DR - Distributed reserves

This form of collateral can’t be used to support the “last 15%” of tier 4, so won’t be an ultimate solution. But it can be used to reduce risks, in the short to long term, for the residue amount above 15%. Though I still prefer a higher collateral like 5%-10%.

Higher collateral would come at a higher cost in the form of contractors requiring more interest from Nu though. With the current situation I would really like say a 10% reserve similar to Nusafe that is in USD and value stable an additional 5% reserve in BTC (since Nusafe similar reserves come with an monthly interest rate). And an additional 10% reserve in DR, this last reserve would be “cheap” in terms of interest and gives Nu a net profit when contractors default and is still backed by the 15% reserve in Nusafe/BTC. This would give us a total 25% reserve relative to outstanding Nubits.

I think for the current Nu situation this would be the best model.

@dysconnect comment is exactly what I am talking about with “potential downsides that in the end are not”. We have already discussed why that is not the case, yet the “potential downside” is brought again with conflicting premises:

When that contractor in this scenario:

“reinvest” the money, it transitions to this other one:

so the issue we are talking about:

never even exists in the first place, cause at that point we don’t have to support anything with the 15%.

This mathematical model I am talking about would help avoiding this, the 60+ redundant posts in NuSafe discussion, among others like DR, revenue allocation, T4 funds management

The problem is that we are expecting to have that 50k in Tier 4 reserve not on the buy side on exchanges. When the buy side gets eaten we’re under the assumption that we still have 50k in tier 4 not realizing its actually already used on the exchange. So when 50k buy side buffer is gone, and additional decline happens suddenly we’re faced with contractor default and no reserve.

And again, “potential downsides that in the end are not” that have already being discussed.

There is no difference having BTC in T4 as having BTC in buy side.

When that 50k buy side buffer is gone, we have had a dump of 50k, thus having had 50k in T4 as BTC would have been used to pay for that dump and cancelled each other. So with that premise of 50k dump we are broke with 50k in T4 and with 50k in DR.

We don’t even have/need to recover the collateral. The 50k dump have been saved by that collateral in form of buy side. We lost 50k of collateral, we won 50k not spended from T4.

Summarizing, if we can’t recover the collateral dumping it into the buy side, we don’t need to recover the collateral.

[quote=“ttutdxh, post:58, topic:3163, full:true”]
There is no difference having BTC in T4 as having BTC in buy side.[/quote]
There is, the T4 reserve (especially the 15% part) is supposed to be a backup reserve, thus when buy side swindles it can be balanced and T4 restocked with NSR sales.

This is true, that is why when contractor defaults there is no loss but a net profit for Nu. The only problem is that this is akin to putting our T4 reserve actively on T1 buyside. Normally we would restock T4 with NSR sale if it gets eaten by balancing with T1, in the DR there is no “loss” of funds when contractor defaults but an unexpected decline in T4 reserve. That is why it’s not suited as a true “emergency” reserve (which the 15% reserve is supposed to be).

You can’t assume that the buy side is larger by 50k for the entire duration of the contracts. It is a very long period of cash flows that you have to model. I also think that 1 NBT in the wild for 1 USD in tier 4 buy side is a bad trade unless in specific circumstances.

The entire liquidity structure has a lot of robustness to begin with (which comes at costs), to account for the unknowns and unquantified. We cut costs by finding ways to remove a few buffers at a time.

Now if you really want to cut into the 15% by this untested mechanism, you are removing more than a few blocks in a game of Jenga. Moving slowly has its own risks and costs, but I don’t think there has been analyses and modelling comprehensive enough for moving more quickly.

That is why I propose the mathematical model. That risk is true for any proposal, as simple as it may seem.
Would you support a grant for the development of that model?


@Dhume We can assume that we need to start the NSR sale when that buy side buffer is lower than the DR reserves. It is common sense. The DR reserves that now are buy side buffer are less than 15%, so T4 effectively has less than 15%

Edit: The first step when the 15% is reached would be ask DR contractors to buy from sell side, burn and proof. And if that is defaulted start with the NSR sale, as we only have the <15% buffer now.

Yes but imagine the DR contractor uses his BTC to buffer our active T1 buy side, thus getting double interest. Suddenly something happens (his house burns down w/e) that requires him to pull all his funds. Not only are we suddenly faced with a strong decline in our T1 buy side but also a similar reduction of T4 reserve (since he defaults). That means on our net balance we’re suddenly confronted with a “double” reduction in buy side liquidity.

Of course this still means a net profit of the 1% collateral for Nu, but that comes at the tradeoff of having unreliable liquidity. This is okay for an extra layer of reserve but not an actual reserve.

Forget about the 1% collateral/profit, that does not change the outcome, it is only a plus and a incentive.

Lets say this DR contractor uses 50k, and that T4 is 100k (50 this DR contractor, 50 BTC), later I will do the example with 50k T4, all in DR:

That scenario is, again, not only a default, but also a 50k decline in demand. In the first place, there is the decline in T1-T3 buy side. That requires us to re balance, yes. We pay for that. T4 down by 50k BTC, now have 50k in T4 as DR, which is actually in the T1-3 buy side buffer we just repaired.

Lets see what happens if we do not use DR, and there is a 50k decline in demand (50k liquidity disappearing):
We rebalance liquidity by 50k, T4 down 50k. T4 now 50k in BTC.


Now the example with 50k DR contractor and 50k T4 (all in this DR):

Again, not only a default, but also a 50k decline in demand. Buy side liquidity down by 50k, from the buffer of +50k. We try to rebalance to cover DR requirements, and ask the DR contractor to: put up those 50k in buy side/buy 50 from sell side and burn. He is in default, so we lose that 50k “buffer”. We now don’t have any DR contract, and a balanced (equal) liquidity without that 50k buffer. T4 is now 0.

Now the example with 50k, all of them in T4 as BTC:
Demand/liquidity down by 50k, we spend those 50k in T4 to rebalance. T4 now 0.

Magically, no diference…

[quote=“ttutdxh, post:63, topic:3163, full:true”]
That scenario is, again, not only a default, but also a 50k decline in demand. In the first place, there is the decline in T1-T3 buy side. That requires us to re balance, yes. We pay for that. T4 down by 50k BTC, now have 50k in T4 as DR, which is actually in the T1-3 buy side buffer we just repaired.

Let’s see what happens if we do not use DR, and there is a 50k decline in demand (50k liquidity disappearing):
We rebalance liquidity by 50k, T4 down 50k. T4 now 50k in BTC.[/quote]

So far so good, in your example in both cases T4 is down to 50k BTC worth, but in DR there is no actual BTC held by Nu but only a contract. Let’s say after the initial 50k decline in demand another say 10k decline occurs. Our T4 would go from 50 to 40k BTC held by Nu, in DR if the contractor defaults suddenly this 50k T4 reserve is gone and we are faced with an immediate 10k deficit for peg maintenance and no further reserve to speak off.

[quote=“ttutdxh, post:63, topic:3163, full:true”]Now the example with 50k DR contractor and 50k T4 (all in this DR):

Again, not only a default, but also a 50k decline in demand. Buy side liquidity down by 50k, form the buffer of +50k.[/quote]
There is no guarantee that this buffer necessarily stays on the order books, it might get pulled. If the contractor is the one essentially double providing on both T1 and a DR contract he could pull his 50k BTC and we are faced with a 50k decline in T1 buy side AND a 50k decline in T4 reserve (the DR contract).

Or like I stated above, he pulls the 50k from T1 and we are faced with a 50k decline on top of a 50k default in T4.

I have covered every potential scenario you describe with examples, in 20 lines. You have not found any inconsistency/error in the calculations and the premises/circumstances are the same for all the possible events.

Yet you keep altering the circumstances more and more, and describe potential problems that may be “potential downsides” but fail to prove your points with a comparable example between outcomes.

I am going to make a last example applying your last “potential downside”:

I hope you notice this is a change in circumstances/premises which leads us to another outcome with or without DR.

Another 10k decline occurs: We only have 50k in DR, no BTC left. Since we still have a DR contractor, buy side have a buffer of +50k. With the decline it is now +40k in buy side. We ask DR contractor 10k. Contractor defaults. We now have +40k buy side buffer. Bonus outcome: We need that $40k back in BTC to pay for something. No problem. Just dump 40k NBT and you will get the $40k in BTC in T4. Walls then will be totally balanced, no buffer. Everyone’s square.

Another 10k decline occurs: We pay for that with T4, T4 is now $40k in BTC.

Another 10k decline occurs: We have no money, we have to sell NSR before PEG breaks. The moment NSR sale should start is just when the first decline of 50k demand starts, and we notice the limit of 15% have been reached.

Another 10k decline occurs: We have no money, we have to sell NSR before PEG breaks. The moment NSR sale should start is just when the first decline of 50k demand starts, and we notice the limit of 15% have been reached.

As you notice we have the same money, and the same time to make the NSR sale.


If you don’t do a effort to prove your points before saying them out loud, then we have no way of moving forward.
I will now stop considering “potential downsides” if there is not, at least, two examples. One using DR and one without using it, showing the same premises/circumstances, and showing different outcomes.

I’m sorry for apparently not being eloquent enough to explain what I mean to you. Other people seem to understand but I’m failing to explain it to you. The fundamental problem in your scenario’s in favor of DR is that you don’t adequately calculate in where the burned Nubits come from.

Are they stale (aka already held by someone off exchange in a private wallet), are they currently part of our sell walls, are they bought of these sell walls or removed by the contractor who already owned Nubits and pulled them from T1 to burn them, are they bought directly from tier 4 by a private exchange between FLOT/Jordan with the future contractor etc. All these eventually end up having different effects. Also define what you mean by burning them, from what I understand burning means the coins disappear from the blockchain and thus also out of the effective money supply. So define your meaning when you say burning.

For example when they are purchased from T4 no effective change to demand occurs, unless they are not actually burned. You see me changing circumstances and premises because I just have to assume what you mean in different examples. So I urge to make an example for DR while accounting for all these factors.

So when you make an example of the workings of DR you should start with:

  1. Starting premises (how many outstanding Nubits, how many T4 funds, etcetc)
  2. Where do the Nubits to be used for the burning come from (already out in the wild, T1, T4 etc)
  3. What do you mean by burning, since different definitions have different effects on the money supply.
  4. What are the net effects of DR when taken into effect (these effects are different based on the different scenarios in step 2)
  5. What happens after DR is in effect and a decline in demand and DR contractor defaults.

If you can state a clearly defined example accounting for all these factors then I will try to show you what I mean.

I am yet waiting to see that supposed people, or their points, backed by a complete example:

None yet.

Already discussed

  1. Premises:

  2. Buy side is not constrained to be max 60% of total liquidity. By that, I mean that there is lower bounds so peg is not broken, but if 600.000 USD are put on buy orders that does not force Nu to put 400.000 NBT in circulation, like we should do with the current 40% rule. The problem and a potential solution to that current Nu problem is described here.

  3. DR reserves combined value is, as a maximum value, the difference of buy side from sell side:

* If that difference is lower than DR contracts, that difference have to be recreated by DR contractors buying directly from sell side and burning or putting up orders on buy side (by using NBT entry gateways set at less than 1 USD for example). Their debt will be adjusted accordingly. If that is not possible (DR contractor default), the contract is finished.
  1. 800k NBT in circulation

  2. $150k in T4 as BTC.

  3. Current liquidity is 60k USD buy side 40k NBT sell side.

  4. FLOT have supply of 25M NSR and 150k NBT to be injected on short notice, and unlimited more on shareholder request.

  5. Already discussed and example made. Feel free to make example with both premises.

  6. Burning:

I only know one definition of burning,
Are you aware of https://docs.nubits.com/rpc-api/#burn?

4 . Net effects to take into account:
Any net effect difference in the global value of Nu (G0), calculated as:

  • At the beginning of the example: G0: ((Executing all buy side and all sell side) + circulating NBT) - reserve money.
    Example with premises:
    G0 = (60k - 40k) + 800k - 150k = 670k
  • At the end of the example: G0 + (changes in global demand as events in the example).
    Everybody have to end with the same NBT they started, if they not, that is a change in demand.
    Example with premises and someone dumps 50k NBT:
    G0 = ((60k - 40k) + 800k - 150k) + (-50k) = 670k - 50k = 620k

in comparison to not using DR.

5 . What happens after DR is in effect and a decline in demand and DR contractor defaults:
That is part of the example.

It is very likely that I have made a mistake with the points, please check for contradictions or omissions before elaborating the example.

I think that a google spreadsheet is preferable to show such examples, but anyone is free to make the example in the way they like.

The point of all this is to make pairs of examples:

to prove that usage of DR is (or not) a positive addition to Nu.

[quote=“ttutdxh, post:67, topic:3163, full:true”]
DR reserves combined value is, as a maximum value, the difference of buy side from sell side:

If that difference is lower than DR contracts, that difference have to be recreated by DR contractors buying directly from sell side and burning or putting up orders on buy side (by using NBT entry gateways set at less than 1 USD for example). Their debt will be adjusted accordingly. If that is not possible (DR contractor default), the contract is finished.[/quote]

Does this mean you now want to limit DR to the difference possible in sell and buy side? Since that changes the whole proposal. This also means that DR in this light is limited in how big it can get, does that mean you are not proposing to use DR for the full 15% reserve? I have to think about examples since this is complete overhaul of your earlier proposal.

Secondly I want to address your point in the Demand based liquidity topic, you state:

[quote=“ttutdxh, post:67, topic:3163, full:true”]
But what about buy side? To fulfill an imaginary demand? A lot of people could be putting together millions in buy side orders just bellow 1 USD because they realize that buying NBT for less than that is always a wining trade. Does that mean that demand have increased? No, it just meant more people are aware of a profit opportunity, but they will immediately sell the NBT at more than 1 USD (probably at another exchange walls) to make their profit.[/quote]

For as far as I’m aware liquidity is only counted as what is reported by the pools. So only the funds actually earning interest and being placed by Nubot are being counted as liquidity. Someone placing a 100k buy Nubit order below or even at the same price as our buy walls is not being counted towards the liquidity, or used in the 40% criteria. For as far as I’m aware that is. Order placed outside of Nubot are not relevant for our liquidity calculations.

Arbitrage should not work as we maintain the same prices across exchanges using the same bot.

[quote=“ttutdxh, post:67, topic:3163, full:true”]
I only know one definition of burning,
Are you aware of https://docs.nubits.com/rpc-api/#burn?[/quote]

Yes this is also my definition, maybe I misinterpret but as I understand it that means burned Nubits are gone forever and no longer count towards the calculation of outstanding Nubits.

Since most of you examples implied huge demand drops, defaults, liquidity disappearing in a second, and in summary, a lot of concurrent huge problems, that premises would get a ideal state where DR isn’t affected by anything at all, even in those extreme situations (I hope). We will see if it theoretically works.

If it does, we can then fit the premises and rules to our current model wherever we want, and control where we put the risks in the system, and how big we want them. I will explain more once (if) this ideal model is proved robust.

As far I am also aware we use all liquidity, example here balancing polo from exchange data. But if it is as you say, even better. A premise less to comply with, and one problem less, because nothing would be in the way for creating those buffers.

Yes, I can’t think of confusion with other definition.

These changes are drastic compared to the earlier version, the only problem I see is that it makes it very likely that DR contractors will be called upon very quickly to add additional liquidity to our T1 buy side. It feels to me that it’s more of an addition to our current liquidity provision system.

It certainly is, but your circumstances were extreme, so that would be what it takes to fully cover Nu and maintain stable USD reserves in a ideal way.

Now if we just replace premise #1 and #2 for these new ones, we drop those drastic changes and we are back with the original proposal:

  • T4 reserves are formed by a smaller amount of DR (T4-DR) than the amount of BTC (T4-BTC).
  • Liquidity must be balanced first with DR money. We use T4-BTC only in case there is no DR money left.

Now the only drawback we have (not existent in the ideal version):

  • In the case all DR contractors defaulting at the same moment of the demand drop, the immediate demand drop Nu can sustain without breaking the peg temporarily is limited to the value of (T4-BTC + combined DR collaterals), instead of the total T4 value without using DR.

Let’s use T4-DR = 10% of T4, and T4-BTC = 10% of T4, as starting point for the examples (as you suggested).


I hope the first newly added premise changes your opinion that DR was “fundamentally flawed”, and “it makes no sense” to a more positive outlook. Either way it will be useful to move forward.

Anyone else noticing something not taken yet into consideration, please contribute your thoughts. Constructive criticism is always welcome.

I will properly elaborate a document with the complete detailed proposal in it’s various forms. This discussion is surfacing a lot of details required to handle DR correctly, which I initially haven’t considered, so it is being really useful for moving forward. Thank you @Dhume.

3 Likes

Another possible reserve could be constructed with BTC futures. For that purpose Nu could fund custodian or set up bounty for insurance.

I saw the first post come by at some point, looked interesting. Was there a specific reason why it never took off?

@ttutdxh As my understanding of DR has grown combined with your proposed changes I feel it’s getting better and better. I lack the time this weekend to delve more into it but I’ll try to contribute some insights (barring I get them) Monday/Tuesday.