were an order of magnitude out due to one of the underlying inputs being an order of magnitude out. The worst case scenario is that things could die off that wrong number. If nor humans, at least a lot of rats...

Reply on news.ycombinator.com