Let battle commence !

I predict a scientific cat fight on arXiv in the next few days.

A few weeks ago there was a claim that the rotation curve of galaxies scales very precisely with the density of the observable matter. This isn't a new result (see point 2 here) but the new paper apparently confirms it with much greater precision, so it can't be dismissed as an observational error. As an observer, I'm skeptical of that because errors are always larger than people claim.

But let's assume it's true. If so, it's no doubt very interesting. There's no obvious reason why the density of the observable matter should correlate so well with the rotation curve. Density of normal matter should make very little difference to how fast it's rotating around the centre of the galaxy since that should be dominated by the much greater mass of the dark matter. And it's not obvious why the dark matter should have strong enough density variations to cause wiggles in the rotation curve anyway - especially drops in the rotation speed, which are really hard to explain if the dark matter is the smooth distribution of particles it's supposed to be.

So could the whole dark matter edifice just be wrong ? Sure, that's always an option - but it's not a very likely one. The reason I don't often comment on major press releases is because they're invariably overblown, and this case was a particularly bad example ("...argue that the relation they've found is tantamount to a new natural law." - oh please, _seriously ?). And a mere month later, this paper challenges those claims very strongly.

The problem is that even without knowing the details, galaxy formation is messy. Gas can be heated by collisions or from stars and supernovae. Its cooling depends on its density and chemical composition, which depends on the star formation activity. It's got complicated magnetic fields doing all kinds of jiggery-pokery. And if there's dark matter, it's also being strongly affected by a much more massive component. So the gas and stars are subject to (potentially) very strong selection effects - what we're seeing is just the tip of the iceberg, biasing our view of the Universe.

The problem has been, for many years, that it's computationally very difficult to simulate both the complicated physics of normal matter and the apparently much simpler behaviour of the dark matter, on a large scale. That's starting to change as computational power improves, although for sure the physics used is still limited. This new paper uses these improved simulations to show that the "new law" can actually be very well explained entirely using conventional physics. In fact it appears that this can be explained using only the collapse and cooling of the gas - nothing else is needed.

Is the debate over ? Hell no. I'd be amazed if there wasn't a response from the original authors very soon. The new paper predicts there should be a strong dependency of this relation on the distance of a galaxy (since galaxies earlier in the Universe were forming stars at a faster rate). The original authors didn't comment on that. The new paper also doesn't really comment on the small-scale wiggles in individual galaxies, which modified gravity theories have no problems with (by design rather than due to predictive power). But their main result is very, very close to the earlier observation.

So CDM survives yet another challenge, and I maintain my 80:20 balance in favour of dark matter over some other explanation.
https://arxiv.org/abs/1610.06183