You didn't answer these questions:
1 - How is DPoS an "improved" version of PoS seeing that it is clearly a more centralized consensus mechanism?
2 - How does Bitshares rationalize that it is "decentralized" when it is susceptible to and
previously undergone a Sybil attack, where actually, Bytemaster voted ONE person into FIVE delegate positions?
Someone having multiple delegate positions is not an attack. It increases centralisation so is best avoided, but is not an attack. Would you consider the moment gigahash.io briefly had over 51% of the hashpower on bitcoin an attack even though they didn't actually attack? Having 5 delegates is no where near enough to even potentially attack the network so please refrain from making wild exaggerations.
Hmm, I don't know. I'm pretty unbiased here and somewhat interested in Bitshares in general, but reading what happened there looks like a pretty clear case of a Sybil attack. Having one guy secretly holding five delegate positions until he was outed by his former group sounds like it fits that definition to me. That's assuming the intention of DPoS is one delegate = one person(or more). Don't have to attack the network itself to succeed in gaining more delegates through a Sybil attack.
I think theoretically you could get a pretty reasonable amount of confidence in 101 DPoS delegates if each delegate was forced to provide significant proof about their real identity including live streaming. Tax documents, IDs and such too. I'm someone who supports the right to stay anonymous in general and I don't put a lot of stake in people providing their real name and such. But some really invasive protocol to prevent Sybil attacks seems like it could plausibly work.