The more I dig into the history of dark matter as a theory, the more irritated I get.

The Bullet Cluster is seen as evidence for collisionless dark matter. Great.

Meanwhile, Abell 3827 is seen as evidence for self-interacting dark matter. Fantastic.

Both are put forward as evidence for dark matter – even though, taken together, they show that dark matter can't be the solution to the relevant observational problems, because these two "answers" point in completely different directions.

It's collisionless when we need it to be!

But it collides when we need it to!

It's a fish!

It's a horse!

It's a croissant!

Then you've got Abell 520, where the main mass peak coincides with the gas rather than the galaxies – the opposite of the situation in the Bullet Cluster.

And the ultra-diffuse galaxies DF2 and DF4 in the NGC 1052 group, with very low velocity dispersions that indicate little or no dark matter. A fact which gets used, somehow, with logic that would embarrass a toddler, as evidence that all other galaxies MUST have dark matter!

Completely selective interpretation. Never a synthesis of all the observations, which would reveal the total incoherence of the theory.

Yes, we have real observational phenomena which need to be explained.

No, the current approach does not explain them.

Dark matter my hole.