Conversation

SBF
@SBF_FTX
2) that being said, there are some improvements you could make. First, there are still some fake exchanges here. They're a bit tricky to catch--their orderbooks often look real! They tend to be quite thin though. I suspect you may be overweighting spread relative to depth.
3
7
SBF
@SBF_FTX
3) I'm confused about how you're combining markets. It's definitely not a straight average, not is it the top, or anything like that. What is it?
2
6
SBF
@SBF_FTX
4) I think it's weighting stablecoin-stablecoin markets too much. E.g. the PAX creation portal has a 0-slippage infinite size PAX/USD market at $1; so providing liquidity in that pair isn't as important. Does this mean paxos.com should have a score of 1000?
2
1
SBF
@SBF_FTX
5) And, e.g. FTX actually has this too! You can deposit USD and withdraw PAX 1:1 for infinite size, which is basically a 0-slippage PAX/USD market. But we don't call it a market, because it's.... basically a USD/USD market.
1
13
SBF
@SBF_FTX
6) The order size you check for is mostly up to $10k. That makes sense for many coins! But not for BTC/USD. A significant % of BTC/USD volume comes in lots of size > $10k. You probably want this to be a % of ADV instead (maybe up to 0.01% of an ADV).
Twitter Web App
9
Likes
CoinMarketCap
@CoinMarketCap
Replying to
@SBF_FTX
and
@SBF_Alameda
Based on checking trade size distribution of BTC USD pairs across various exchanges, the 75% quartile is still relatively small. We used this distribution to inform the different order sizes we simulate to produce our score.
1
1
SBF
@SBF_FTX
Replying to
@CoinMarketCap
Is that 75% quartile of trades or of volume? I.e. if there are 6 trades of sizes 1, 1, 1, 1, 1, 90, would you say the 75% quartile is 1 or 90? I think the answer here should be 90.
1
Show replies