Introducing the Pre-diabetes in Sugaryield.com (Part 2)

SugarYield.com ®
5 min readJan 20, 2023

--

The Pre-diabetes vault is further divided into two different vaults, that represents the buying and selling of Insurance. In simpler words, that represents your stance in the trade. The two vaults that you will be coming across are

  • Caramel Vault: represents hedging against a particular pegged asset
  • Saltish Vault: represents selling of insurance, (Pro-risk) position

Payout Calculations

Here is the methodology via which Payouts are calculated for the Caramel and Saltish Vault.

We define the following variables:

  • Bi is the premium paid by each insurance buyer
  • Sj is the collateral posted by each insurance seller
  • NB is the number of insurance buyers
  • NS is the number of insurance sellers, and
  • c is the trading fee taken by the protocol.

In the event of no depegging event,

  • Each buyer i pays Bi, and receives nothing
  • Each seller j receives
  • in addition to the return of their collateral.

If there is a depegging event,

  • Each buyer i net receives
  • Each seller j net pays

Numerical Example

Let B1 = 1, B2 = 3, S1 = 150, S2 = 200, S3 = 50, c = 0.005. It is clear that, before accounting for trading fees, the yield in this scenario for a single epoch is 1%.

In the event of no depegging event,

  • Buyer 1 pays 1 BNB,
  • Buyer 2 pays 3 BNB,
  • Seller 1 receives 1.5 × 99.5% BNB
  • Seller 2 receives 2 × 99.5% BNB, and
  • Seller 3 receives 0.5 × 99.5% BNB

If there is a depegging event

  • Buyer 1 pays 1 BNB and receives 100 BNB, leaving him with (100 × 99.5% − 1) BNB,
  • Buyer 2 pays 3 BNB and receives 300 BNB, leaving him with (300 × 99.5% − 3) BNB,
  • Seller 1 receives 1.5 BNB and pays 150 BNB, leaving him with (−150 + 1.5 × 99.5%) BNB,
  • Seller 2 receives 2 BNB and pays 200 BNB, leaving him with (−200 + 2 × 99.5%) BNB, and
  • Seller 3 receives 0.5 BNB and pays 50 BNB, leaving him with (−50+ 0.5× 99.5%) BNB.

Determining Strike Prices

SugarYield aims to eventually expand to provide insurance markets for various different stablecoin protocols. Here we provide a sample methodology for determining strike prices, which someone may seek to use when creating new SugarYield vaults. At genesis, SugarYield epochs will be assigned on a per-month basis; that is to say, there will be an January epoch, Feb epoch, etc. This is subject to change according to protocol governance votes.

The methodology presented below is for a stablecoin pegged to 1 USD. In this framework, we assign three strikes to each stablecoin:

• K1 is the riskiest strike, which is expected to be breached every 3 months.

• K2 is a “medium risk” strike, which is defined as being breached every 18 months.

• K3 is a “low risk” strike, which denotes black swan events. These are the lowest yielding but provide protection against events that are unexpected to ever occur over the lifetime of a stablecoin.

We assume that price deviations from $1 are independent and identically distributed random variables.

Where Si is the stable coin price at a given time Ti. It is well known that in times of mass decollateralization spirals, these variables become correlated and the i.i.d. the assumption above does not hold. We deal with this by assuming that the discrete time series of stablecoin prices is sampled from a “continuous” (block-by-block) series

. It is well known that in times of mass decollateralization spirals, these variables become correlated and the i.i.d. the assumption above does not hold. We deal with this by assuming that the discrete time series of stablecoin prices is sampled from a “continuous” (block-by-block) series Ŝ,

Where we assume that any correlation spirals happen within each interval

.

Each strike Kk∈{1,2,3}has an associated rate rk , which is defined as the probability that the strike is breached within a given ΔT . The rate is calculated using an indicator function from the discrete time series

, as

and can be used in a binomial distribution to find the probability Pk of a particular strike being breached within a given month:

Where f is the sampling frequency and d is the number of days in a given epoch. The equation above is then solved for each rk

on the interval

given the desired values of Pk. This can be done using a variety of root finding algorithms. Once

rk is determined, the set of all Xi can be iterated through for varying strikes until an appropriate K is found. For the cases K1 and K2 , each Pk, respectively, is 1/3 and 1/18

This wraps up our series covering Diabetes mechanics, simulations for the protocol’s pricing models can be found in our gitbook. You can follow our Twitter account @SugarYield and join our discord for updates and announcements.

For more information, Visit:

Website | Twitter | Telegram | Reddit |Medium | Documentation | Github | Smart Contract | Youtube | Contact

--

--

SugarYield.com ®
SugarYield.com ®

Written by SugarYield.com ®

SugarYield.com is a (DeFi) platform offering a wide range of financial services, including lending, borrowing, Hedge or speculate on the risk of pegged assets.

No responses yet