Help us by sharing this post on 𝕏!
Hey Curve, you might wanna take a look…
Oh noes! Is this finally the end of $crvUSD?
From the screenshot, it appears $crvUSD is off by a whole half a penny. If the whole ecosystem hasn’t already collapsed, then rates might have to spike to infinity if there’s any hope of keeping the wheels on the bus! TIME TO PANIC SELL…?
But wait just a second… we’re up to date on our Ape tutorials… why don’t we quickly scrape the chain history…
That’s odd… can you show me on the chart where the depeg touched you?
Curve would later confirm no depeg existed, on this or any chain.
We were promised a 0.5 penny discount, and the best you can do is 0.15 cents? That’s smack dab in its usual trading range…
What happened?
If you’re not familiar, the concept behind Y2K Finance is a sharp idea. Their docs describe it as a “suite of structured products designed for exotic peg derivatives.” In other words, if you’ve got money in stablecoins, you could hedge against the risk of depegs… making it a sort of perfect counterpart to a variety of DeFi products.
We’ve written positively about the service before. Depegs are a real and potentially devastating event. The existence of a service to hedge against such risk could be huge!
Incidentally, congrats to $MAI, which is on track for repegging!
A challenge for Y2K Finance is the oracle problem that’s faced many protocols. If the oracle lacks precision, then you end up with a service to bet on oracle performance, as opposed to the underlying depeg risk.
How did it miss the marker so badly and sound the alarm here? Let’s flip through their gorgeous UI…
First thing we notice is that they’ve got a lot of adoption… 9 figures is really good!
But… DeFi Llama reports just 0.15% of this TVL?
One guesses the phrase “has been deposited” is doing most of the heavy lifting on the website. Counting TVLs cumulatively is a nice trick though… they’re just a few flashloans away from becoming the first trillion dollar protocol in DeFi…
Zooming into their markets page, we see the DEPEG ALARM is still firing off for $crvUSD, even though the asset price is listed at a healthy $0.9983 — it looks like the website UI appears to rely on a different (and more accurate!) price feed than its own depeg market.
These dueling price indicators would also show up in the automated alarm that got posted to 𝕏, somewhat undermining its power to FUD efficiently.
This post also has some other curiosities. We’ll ignore the mention that $crvUSD depegged to 0.995 (WETH) — we assume it’s an unclear reference to the fact it’s a WETH-based market, and not an updwards depeg of several thousand dollars.
What’s strange is the linked transaction, which doesn’t seem to have anything to do with $crvUSD (or maybe we’re left-curving again?).
Heading deeper into the site, we can get a sense of just what is going on.
It turns out that the $3600 worth of bets in this market are betting on the price action reported by Umberlla/Umbrella Network, but it’s not immediately clear where this oracle is getting its price from.
Word of goddess has it that cross-chain issues from the early days of $crvUSD are to blame.
Hey Umbrella, you might want to take a look…
Oracles of Hell-Fi
The core issue facing Y2K Finance is emblematic of the broader issue that’s been faced so many similar protocols that try to build atop oracles. Oracles are great, but remain prone to occasional blips of errant data. The more your protocol is reliant on oracles, the more you find yourself at the mercy of getting felled by oracle misbehavior, not the underlying truth it’s trying to report.
If you’re playing the prediction markets, you need to check what counts as the resolution… the actual event, or the reporting/misreporting of it…
Y2K users were not betting on a $crvUSD depeg, but betting on an oracle’s accurate reporting or misreporting of a $crvUSD depeg.
In this case, it turned out to be harmless fun — the handful of people and $3K of liquidity in this market surely didn’t mind making a little extra money off the false positive. And we doubt Y2K went broke even if they were financing the losing side of that trade.
Oracle overreliance can turn into trouble if you were, say, building an automated service to try to hedge such a trade with big money. There was no actual countertrade anybody could make off of the misreported price. If users had built this into a dApp moving serious money based on this trade, it could have backfired pretty badly.
Curve ran into similar questions of construction when building $crvUSD. The stablecoin needed a reliable oracle price that was resistant to manipulation. The integration involved a Chainlink failsafe, which ultimately was disabled as the failsafe only made things worse.
Zoomers are too young to recall, but according to boomers the year 1999 was a perilous time for nerds. Computer storage being expensive back then, years were often only encoded with two digits. The concern was computers would misread 2000 as 1900, which would somehow cause airplanes to fall from the sky. Maybe the computer would claim jet engines weren’t invented in 1900 and thus disintegrate?
The “millennium bug” was an old-fashioned oracle problem of sorts. The issue at stake was a computer asserting a different truth value from reality. The ultimate concern was not so much a single program getting confused, but the emergent nature of embedding systems atop systems. Fortunately the calendar flipped to 2000 without any issues, although concerns about overreliance on interdependent systems quietly persist as technological entanglement has grown exponentially.
Fortunately, for anybody eager to experience the hysteria of 1999, we can feel relief that the spirit of the millennium bug lives on today, in the aptly named Y2K Finance.