AIP2: Extend Epochs to 24hrs and Allow protocol to control the liquidity

The last few days have been seriously great for ARTH. We’ve finished one full expansion and half a contraction cycle. However, there are a few things identified by the team out which we’d like to propose the community two major changes.

The suggestions presented here will not only help control debt but will also further help bring the system back into contraction.


Suggestion #1: Motion to increase the epoch interval from 12hrs to 24hrs.

As mentioned before, the protocol needs to work in such a way that the growth in demand needs to meet the growth in market cap. However, in the early days of the protocol (as is what we are in today), it will take time for the market to catch up with the changes made to the protocol.

Now that we are in the contraction phase, it’s imperative that we avoid printing too much debt since it takes time for the market to catch up with the protocol.

Conclusion: Increase the Epoch time from 12hrs to 24hrs


Suggestion #2: Allow the protocol to control the liquidity to make it cheaper to hit 1$
While it seems at the very first historic contraction phase, the protocol burnt through almost 200,000 ARTH tokens and instantly brought the price of ARTH back to 1$, we have found two core issues that relate to maintaining the 1$ mark.

  1. Buying the 1$ is expensive as the liquidity of Uniswap is still high very high which means that we need to spend more Dai to get the protocol back to 1$

  2. ARTH bonds are being manipulated by bad actors who buy before ARTH bonds are issued and sell when the price is brought up by ARTH bond holders.

To resolve this we need to figure out ways to penalize whales that bring the price down and instead reward those who help bring the price back up. The solution mentioned here is a very simple one however it does not eliminate the possibility of whales manipulating the protocol, it only deters them.

In this suggestion we propose allowing the protocol to control the liquidity of the ARTH pool. The way this works is quite simple, in that ARTH LPs simply deposit their LP tokens to the protocol’s smart contracts and if the protocol finds that it has entered into the contraction phase, it’ll pull out the liquidity from Uniswap allowing making it cheaper the price to get back to the 1$ mark. Once the protocol hits the 1$ mark again, then the smart contracts add back liquidity.

The smart contracts should ideally use the 12hr TWAP price to avoid market manipulation.

Any excess ARTH from IL is given back to the LP provider along with a small reward for helping the protocol get back to the 1$ mark. This is an additional incentive for LPs to help bring the protocol back during a contraction phase aside from just bonds.

Conclusion: Allow the protocol to control the liquidity and reward those who remove liquidity during a contraction


This solution should help in getting the price back to 1$ and paying off the debt of ARTHB holders, however I do believe that we still need some more debate on how to prevent whale manipulation (point 2)


Inspiration for having a protocol controlled liquidity comes from https://fei.money/

3 Likes

We need these countermeasures in this moment, don’t like so much 24h epoch but we need that right now.
I still believe that the bond should be the second option.
During a contraction we use dai to buy back arth and stabilize the price at GMU, I believe we should do the same in expansion mode, allow the protocol to use a part of the new minted Arth and seel them for Dai and use this Dai like first option if arth <1$.

3 Likes

Nice suggestions just a few things here.

Regarding the first one, while it’s important not to continue printing while we are in contraction we need to also look at Arth’s current expansion too. We need to continue to expand at a growing rate. Given that Arth is a stable value coin lesser supply is not going to do any good. So for this reason I’d like to suggest a tweak to the first point.

If the number of Arth B tokens left to redeem is less than 3-4 % of the circulating supply of Arth we can go back to a 12 hour epoch system as we should be entering the expansion phase soon as the reduced number of Arth B tokens left for redeeming would imply that debt has gone down and now it’s time to expand.

If the number of Arth B tokens left for redeeming are still significantly higher the protocol takes the 24 hour epoch route to avoid incurring excess debt by minting these Arth B faster.

With regards to the second suggestion,

If the protocol does remove the LP it will aid in faster gain of the price but I’m not sure if that is enough to deter the whales.

Probably the only thing that can deter the whales are a percentage of transaction taken as a stability fee since they are working against the protocol.

For fairness if we can find a way to apply this fee on days of contraction it is even better or no expansion/ contraction then it’s even better.

An alternative solution would be to set a fee depending on the price of Arth when it’s being dumped by the whales. If Arth is being dumped between 0.98 to 1.03 for example the fee can be applicable (PS. I’m not sure if it’s possible to write a smart contract to do this based on the price of the asset but if it can be done then amazing)

Secondly, we can set a cap on the amount of Arth B a user can redeem per epoch. This way the whales cannot abuse the system and will have to wait in line for everyone else to finish redeeming their arth B tokens.

An alternative approach would be to have some sort of a vesting period for Arth B holders where a percentage of the bonds are released once in every 4 hours or 6 hours. For example Alice buys 1000 bonds and gets 40 % of the bonds ie. 400 tokens instantly which can be redeemed for Arth as it is minted. However the remaining 60 % are given out 15 % every 4 hours or 20 % every 6 hours.

This way we can effectively control the amount of Arth B held by individual at the initial epoch and the 12 hr or 24 hr epoch that is set to happen.

P.S : some of these suggestions may or may not be feasible to execute from a smart contract point of view and I might be suggesting impractical things too so feel free to take them with a pinch of salt.

5 Likes

Maybe I’m wrong but still, I decided to share my point of view. Don’t you have a feeling that increasing an epoch interval from 12hr to 24hr will increase instability of the algorithm?

What we have here is a closed-loop controller:

  • we have system input which might be considered as all the actions taken to stabilize the price, i.e. issuing bonds, minting ARTH etc.
  • we have system output which is current price
  • we observe the output (price) and use it as a feedback
  • we calculate error comparing reference (target) price and the feedback
  • based on the error we issue system input and we do it in intervals (with subsequent epochs kicking in)

The current control system has proved that it may lead to instability. Now we have to take measures to decide and apply the best control strategy to control and stabilize the system. What is proposed here is to change the epoch interval, however I think that having the system input issued less often will not solve a problem of instable output - it might create delay and even greater instability. To me, the interval should rather be shorter, but the amplitude of the control signal should be lower (not to overshoot the target). Maybe the corrections should be less in power, i.e. in the expansion phase not 30% but 10% of current supply should be minted and 2% should be available for a buy back in the
contraction phase, but the epochs should be, let’s say, 4hr? It seems to me that this will make the dumps and pumps less intense. And that’s what we need to have a stable system.

Imagine that you drive a car. Your road is straight and your main goal is to keep the car on the center line of the road. You assume that to achieve your goal it’s enough to touch the steering wheel and correct your steering every X (where X is unit of time). You try, but it appears that you drive from one band to another as every time X passes after you’d made corrective action you end up not on the center line but on the other side of the road. What you would do? Would you decide to touch your steering wheel every 2X assuming that from now on you’ll be able to issue appropriate control signal? Or you decide to check and correct your steering more often but with lesser power? I’d do the latter.

1 Like

So on this note; yes the expansion rewards should be decreased now; to something like 10% and further it should also be vested so that the participants do not dump into the coin. (mentioned here: AIP4: Option for re-staking rewards / Vested expansion rewards to reduce dumps - #2 by enamakel)

Reducing the epoch interval I might not be a fan of but I think the alternative which is issuing fewer bonds is a better option. Reducing epoch time right now will cause the market to not be able to catch up with the protocol.

Further, the main problem here is that there are too many thinking participants that work against the interest of the protocol. These mainly include whales manipulating arth bond buyers and further arth bond buyers cashing out.

I believe the protocol should in some sense have better control of the market hence the suggestion #2.