Hedging the crab auction

Joseph Clark
Opyn
Published in
6 min readNov 7, 2022

--

Market makers participating in the auction might try to hedge their oSQTH exposure with perp futures or options. This post describes some trade flows of auction participants to show how traders can manage risk and earn a return.

**************************************************************************

Herein we describe some common trade flows. It is not suggesting that these are a good idea. It is not financial advice. There are many other factors and uncertainties. If you want to have your fortune told, go to the fair.

**************************************************************************

(joint work with @reegannotraygan)

Each mon/wed/fri at 16:30 UTC, the crab vault rebalances by trading WETH for oSQTH. It trades here: https://www.squeethportal.xyz/auction

A completed crab auction trading oSQTH for ETH

Market makers participating in the auction might try to hedge their oSQTH exposure with perp futures or options. This post describes some trade flows for this.

Meet Jack

Jack just bought 1000 oSQTH in the crab squeeth auction for 0.06 ETH each (60 ETH total spend). At the auction settlement time:

  • eth price is $1500
  • squeeth implied volatility is 60.5%
  • premium (funding) is 0.1% per day

What happens now?

Trading the squeeth auction is a lot like a cake. There are many layers…

To make things simple, say Jack holds the squeeth for 2 days. What kind of things can Jack do to reduce his risk? How many layers are there in the cake?

Layer 1: Delta hedge (managing ETH exposure with perps on dydx)

By hedging delta, Jack is betting that volatility is underpriced over the two days that he holds the oSQTH. This is a gamma vs theta trade.

Jack goes to dydx and hedges the auction by shorting 120 eth on dydx at $1500/ETH.

After holding this position for 2 days Jack’s payoff looks like this:

Hedged long oSQTH

Say the ETH price moves to $1300 after 2 days and Jack sells the 1000 oSQTH back to the auction at 80.5% squeeth vol: :

  • The squeeth pnl is (90–68.27)*1000 = -$21723
  • The dydx perps pnl is -120*(1300–1500) = $24000

Overall Jack makes $2277

SqueethLab price and greeks calculation

Attribution:

(see appendix for details on squeeth attribution with greeks for more details)

Squeeth:

  • Jack loses 2*(1300–1500)*60 = -$24000 on squeeth delta
  • Jack makes (1300–1500)²*60/1500 = $1600 on squeeth gamma
  • Jack makes approx 2*0.605*17.5/365/100 * 20 * 60*1500 = $1044 on squeeth vega
  • Jack pays approx 2*0.605²/365*60*1500 = $181 in squeeth funding (theta)

Perps:

  • Jack makes -120*(1300–1500) = $24000 on dydx perps delta

Jack makes money here for three reasons:

  1. The squared ETH move was more than squeeth premium (because squeeth was underpriced). Jack pays $181 in premium but gets $1600 in gamma pnl.
  2. Squeeth implied volatility increased by ~20% and Jack was long volatility. Jack gets $1044 in vega pnl
  3. Higher order effects account for the difference between squeeth pnl and the addup from the greeks $2463–$2276 = $187

Layer 2: Single-strike gamma hedge (managing delta exposure with perps on dydx AND gamma exposure with at-the-money call options)

By hedging squeeth with futures and options, Jack is locking in the difference in funding between squeeth and options.

In the last example Jack hedges his delta, but is still exposed to convexity from oSQTH.

Suppose Jack doesn’t like leaving things to chance. Jack decides to hedge with futures on dydx AND by selling an at-the-money call option.

A 1-month option is trading at 80.5% implied volatility. Jack sells 70 1-month atm call options. The ratio of options to squeeeth matches the gamma. Jack uses this sheet to calculate that he needs 70 options to hedge his 1000 oSQTH

Jack also sells 82 futures on dydx to hedge the options (selling 120 to hedge the oSQTH and buying 38 to hedge the calls ).

The option exposure offsets the squeeth exposure.

Options hedge squeeth convexity

After 2 days vol is unchanged for both squeeth and options.

  • The squeeth pnl is (90–67.47)*1000 = -$22536
  • The dydx perps pnl for the squeeth hedge is -82*(1300–1500) = $16370
  • The option pnl is -70*(48.9–137.8 ) = $6215

Overall Jack makes $49

Option before and after
Squeeth before and after

Attribution:

Option trade:

  • Jack makes -70*(0.546)*(1300–1500) = $7630 on option delta (calcs generated in OptionLab)
  • Jack loses approx 70*0.5*0.00114(1300–1500)² = $1600 on option gamma
  • Jack makes approx 0 on option vega (vol doesn’t change)
  • Jack makes 70*(-835)*(2/365) = $320 on option theta

The total from greeks is -$6350 so higher order terms contribute 6215–6350 = -$135.

Squeeth trade:

  • Jack loses 2*(1300–1500)*60 = -$24000 on squeeth delta
  • Jack makes (1300–1500)²*60/1500 = $1600 on squeeth gamma
  • Jack pays approx 2*0.605²/365*60*1500 = $181 in squeeth funding (theta)

Total from greeks is 22581 so the higher order terms contribute 22536–22581= +$45

One way to look at this is that holding options vs squeeth locks in the difference between the squeeth funding and the theta for the matching amount of options (receiving 320 from the options and paying 181 on squeeth).

This is a squeeth theta vs option theta trade.

Layer 3: Multi-strike gamma hedge

Jack can keep adding options across strikes in the same way to keep improving the hedge. In this limit he will hold an equal amount of all option strikes scaled so that the total option gamma is equal to the total squeeth gamma.

Intuitively, this hedge works because a constant amount of options across all strikes replicates a squared payoff. See the power perpetuals paper for more details.

An even strip of call options replicates a squared payoff

Level 4: Multi-strike, multi-expiry

The ‘real’ replication of squeeth with options uses many different expiries as well as strikes.

If Jack wants to be precise. He can buy options with the weighting function in the paper, and rebalance frequently to maintain these weights. But honestly who has the time for that?

The utility of the extra hedge is that you have proper exposures to the implied volatility term structure of options (the difference in implied volatility across expiries). This will matter a lot in certain kinds of shocks when very short term options implied volatility moves a lot more.

Appendix: Squeeth greeks attribution

It’s sometimes useful to approximate pnl for some value V with

dV ≈ delta *dS + ½ * gamma * dS² + vega * dVol + theta * dt

Where

  • dV — change in value
  • dS — change in underlying price
  • dVol — change in implied volatility
  • dt — change in time

For squeeth, the greeks are:

  • Delta = 2 * (eth price per squeeth )
  • Gamma = 2 * (eth price per squeeth )/($ price per eth)
  • Vega = 2 vol * (17.5/365)*($ price per squeeth)
  • Theta = vol² * ($ price per squeeth)

So, for example if squeeth price is 0.06 eth, vol is 0.605, eth/$ is 1500 then

- ETH moves to 1300

- Squeeth implied volatility moves to 0.805

- 2 days passes

dV ≈ 2*0.06*(1300–1500) + 0.5 * 2 * (0.06 )/(1500) * (1300–1500)² + 2 *0.605 * (17.5/365) * 90 * (0.605–0.805) + (2/365)*0.605² * (90)

= delta pnl + gamma pnl + vega pnl + theta pnl = 24 + 1.6 -1.04 + 0.18 = 23.26

This actual value change repricing squeeth is 90–67.46 = 21.72. The difference is due to higher order terms.

The approximation is used in SqueethLab

--

--