Conversation

There are some very cool things that you can do (e.g. replace young

with a smart contract) with CFMMs if you understand how they intersect with control theory, PDEs, and some differential geometry Few? @bluekirbyfi never taught me how to use this meme correctly

Quote Tweet

fiskantes in his true form

@Fiskantes

Sep 24, 2020

Replying to @Fiskantes @mrjasonchoi and 2 others

Market Makers would prefer the simpler form if they could but they are forced to adjust in adversarial environment. Ofc there are many nuanced counter takes to this, for example in some of @tarunchitra papers on CFMMs

Replying to

and

By young Matt Levine do you mean IB, analyzing news, or something else?

Replying to

and

Replacing derivatives lawyers at an IB with a smart contract 🙃

Replying to

and

Oh boy do explain

Replying to

and

Step 1: Find an integrable derivative payoff, g, that you’re interested in Step 2: Normalize by L1 norm g -> ghat Step 3: Construct CFMM that replicates ghat (see below) Step 4: Recursively borrow/lend LP shares on

@AaveAave

@compoundfinance

to get g NOTE: no lawyer(s) needed

Replying to

Control theory: Ensuring |LP share price, ghat| < epsilon PDEs: Controlling the Greeks for LP share Differential Geometry: Liquidity constraints, plus: do you really think you’re doing any math if someone doesn’t say, “didn’t Gauss prove that in the 1800s?” cc

Replying to

what role are you thinking of a lawyer playing an an IBank?

2:30 PM · Sep 24, 2020Twitter Web App

Replying to

The parts that Levine often mentions are: 1. Constraining the payoff function g to meet laws, regulations, clearing rules, etc. 2. Structuring an SPV whose shares represent the derivative asset (which is now replaced by CFMM LP shares) 3. Making sure there is liquidity at genesis

Replying to

Yeah the thing I'd say, though, is that the hard part of (1) isn't the math -- it's understanding what the constraints are.

Show replies