A couple questions that I could probably look up myself but since you're here and I'm lazy:
If a user converts SBD to STEEM while the debt ratio is over 10%, then they will get back less than their ~$1 USD worth of STEEM.
What happens to conversions already in process when the haircut happens? Do they convert at the rate of the enacting time, at the rate of the fulfillment time, some way of averaging it over the 3.5 days?
What is interesting is that when liquid rewards are paid in SBD (instead of STEEM), those SBD tokens essentially have the same economic benefit as a user buying STEEM and not selling it, i.e. HODLing.
Is the rewards pool based on the actual Steem supply or the Steem-equivalent supply of Steem and SBD? (What Steemworld calls "virtual supply" and Steemd has as the larger number in "current supply.") If it's the latter what effect does SBD holding have on the pool and the overall economics especially if SBD remains near $1 while Steem continues to fall?
It will be based on the median across the 3.5 days.
It is based on the virtual supply.
It's not possible to give an answer that is correct under all cases because the difference between what the virtual supply would be depends on what happens to the STEEM price. (The short answer is that it depends.)
If you look at the leverage section of the post, there are scenarios where more users holding SBD could cause the rewards pool to increase, and there are other situations where it could cause the rewards pool to decrease.