Assuming you are talking about Poloniex, each new deposit of a customer is a new XMR address. And if I am not mistaken there is change address for each succession of withdrawals from users meaning the coins are never always sent from the same XMR address.
Did I miss something here concerning your thoughts?
I was under the impression that the exchange address for plx was the same for everyone, and it was a different txid field that they gave to each person. I'm going off of memory here from when I made a second account there, and I remember the address being the same as on my other account, and only the txid field changed. I could be wrong, but I'll go make some more accounts there now to look.
Maybe, in a sense, when withdrawing from plx to you the address is different becase of stealth addressing, but would having the same private key for that stealth address still have an effect on benefit in the addressing/rs's if the tx came out of that address for one tx, and was sent then to one more address (say someone bought something), and then the money goes right back to the exchange address?
I was just wondering if there was a way to account for that (statistically? some other way?) when looking at the blockchain, and if there were any large resulting effects that could be mediated with having an actual different address rather than just a different txid?
So, if I were poloniex, what mechanism prevents me from looking at the blockchain composed of mostly transactions with me and saying 'that's my mixin' or 'that's not my mixin', at some point in the future?