The merkle article oversimplifies the real findings of the reseach work : https://arxiv.org/pdf/1709.02489.pdf
In particular for zcash, only transparent adresses can be analysed, which is the expected behaviour since user need to activate the z adress to preserve anonymity.

Extract from the original article:
Zcash is also supported, at least to the extent that Zcash blockchain analysis is even possible:
it introduces a complex script that includes zero-knowledge proofs, but these aspects are parceled away in a special type of address that is not publicly legible by design.

I still prefer monero than zcash for the same reasons as many (massive premine, trusted setup...), but  we need to be honest when reporting these findings.
For dash, the conclusion of the merkle article is still valid.