Conversation

Jesse Powell
@jespow
Oracles are only responsible for reporting the truth, not for controlling events or even what people do with the truth. To spot exchanges, the behavior is neither good nor bad. A robust index, like
@CFBenchmarks
is much less vulnerable to being influenced by a market anomaly.
2
3
21
Jesse Powell
@jespow
So, it is not so much an attack on the oracle as it is the exploitation of a naive counter party. The sucker accepted an unknown, variable price, which turns out could, under the terms, be set by the other contract participant. Oops. Not the oracle's fault.
1
1
21
Zhu Su 🔺
@zhusu
I think the word "fault" here isn't really a useful one, I agree. Main question is do underlying spot exchanges that are often used as oracles have any real or perceived esponsibility to the community to prevent this behavior?
5
13
Zhu Su 🔺
@zhusu
If the price on kraken goes from $10k to $1 and then $10k while nowhere else moves and all margin longs on kraken itself are liquidated, would we agree that their order book malfunctioned and that something needs to change?
2
2
16
udiverse
@udiWertheimer
Depends on the circumstances. If there was an actual malfunction, yes. But if the markets are just inefficient because some large prop firm randomly shut down, and you think there’s such a sweet arb opportunity lying around, why complain to kraken? Go ahead and take it
2
11
udiverse
@udiWertheimer
That said, an exchange has responsibilities toward its own customers, so it would be in their best interest to sort things out somehow if they want to keep the margin product alive. They have zero responsibility to keep bitmex plebs happy though
1
4
Zhu Su 🔺
@zhusu
Agree. The key philosophical question I find interesting here is whether exchanges (esp those in the CME index) have any responsibility to improve the smoothness of their trading because of derivatives based on them.
1
4
Jesse Powell
@jespow
Anyone is welcome to offer any exchange incentives to adjust operating procedures/performance. If they're not compensated, there is no responsibility/obligation. I think this is a bit like asking your ISP to filter all the large packets because you want to use a short password.
1
1
3
Zhu Su 🔺
@zhusu
Yes exactly the point of my initial tweet. You would be more concerned w margin longs getting liqed on kraken itself v margin longs on deriv exchanges which use your data but don't compensate you. I'm not arguing that you will or should care--only that you could care.
1
4
Zhu Su 🔺
@zhusu
Replying to
@jespow
@udiWertheimer
and 2 others
Exactly, hence the onus will be on the clients of the affected derivative exchanges to push for better index methodologies on those exchanges as opposed to expect oracle exchanges to discourage these attacks (b/c they aren't incentivized to in any way).
1
4
Jesse Powell
@jespow
Yes. If not better index methodologies then different liquidation rules, or both. You could use a single exchange's price but only trigger liquidations after X time (though the delay is also risky since the move could be global). Have to design contracts to sustain attacks.
1
1
5
Show replies