RE: The dangerous reverse conversion proposal, and how to REALLY fix the peg
After some further consideration, your proposal is hardly different from the witnesses' proposal once we reach the intended/expected end state of SBD=$1. Consider your conversion formula: sbd_received=steem_price*steem_converted*(1+SIQ)/sbd_price. Given that SBD=$1, SIQ=0 and therefore this reduces to sbd_received=steem_price*steem_converted which is exactly what witnesses are considering!
Since you wrote elsewhere that "Mallory's attack necessarily takes place once SBD has stabilised at $1", it would seem that your proposal adds nothing other than the rate-limited qualification which you have not fleshed out into an actual proposal. (FWIW, I have also considered rate-limiting, but as you note, the problem of how to establish the rate is quite difficult. Since the existing payout mechanism is a sort of STEEM-to-SBD conversion, we can say that this rate-limited conversion already exists, but the problem is simply that the rate is too low.)
One last comment. In a comment you mentioned this precondition for your Mallory attack scenario.
the majority of the external market to pay no attention to the conversion queue
First of all, I would point out the need to clarify what is meant by 'the majority' of a market. In any market there is a mix of informed and uninformed traders. If there is too much capital deployed by uninformed traders, then prices will simply not be rational, until a new equilibrium is reached where: a) more informed traders are attracted to the market by profit opportunities; or b) uninformed traders have lost enough money such that they no longer are capable of setting prices.
Second, I can say that from experience the premise of at least informed traders being unaware of the conversion queue is completely false. There have been numerous occasions of large conversions using the existing SBD-to-STEEM mechanism that entered the queue and they were promptly noticed and discussed by traders. There even a few posts on this platform about them which can be found, if one cares enough to go dig them up.
I do recognize that none of these mechanisms are perfect, but we must also not let the perfect be the enemy of the good. Overall most of us who have looked closely at this do not find STEEM-to-SBD conversions to by significantly more dangerous than the existing SBD-to-STEEM conversions (which also raise the possibility of platform-provided liquidity being exploited), and further we find that the existing mechanism is reasonably effective (though, again, not perfect) in limiting these risks.
Considering a variety of factors including risks as well as complexity of implementation, it seems
STEEM-to-SBD conversions is the best plausible method in the short- to intermediate-term of remedying the clear and massive manipulation associated with SBD being pushed very far from its intended and most useful price of $1. If and when we are able to identify and even better mechanism, the possibility exists for swapping that in.
Please understand that I am reading your proposal with the utmost respect for your perspective and adversarial approach to analysis. In fact, my view is that we as a community need more of this, not less. I hope you are able to continue to contribute to these discussions.
Our discussion here aside, I have just discovered that you are, in addition to a top Steem witness, also a Monero core team member. Thank you for your work on Monero. After I send this post, I will cast my vote for your witness. I am convinced through others vouching for you as well as your involvement in Monero that you will represent me well.
Now, you were saying:
I agree, and you have provided sufficient argument that I am at least partially convinced STEEM->SBD conversion will not be the absolute end of the world. I do trust the existing and dominant stakeholders to solve the problem in a manner they believe to be safe after all this discussion. However, I'd also like to address this point you made:
My proposal doesn't only deal with Mallory's attack; it also proposes a gradual, controlled drop in SBD's price, and a fix to the 50%/50% algorithm in the meantime. You seem to have focused in on the attack aspect of my proposal, but it was to be taken as a whole. The attack was only part of the problem with the proposed conversion. The reason for fixing the 50%/50% algorithm first is:
The rate-limiting part is to prevent Mallory's attack and it is the only mechanism for doing so. You are correct that the attack mitigation aspect of my proposed solution is quite gutted without it. The closest thing to an actual proposal I have here is to rate-limit the entire platform to something we don't think could significantly impact the market, and divide that capacity according to vesting stake. Just as they are rewarded in other ways, those with high stake in the platform are rewarded with a higher profit potential from the big SBD sell-down.
More on point 2: Many have said that people who would not be here without the high SBD should not be here. I mostly agree. I think that anyone who is selling the excess SBD they receive and exiting the system with that value is stealing from the platform (to the extent that "theft" as a concept has meaning in a decentralised society). At least 50% of the "real value" of each post should re-enter the economy as vesting shares and any noncompliance with that mandate is an exploit and should be patched.
However, I also think that the high SBD has brought growth to our platform in the short term and I think it's silly to crash the value of SBD all at once. By winding it down over time, I believe the chances would be greater that more of that value would stay in the Steem ecosystem.
Thank you for your kind words. I appreciate the discussion we have had over the past couple of days. I am glad you see the value in this type of thinking, and will certainly continue to involve myself in discussions concerning the platform's direction as I see fit. Thanks for hearing out my concerns.
I'm not sure this would work. It might result in a gradual drop. But it also might lead to an immediate price collapse as soon as it is implemented, as speculators extrapolate the progressive effect of your 'gradually' discounted conversion (of course, this would require that they are actually paying attention, a questionable assumption as we have both previously noted). The only way I see to somewhat ensure a controlled drop would be to actually reset the peg higher and then gradually reduce it. Or perhaps to introduce rate limiting with a very low rate at first, and then slowly increase the rate (though the latter idea still risks a rapid price collapse). Both of those are somewhat complicated, and implementation complexity is definitely a consideration especially for code that in all likelihood would only ever be used once.
The ideal case here would be for the existing mechanism and/or changes to market dynamics to bring the price down on its own at least closer to $1, and then strengthen the peg from there. I'm not sure if or when that will actually happen but I wouldn't rule it out. A week ago SBD was trading at $3, having come down from $14. A huge amount of SBD is being created now. So we will see how this works out for now.
I see at least three issues with this idea. 1) It is difficult to determine how much of a rate will significantly impact the market (and this likely changes over time, mandating an ongoing monitoring and adjustment task); 2) In order for any real influence on the peg to be effective, it would have to at least have some effect on the market (in effect, what we see now is precisely the effect of excessive rate limiting); and 3) clearly many if not most stakeholders would not participate, therefore this mechanism would either result in a tiny rate being applied in practice, or would need a significant overcommit to allow those stakeholders who do participate to have any meaningful effect. I'm not sure how to apply that overcommit in a manner that doesn't allow larger stakeholders to perform extremely large conversions, and also that does not become extremely complicated to implement and understand.
That is a reasonable way of putting it, and I mostly agree. Again, though, we are somewhat confronted with the reality that code to patch it would in all likelihood only be used once if the upper peg issue is addressed, and that (including not only development, but review, testing, and deployment, all of which have significant associated costs and time lags) has to be traded off against the severity of the exploit. There also isn't really widespread recognition of this as an 'exploit', in fact as you have seen quite a few people see it as a benefit!
Thank you for the witness vote. Be assured I take those duties seriously and do my very best for the community both operationally and in terms of policy.
Hi @smooth, though I would much prefer a strong SBD, I'd just like to say how much of a relief it is getting your thoughts on this first hand.
Prior to reading this, I was under the impression you were aggressively dovish, seeking an instantaneous crash to $1. The prospects of this had me very worried, as IMO there would be collateral damage in the form of reciprocal panic selling on STEEM, and perhaps even a permanent hysteresis effect stemming from burned investor confidence.
Completely agree, if a peg must be enforced, this would be ideal. The problem is, that decline to $3 was in the context of a market-wide decline, and both STEEM and SBD rallied harder than BTC once the bottom was in. Both coins are not only tracking the market, but outperforming it.
What I'm seeing at the moment, on both STEEMUSD and SBDUSD, are Inverse Head and Shoulders reversal patterns. As such, if left well alone, I think we'd get a return to the highs when/if Bitcoin resumes higher. In all honesty, I think the chances of an unaltered SBD declining solo from here at virtually 0.
Therefore, if SBD must be pegged, I think we need to be looking at controlled decline scenarios.
This would make a whole lot of sense, though sounds like implementation could be quite the process. I've heard you've made a strong case against this , but is not the envisioned and most straight forward solution the price bias parameter?
Update
Though the market charts still look bullish, after looking at these supply charts from @buggedout, I think this thing will almost definitely sort itself out as STEEM trends higher. Matter of time. Still very interested to hear your argument against a price bias controlled decline though.
I have no arguments against a controlled decline but as I wrote above I don't think it can be reliably implemented without explicitly pegging higher than $1 and gradually reducing that target. Anything else (or doing nothing at all) still risks a price crash even though it may not force one on a particular schedule.
Apologies, I'd heard you'd made an argument against using the witness Bias parameter specifically? I note only one top 20 witness is currently applying a bias.
The bias parameter increases the supply of SBD with no compensating reduction in the supply of STEEM. Since SBD is ultimately converted back into STEEM, this results in higher (potentially much higher) inflation than is supposed to happen (currently about 8.5% with a slow decline down to 1%). This is not something that many stakeholders, particularly the larger ones, support, and therefore it is unlikely to happen.
By contrast a conversion mechanism creates SBD by converting the corresponding quantity of STEEM into it, with no net effect on the overall inflation rate.
Sorry about the delayed response. Makes perfect sense. Thanks mate.
After talking to you and Luke Stokes, I am confident we are in safe hands and will be re-voting the majority of witnesses I un-voted due to seemingly unfounded rumours of reckless intervention.