Thanks for chiming in. I do recall coming across hashpool, and it may be appropriate and within the risk-tolerance for some users. However, due to the custodial nature of ecash, I have not investigated hashpool in relation to BMAX at all.
One nice thing about BMAX, is that because it is its own network, and because BMAX nodes are also Bitcoin nodes, then BMAX consensus can serve as a sort of “observer” of relevant Bitcoin on-chain metrics (feerates, difficulty, etc), and can settle contracts/bets about these things, all without needing an oracle or federation.
We may simply be squabbling over subtly different definitions, but I do think we agree on the general effect: it is not generally rational for someone to pay more than R sats now to accept an offer whereby they will receive R sats T time-units from now.
Rather, the amount someone might consider paying is simply the present value P of that offer, where P = R e^{r T} and where r is the buyer’s “discount rate.” The value of r implicitly accounts for whatever the actor cares about, including, but not limited to: counterparty default risk, inflation risk, and opportunity cost. For the latter, and at least for a subset of actors, it includes peeking at the current offerings of the traditional finance world, such as “what interest will the bank or a treasury bond give me?” and because that question is not denominated in sats, the buyer must then take a guess at translating that other unit into sats (at a future time), discounting accordingly, and so on.
We do not get to directly observe what exactly is the buyer’s calculus with regard to r, but if we are lucky and can observe the prices paid for many similar such claims, we can begin to infer some information about the collective time preferences.
When T is small (seconds to days), the inferred information about time preference is not very useful to anyone other than hedgers and speculators who want to act in such short time spans. However, when T is large (months, years, decades), the inferred information about collective time preference becomes much more relevant for economic actors who have longer planning horizons.
Miners themselves are good examples of these types of actors, but they are not the only ones. Any person or business which seeks to embark on a long-term capitally-intense journey should ask them selves: “what if we do not do this, and just buy bitcoin instead?”
Right now, trying to answer that question is, in my opinion, too reliant on the traditional fiat world (and even then, the horizons in the fiat futures markets for bitcoin are not long), and we can do better.
I think you already know all of that, but I wanted to try to lay it out for others who may come across this thread in the future.
I suspsect there is demand for something like this too. While miners are the obvious initial “customers” / users of such a thing, I see no reason why it would not very quickly expand beyond that segment, and that is a good thing.
The BMAX design needs to be robust to aptly raised concerns like this. Thank you @ajtowns for providing such helpful comments and ideas! Some of what you elucidated I had thought about, but had for some (clearly wrong!) reason dismissed at some point. Nevertheless, I am working to update the design so that these concerns may hopefully be alleviated while still preserving the desired underlying properties of incentive compatibility, oracle-free, non-custodial.