Also, this is undesirable if the HIVE price was lower in the past, but has risen say in the past day or so, which will regularly happen due to typical price fluctuations. It will discourage creating new HBD even if HBD is overvalued and more is needed, since the minimum will take the lower price, not the current price, and the converter is penalized.
Now in this case, is is clearly no worse than simply not having the function at all (which is where we are now) so in that sense it can't be seen as an objection to the proposal.
But I see see no gain to using this proposed method over inverting the existing conversion mechanism, which has proven to be quite effective (apart from situations where the haircut is hit, which of course is a different issue).
Everything you've said here I generally agree with, but I think you wrote this prior to my last comment, so just to note as per my last comment, the reason I favored this over the 3.5 day delay method is ability to respond quicker to a price pump. But there is definitely the tradeoff that you mention in your first paragraph with the worst-case price in last hour term.
Yeah, I think rather than wondering about different tradeoffs, we should just start with the inverted 3.5 day mechanism. That addresses the fundamental supply imbalance issue. If it then turns out we need something faster to provide more stabilization against short term moves than what people doing short term trading and market making provide, then we can add something else.
Let's start with the proven mechanism and see if it is enough. It think it might be.
(Yes I wrote the above before your last comment.)