# Gauntlet Liquity report findings
ref: https://liquity-report.gauntlet.network/

## TL;DR and key points
- **Stability pool needs another form of incentives(such as LQTY tokens), otherwise there will not be enough liquidity to protect from insolvent state.**
- **It is beneficial to have CCR higher in relation to MCR. This will result in less time spent in insolvent mode for the price of being in recovery mode**
- **The gas compensation for Liquidators appears to be sufficient for preserving the functionality of the protocol**
- **Borrowers should be incentivized as much as possible to maintain their ICR more frequently. This improves protocol health and decreases time spent in recovery mode**
- **Collateral price volatility has almost no effect on amount of redemptions**
- **Borrowed asset price volatility has significant effect on amount of redemptions(arb opportunities)**

---

## Overview about simulation

### What type of modelling is it?
Gauntlet uses Agent-Based Simulations.
Agent-based models are a kind of microscale model that simulate the simultaneous operations and interactions of multiple agents in an attempt to re-create and predict the behaviour of a complex system(protocol), such as Liquity.

Each Agent represents a micro part of the system. Agents can be: borrowers, eth volatility, gas prices, redemptions, network congestion, etc; each playing it's role in protocol behaviour.

NOTE: Gauntlet used data for loan openings from Maker CDP

NOTE2: Simulation assumes some parameters of the system to be perfect. For instance: when redemption happens, system expect borrowers redeem LUSD for ETH and immediately sell ETH for USD to lock in profits. This might not be the case in real world scenario

NOTE3: Simulation doesn't take into account LQTY rewards for Stability pool at all

## Effects of ETH volatility on system insolvency

Simulation ran for 3 scenarios:

1. low: ETH vol=41% annualized
2. normal: ETH vol=97% annualized
3. high: ETH vol=173% annualized

Input parameters:
- 14 days time window
- 365gWei gas price(why tho?)

All heatmaps have:
- X-axis as MCR(Minimum Collateral Ratio)
- Y-axis as CCR(Critical Collateral Ratio) - MCR
- Z-axis as % of time spent in specific state against total time(14 days)

### System insolvency sims against ETH volatility

#### Insolvent state
This simulation shows that there is minimal risk for the protocol when ETH volatility being "low" or "normal".

However, when ETH volatility is "high", protocol can spend up to 3.5% of time in insolvent state.


#### Recovery state

Heatmap below demonstrates(in relation to the chart above), that the bigger CCR-MCR value is, the more protocol might spend in recovery mode.
However, this can significantly decrease time spend in insolvent mode

#### Stability pool returns

One more interesting sim for Stability pool returns.
It seems that pool returns are low under "low" and "normal" scenarios and only rise when ETH volatility is very high.


### Conclusions:
- **Stability pool needs another form of incentives(such as LQTY tokens), otherwise there will not be enough liquidity to protect from insolvent state.**
- **It is beneficial to have CCR higher in relation to MCR. This will result in less time spent in insolvent mode for the price of being in recovery**

---

## Effects of dynamic borrower behaviour on protocol
As in previous sim, this simulation runs against 3 different scenarios:
1. No dynamic behavior: Borrowers never top up or withdraw collateral
2. 2-week mean: Borrowers check their troves every two weeks on average
3. 3-day mean: Borrowers check their troves every three days on average

Obviously, the percentage of time the protocol is insolvent significantly decreases as borrowers maintain the CR of their troves on a more regular cadence

Also, time spent in recovery mode decreases significantly if borrowers check their troves more frequently

### Conclusions:
- **Borrowers should be incentivized as much as possible to maintain their ICR more frequently. This improves protocol health and decreases time spent in recovery mode**

---

## Effect of gas prices on protocol
For this sumulation, again, there are 3 different scenarios:
1. 15 GWei
2. 80 Gwei
3. 315 Gwei

ETH Volatility is set to high number of 173% annualized

We can observe, that protocol has some insolvent time over 14 days,

but the time spent in recovery mode is relatively unchanged

### Conclusions:
**The gas compensation for Liquidators appears to be sufficient for preserving the functionality of the protocol**

---

## Analyzing effect of Redemptions on protocol
The effect of LUSD and ETH volatility on redemptions. Very high gas price is used (315 GWei) with mean_borrower_reaction_time of 14 days for these sims.

**NOTE**: Higher base rate indicates increasing number of redemptions at a given time

On the heatmap below you can see that num of redemptions, redemptions volume and base rate are dependant only on LUSD volatility on the market. Obviously, unpegging LUSD creates arb opportunities.

Interesting, that ETH volatility has almost no effect on num of redemptions

### Conclusions
- **Collateral price volatility has almost no effect on amount of redemptions**
- **Borrowed asset price volatility has significant effect on amount of redemptions(arb opportunities)**

---

## Outcome: Recommended Parameters

| Parameter | Current | Conservative | Aggresive |
| -------- | -------- | -------- | -------- |
| Minimum Collateralization Ratio (MCR)| 110% | 130% | 120% |
| Critical Collateralization Ratio (CCR) | 150% | 140% | 130% |