Conversation

1) twitter.com/cyounessi1/sta challenge accepted ;)

Quote Tweet

11) Now just waiting for big

to tell me why this entire thread is dumb and/or horribly wrong. cc @SBF_Alameda @danrobinson @abandeali1

Show this thread

5

2

18

2) So, I *do* think AMMs make some sense for mean-reverty spreads. Like some premiums. Also, like USDC/TUSD/etc. (Curve!)

2

1

3

3) But you could do the same thing on a DEX. the tricky thing here is that the two cases

presents aren't quite the same. Say there are two things--A and B--and over time A/B --> 1 (i.e. they converge). In his case, treasuries and their futures.

2

4

4) In the CEX case, Cyrus has these markets: -- A / USD --B / USD In the AMM case, Cyrus as this pool: --A / B The reason he gets the "better" result in the AMM case is because it listed the right market.

4

3

5) But in the CEX case-- --they could have just listed a market of A / B, and then his firm could have put out lots of limit orders to provide--buying A/B @ 0.999, and selling A/B @ 1.001

5:20 AM · Oct 16, 2020Twitter Web App

Replying to

6) And in the AMM case, if they instead had A/USD and B/USD pools, then when things moved the LPs would get picked off in both at the same time. They key here is having a single market that expresses the mean reversion rather than a two legged trade that is latency dependent.

4

1

7

Replying to

and

So the A/B market (guaranteed implied spreads) never trade for profit, are horribly illiquid, horribly inefficient! And trying to trade A/USD vs B/USD leads to slippage. Doesn't an AMM fix that (or at least, that's my argument). So it seems to me that the AMM is very... (1/2)

1

Replying to

and

...similar to the former (which would be bad), with one caveat being that you get part of the profit fee upfront, whereas with guaranteed spreads you have to actually cover the position to profit. If AMMs are identical to guaranteed spreads then I'd be more open to your viewpoint

1

1

Show replies