Dreamed up a new HFT Market Maker on Deribit options

Jarett Dunn

2 min readSep 27, 2020

Hey,

Dreamed up a new High Frequency Trading market maker on deribit options.

Highly alpha.

Process:

1. to be included, options must have an implied volatility not equal to 0 (to exclude low-volume instruments)

2. we measure both the black scholes estimated value of a put or call, and the spread on bid/ask

3. if the black scholes determines the price is undervalued by a minimum threshold (currently 10%):

a. we take a look at those black scholes inefficiencies greater than the average

b. and the spreads over the spread average

4. trade them all

5. if our best bid is no longer the price, we edit the order price accordingly

6. if one of our orders is traded and filled, we enter a new sell order at bbo ask

7. adjusting order price if it’s no longer bbo

We should capitalize on both black scholes estimates for option value as well as the spread itself.

The risks are:

1. Should we enter a buy for an option that goes until expiry without selling

2. Market tending in one direction if we’re more heavily loaded on the other side

This is offset by:

1. if we successfully round trip buy and sell more than 4 options for every one not sold, we’re in profit. We’re avg 15+13%=28% on spread and bs. — deribit is the highest volume of the btc/eth options exchanges

2. random chance has us entering as many puts as calls, all else remaining equal

What’s working:

0. assess market conditions

1. enter into new buy orders, proportional to a % of btc/eth and a % of current # opportunities, leaving wiggle room to enter new opportunities as they arise

2. do not re-enter same buy orders for instrument already in order (probably: or in position)

What’s untested:

1. enter matching sell order (volumes on testnet very, very low)

2. lower # opportunities, higher % yield better? Or more # opportunities, lower % average yield?

3. forward testing in general

Anyone want to take up the bleeding edge of theoretical profits?