Hacker News Reader

Comment by nonbel | Hacker News Reader

>"Based on these data the team discovered that NGC 1052-DF2 larger than the Milky Way, but contains about 250 times fewer stars, leading it to be classified as an ultra diffuse galaxy."

Interesting, the mass discrepancy (ie, "dark matter") has been noted to be proportional to the predicted acceleration due to the visible mass (according to Newtonian mechanics): https://arxiv.org/abs/astro-ph/0403610

How does this galaxy fit into that relationship? If the galaxy is ultra diffuse it sounds like very little deviation would be predicted, but I am no expert here.

Edit:

To clarify since the relationship is somewhat tortured: According to my understanding this observation sounds surprising in light of dark matter, but 100% consistent with what MOND proponents have been saying. I'd love to hear from someone who knows more.

View on HN

Honestly, MOND proponents seem to be saying everything and its inverse.

What is not really a complaint, because they should be looking everywhere. But yes, that will invalidate some MOND theories, while keeping other ones unharmed, and will be evidence of some other set of them. And I don't think anybody should be surprised by any of that.

(But I'm no expert either, so if one wants to correct me, it will be welcome.)

I could be a MOND proponent only in that "dark matter" looks exactly like epicycles to me. I make no claims to understand the details of MOND, so please do not blame them for any of my errors.

That said, what do you mean by "MOND proponents seem to be saying everything and its inverse"?

There are many MOND theories, and you will find some that fit almost any observation you can imagine.

The paper considers the class of MOND theories that made a precise prediction of deltaV = 20 km/s for this galaxy. I guess I am referring to only those.

I see the paper says:

>"[MOND predicts] the expected velocity dispersion of NGC1052–DF2 is ... 20 km s−1, a factor of two higher than the 90% upper limit on the observed dispersion." http://imgsrc.hubblesite.org/hvi/uploads/science_paper/file_...

So they chose to use a 90% CI along with assuming the error is normally distributed, etc. Looking at figure 3b it seems that if they instead arbitrarily chose to use a 95% CI that 20 km/s would be within the interval.

Ie, these results appear to be moderately unlikely if MOND were correct but extremely unlikely if "dark matter" is correct. Thus, according to Bayes' rule, the probability that MOND is correct has increased.

Not sure where the downvotes are coming from. The reasoning that this evidence is moderately unlikely given that MOND is true can be found in the parent post. The authors worked it out to ~1/10 to 1/20.

They don't put a number on how unlikely it would be given dark matter is true, but it sounds pretty extreme:

>Merritt remarks: "There is no theory that predicts these types of galaxies — how you actually go about forming one of these things is completely unknown." http://www.spacetelescope.org/news/heic1806/

>"And NGC 1052-DF2 appears to have none. Or at best, about as much dark matter as normal matter. And that just doesn't make sense." http://www.syfy.com/syfywire/what-is-this-galaxy-doing-witho...

The denominator of Bayes rule for Pr(MOND|data) is just the sum of all Pr(Theory) x Pr(data|Theory) where the main theory is "dark matter" and second is MOND (everything else is some small value epsilon), while the numerator is Pr(MOND) x Pr(data|MOND).

Where is the mistake?

Reply on news.ycombinator.com