Content pfp
Content
@
0 reply
0 recast
0 reaction

Haardik pfp
Haardik
@haardikkk
AMMs: - better UI/UX - largest trading volume - available everywhere - actually decentralized - nicely composable - crypto native - not wintermute why do you think CLOBs are still better?
10 replies
2 recasts
15 reactions

Jorge Schnura pfp
Jorge Schnura
@schnura
Liquidity efficiency
1 reply
0 recast
0 reaction

Haardik pfp
Haardik
@haardikkk
what do you mean?
2 replies
0 recast
0 reaction

Jorge Schnura pfp
Jorge Schnura
@schnura
On a CLOB I don't need to lock up my capital to make the market, which allows me to make the market in a much more capital efficient way. Also, AMMs are decent for spot products, but for derivatives they work worse
1 reply
0 recast
1 reaction

Haardik pfp
Haardik
@haardikkk
Agreed for derivatives If you place an order you do need to lock your liquidity. The equivalent is simply placing that liquidity highly concentrated in an AMM - no?
1 reply
0 recast
0 reaction

Jorge Schnura pfp
Jorge Schnura
@schnura
I can use the same liquidity to place multiple limit orders. The first one to get filled would close the rest. It's not real liquidity (otherwise it would mean leverage) but theoretical leverage. In any case, it's more efficient than AMMs. What really kills AMMs though is fees. There's no way we @turingcapital could operate with Uniswap's fees for example
1 reply
0 recast
1 reaction

Haardik pfp
Haardik
@haardikkk
Good point Re: fees - isn’t 10bps standard on most CLOBs? Uni has several 10bps pools
1 reply
0 recast
0 reaction

Jorge Schnura pfp
Jorge Schnura
@schnura
Not really. Funds like us have fees ranging 2bps to -2bps (rebates) on maker orders and 3.5bps to 2bps on taker orders, depending on volume. 10bps would kill the performance of many of the strategies we run. Basically all the trading strategies, we'd have to rely solely on asset allocation strategies (low volume) and buy and hold (very low volume)
1 reply
0 recast
1 reaction

Jorge Schnura pfp
Jorge Schnura
@schnura
And btw, the reason why fees are high is because the liquidity model is inefficient and hence traders need to compensate LPs for the inefficiency. Otherwise LPs would lose money (most still do across a cycle vs holding the assets separately). It's worth it to keep researching to improve AMMs, but it's also important to note that there wasn't anything broken with the CLOB model, we invented AMMs because of the limitations of blockchains, not to solve an inherent problem
1 reply
0 recast
1 reaction

Haardik pfp
Haardik
@haardikkk
Interesting The 2-3 bps fees you’re talking about, is this on CEX CLOBs or can you find these on “decentralized” order books as well a la Aevo/Orderly/intent protocols etc?
1 reply
0 recast
1 reaction

Jorge Schnura pfp
Jorge Schnura
@schnura
We only trade on CEX because of volume, but there's no reason why you could not find those on on-chain CLOBs. Once blockchains become performant enough for on-chain CLOBs we'll probably see a shift to that type of model with on-chain market making algos that attract liquidity through vaults
1 reply
0 recast
1 reaction

Haardik pfp
Haardik
@haardikkk
For certain top pools AMMs like Uniswap now also have a 1bps fee tier Is the main hesitation there needing to lock up liquidity in that case? Hypothetically an AMM with all 1-2bps fee pools for all top token pairs is still worse for activity like yours because you can’t place multiple orders simultaneously?
1 reply
0 recast
1 reaction

Jorge Schnura pfp
Jorge Schnura
@schnura
No, in that case it's good enough for taker orders (not for maker orders). The issue is that only a few tokens have half decent liquidity in that fee tier pool. Even fewer if you take those on mainnet out (gas is too high). For HFT or UHFT even the fees on L2s are too high + TTF too slow
1 reply
0 recast
0 reaction