The APR argument is a misrepresentation.

The 20% APR contributing to the failure to reestablish the peg isn't about the new HBD being created, it's about existing HBD not being used to help in the re-peg. The HBD is locked away in savings because that is what is incentiviced. The 3.5 days conversion becomes irrelevant if the 3 days saving withdrawel (together with the APR itself) creats a real incentive barier to participation in re-pegging efforts.

There are many other reasons why the 20% APR remains a bad idea, but centralication of algoritmic pegging is one of the bigger ones. Looking at the days around the depeg, only two accounts are responsible for more than 96% of the convert operations.

I did a post on this subject yesterday.

The 20% APR didn't cause the depeg, but it contributes significantly to its length. For a part because as you describe centralized pegging is fragile, and the 20% APR desincentifies decentralized re-pegging in a major way.

If we realy want to create a situation where we keep the insane 20% APR while adressing the centralication concerns, saving withdrawel should become dynamic, maybe dynamicly going down based on the level of the downward depeg.If for example between a 2% and 5% depeg, the saving withdrawel drops from 3 days down to zero something like:

if D <= -0.05:
   Tw = 0.0
elif D <= -0.02 :
   Tw = 100*(0.05 + D)
else:
   Tw = 3.0

Something similar could be considered for the conversion time, but that is a much more tricky one that could incentify manipulation.

The important thing though is that this would be another hack to make up for the fact that the 20% APR incentifies $HBD away from comunity participation in re-pegging. Centralized re-pegging is fragile. Decentralized pegging has halved since the APR went up to 20%, and it was already low at that point.