Sticking to TFA, and not ancillary information, I found this frustrating:
> Instead, the research team ran 75 different models with the initial conditions chosen at random. By chance, some of these created distortions similar to the ones seen in the real-world data, typically affecting only one of the four lensed images. So, the researchers conclude that the distortions in the lensed images are consistent with a dark matter halo structured by the quantum interference of axions.
Essentially the best contender aside from WIMPs is, as subject of TFA, axions. And axions have near-random interference and standing waves, which essentially scatters about light bent by gravitational lensing due to uneven gravity in the lensing object.
Essentially, if you randomize the field enough, you can eventually find, after enough trials, something that seems to match reality. And it just so happens axions are the nice random variables we needed to inject.
That's where we're at: Just randomly select some parameters until the data fits, and lo and behold, it fits.
They claim the parameters that go with canonical particle dark matter won't lead to a fit. This is the point, not necessarily that they could pinpoint axions in particular.
So they searched for a set of parameters for axions that would essentially be a counter example to WIMPs? That makes more sense. But the article is structured as this being supporting evidence for axions, which doesn't make much sense to me.
At least the headline says "No WIMPS! Heavy particles don’t explain gravitational lensing oddities" and also inside the article (near the two colorful figures) they write
"On cosmological scales, WIMPs continue to fit the data extremely well. But once you get down to the levels of individual galaxies, there are some oddities that don't work quite as well unless the dark matter halo surrounding a galaxy has a complicated structure. Similar things seem to be true when you try to map the dark matter of individual galaxies based on its ability to create a gravitational lens that warps space so that it magnifies and distorts background objects."
and later
"It's a relatively simple thing to do with the WIMPs, since there's only one pattern we'd expect: the gradual fall off of dark matter levels as you move away from the galactic core. The lensing predictions based on that distribution do a poor job of matching the real-world data of where the lensed images show up."
So I guess the real article (which I haven't read) has the hypothesis that the WIMPs fit badly and they try to make a case for the axions. Still it is kind of a nice clue to show that the WIMP predictions fit badly even if their axion solution is weak..
>That's where we're at: Just randomly select some parameters until the data fits, and lo and behold, it fits.
When I read that it felt like a Monte-Carlo simulation. A legit approach in modeling.
And indeed - the link to the paper: https://arxiv.org/abs/2304.09895
This seems fine? Planks law was originally just a line fit with parameter tuning. It wasn't till afterwards that theoretical backing was developed with statistical mechanics.
Knowing what types of models and configurations match experiments is useful even if these models themselves aren't derived from first principles.
> Instead, the research team ran 75 different models with the initial conditions chosen at random. By chance, some of these created distortions similar to the ones seen in the real-world data, typically affecting only one of the four lensed images. So, the researchers conclude that the distortions in the lensed images are consistent with a dark matter halo structured by the quantum interference of axions.
Essentially the best contender aside from WIMPs is, as subject of TFA, axions. And axions have near-random interference and standing waves, which essentially scatters about light bent by gravitational lensing due to uneven gravity in the lensing object.
Essentially, if you randomize the field enough, you can eventually find, after enough trials, something that seems to match reality. And it just so happens axions are the nice random variables we needed to inject.
That's where we're at: Just randomly select some parameters until the data fits, and lo and behold, it fits.