Conversation

6) Anyway: There are a lot of problems with the proposal. I'll start just by noting the inaccuracies.

7) a) "3 - If FTT were de-listed, no other users would likely be impacted; there is very little supply volume, and very little borrowing demand for FTT" False. This would drastically reduce borrow demand, lenders would no longer get interest, and TVL would go way down.

8) b) "5 - SBF borrowed YFI which destroyed its price, he sold it on Binance and other exchanges - only once he was caught, did YFI go back up" False. 200 net YFI short over days does *not* destroy it! This is just off by an order of magnitude. The impact wasn't huge.

9) c) The post only cites twitter posts in one direction, which is not in line with "many threads of controversy". d) The third option (partially...) is massively under-specified.

10) Also what are the thresholds for a CREAM vote? Was that determined anywhere? What if there's a plurality but no majority?

11) But lurking underneath: some pretty large, nuanced issues. a) The post is anti-shorting. What exactly is the point of a borrow/lending protocol, if borrowing coins is evil? b) Most of the YFI was borrowed for liquidity and farming, not selling or shorting.

12) c) what are CREAM's biggest risks? Well, single asset risk is one thing. But sometimes it's not the biggest. There's about $70m combined between ETY, YFI, BAL, COMP, CREAM, LINK, LEND, CRV, MTA, SUSHI, and UNI. That's similar to the FTT balance.

13) Which has higher crash risk--FTT, or that basket together? Well, on May 1st, FTT was 15% lower. Half the DeFi coins didn't exist, the other half were on average 60% lower. Similar on 6/1, and 7/1, and 8/1. DeFi is way more volatile than FTT.

14) In fact, it's *so* much more volatile that even though FTT is one asset and the other DeFi coins are a basket, the *correlated* part of the basket's volatility is still higher than FTT's volatility.

6:02 PM · Oct 10, 2020Twitter Web App

15) In both directions. On 9/1, FTT was 20% higher. The *average* of the DeFi coins was 240% higher.

Replying to

16) The data is pretty clear here: The DeFi coins in CREAM pose a greater risk than FTT. Much, much greater. So much that FTT blowout risk doesn't really register compared to the other risks in DeFi right now.

17) If you're worried about a 60% move causing liquidations in CREAM: This has happened about once/month for the DeFi basket together recently. FTT hasn't been at a 60% different price in 2020.

18) So, FTT is probably less risky than the rest of CREAM, but it might get banned. It'd better be useless for it then, right?

19) Well, it *is* less useful per $ than many of the other coins. But I think the community is not correctly understanding the implications of banning that much of the collateral. a) CREAM TVL down 30% b) CREAM borrowing down 40% c) Interest earned by lenders down 40%

20) No one will use CREAM for large size again -- it loses a lot of its future value. (and, of course, some partners, liquidity, etc.) I think this probably nukes something like 20% of the protocol's value.

21) That's a lot! Giving up 20% of value based on an emotional, error-laden tirade to a single large voter for one of the lesser risks of the protocol. Because some people don't believe in borrowing (on a borrow-lending protocol, none the less!)

22) So, yeah, this is a pretty crap-tastic proposal. That being said -- I think some variants would be reasonable! Want to decrease collateral to 40%? Want to cap any single asset at 20% of the total supply? IDK if those are right, but they're totally plausible.

23) But this one-- --this one throws the baby out with the bathwater. Without having done its homework.

24) If that's what you want for CREAM--fine. But also then why use CREAM and not Compound? If you want a small, neatly curated list of assets in a static system, that's what Compound specializes in. In CREAM, I see a dynamic, broader protocol, for better or for worse.

25) But, idk, in the end this is the community's decision. Or, you know, 0xDa495C2Ab0a91623564126778D5AB20fA87C1DFc's. Either way.