[mrintellectual]

The crux of the problem is not Luna itself, but the Anchor Protocol (19.5% yield), which enticed users to burn Luna by investing in UST through a, for lack of a better phrase, Ponzi scheme. I get it - people like decentralization. But without some form of regulation, either by the broader crypto community itself or by governments around the world, situations like the LUNA/UST collapse will keep happening.

There are also rumors of a Terra fork: https://agora.terra.money/t/terra-ecosystem-revival-plan/870.... Thanks, but no thanks.

[hristov]

It is a ponzi scheme. There is no production to provide for this 19.5% yield. It is all taken from incoming investments into the system. And it is well known that ponzi schemes collapse when the influx of new money starts becomes insufficient to pay the yield. It is very natural for ponzi schemes to collapse even without any government involvement.

I think the article should have examined that possibility in addition to speculating about an attack.

It is very possible that the $2 billion that was removed on the weekend was simply a holder that having seen worldwide retreat from risky assets had decided to pull his/her money out.

[nightski]

Except those incoming investments are all matched by an equal amount selling . That’s not something you see in a Ponzi scheme.

[vkou]

The selling is just smoke and mirrors that obfuscates the fraud a little.

Anything promising risk-free 20% YoY returns is either a printing press, or a fraud.

[kristjansson]

Or extremely risky!

[beefield]

Not sure if there is sarcasm there, but it is pretty much a definition of fraud if someone claims that an extremely risky asset is risk-free.

[kristjansson]

Missed the risk-free in GP comment…

[hristov]

That is not true. The supply of luna coins kept increasing, which means that new money was being pumped in without corresponding sellers. Furthermore, while on one side luna coins were being destroyed to support the tera peg on the other side new supply of luna coins was being created and new money brought in. This new money went to the luna foundation which used it to pay the yield on the anchor protocol.

[nightski]

So anything with inflation is considered a ponzi scheme in your eyes?

Just want to mention that I in no way support algorithmic stable coins, I think they are a joke (including luna/terra). But I also don't agree crypto currencies are generally a ponzi scheme (there may be a few exceptions).

[johntb86]

It's the impossible trilemma - you can't have

* a fixed foreign exchange rate

* free capital movement (absence of capital controls)

* an independent monetary policy

As a stablecoin, they claimed to have

* A fixed exchange rate with the dollar

* Generally free movement in and out (except for the 21-day redemption period, etc.)

* an interest rate untethered to risk-free rates you could get on USD

so it was impossible for it to work.

[hristov]

I am not talking about crypto in general but the ancor protocol in particular. The ancor protocol provided 19.5 % yield without having any asset that actually produces that kind of income. It provided the yield from new investments coming in, which is the definition of a ponzi scheme.

[wlll]

> It is very natural for ponzi schemes to collapse even without any government involvement.

I read it as government involvement would stop the collapse of ponzi schemes by regulating them out of (legal) existence rather than by somehow propping them up.

[shawabawa3]

It wasn't a Ponzi.

Yields came from outside investment. It's closer to Ubers incentives in new markets than Ponzi

[nomad225]

What outside investments?

[zucker42]

Patrick Boyle argues fairly convincingly in my mind that the problem with this was the design of Luna itself[1]. TL;DW TerraUSD maintains its price by minting $1 worth of Luna in exchange for one TerraUSD coin. But this assumes that there are always going to be willing buyers for Luna, even if its price is in free-fall. When the price of Luna drops suddenly, desire to hold Luna decreases, causing TerraUSD to lose its peg, causing confidence in Luna to drop further, leading to the death spiral we currently observe.

[1] https://youtu.be/iisPX_xVMV8

[cmckn]

> minting $1 worth of Luna

One thing I’m not clear on: what oracle is deciding how much Luna is worth $1? How do you calculate a market price in a manner that the chain itself can rely on to function? I assume there is a smart contract doing this minting, but if that smart contract depends on an oracle…how smart is it, really?

[initplus]

In this case the Oracle is a consensus protocol where community nodes report the LUNA/UST rate.

But of course to make running such a node easy for users Terraform Labs distributes node software that reports a value chosen by Terraform. So in spite of the distributed consensus design the Oracle is just Terraform.

[lamontcg]

It sounds like the fundamental problem with it is that it was founded on the idea that scarcity necessarily generates demand.

[vba616]

Everybody makes mistakes. Everybody has to learn everything for the first time.

I would suggest the fundamental problem is that nobody considered looking for counterexamples to assumptions in "legacy" financial markets.

Like, how could people who are excited about reinventing finance not be interested in the odd corners of existing markets, like penny stocks and whether/how they go to zero? Or financial crises and panics.

Sure, most people in the world couldn't care less, but by definition, if you are making cryptocurrencies, obscure aspects of finance are your idea of a good time!

[55555]

Because the end users aren’t actually that interested in finance. They just want to get rich quickly without effort.

[netheril96]

One doesn’t need to look at traditional finance. A similar algorithmic “stable coin” Iron/Titan collapsed a year ago. The problem is with people not understanding anything before jumping in.

On the other hand, jumping in without understanding is why a small fraction of people got ridiculously rich from SHIB.

[smrtinsert]

Seems like something that should be so easy to catch in a simulation/test

[px43]

A simulation? This flaw should be obvious to anyone who took 10 minutes to read how the system works.

Synthetics backed by a 1:1 mint/burn are always going to crash and burn the same way. It's happened before, and it will probably happen again. High profile people have been shouting from the rooftops about this exact issue since launch.

I have some friends who lost everything in this, and it sucks. I did my best to explain this exact scenario to people who would listen, and ended up getting some pretty high praise this week from people I saved, but you can't save everyone.

We are now in a world where wealth is being rapidly redistributed from people who don't read the instructions, to those of us who do. The metaphorical sea levels on global finance are rising rapidly. Learning to swim now isn't just a recreational activity, it's a survival skill.

[zucker42]

> We are now in a world where wealth is being rapidly redistributed from people who don't read the instructions, to those of us who do.

Are the people who "read the instructions" really the ones who benefited here? The main winners here, as far as I can tell, are the employees and creators of Terraform Labs—who benefited from people investing in a project that was doomed to fail—and the people who got in early and out early—who are probably more lucky than smart.

[px43]

It's still unclear who initiated the original attack which set this whole thing off, but whoever did that likely took home billions exactly because they understood how this system works.

From the looks of it, there were at least two distinct events around 9am Pacific on Tuesday, and again on Wednesday, where someone very intentionally made massive moves to trigger cascading liquidations. The rest of the damage was caused by a significant loss of faith in the system, followed by a mad dash to the exits.

The people running Terraform were doing a lot better before shit hit the fan. Maybe Do Kwon was looking for an exit or something, but I doubt it was an inside job. Maybe I'm wrong about that. I basically started ignoring Terra/UST the moment I realized how they were trying to back their stable coin. There are enough good projects in the space, wasting energy ruminating about obviously dumb concepts is a liability.

[rocqua]

I can't understand how an attacker could benefit from this without having short-sold Luna or Terra. I figured that would be hard to do. Am I wrong?

[binkHN]

> Synthetics backed by a 1:1 mint/burn are always going to crash and burn the same way.

What are the top few similar synthetics out there right now if you don't mind me asking?

[wtracy]

There's one cutely named Synthetix. If you look into the rabbit hole of "algorithmic stable coins", many of them use the same mechanism.

What made Terra interesting to me is that Luna is usable for paying gas costs. All equivalent systems that I've seen have a backing currency that has no uses other than minting synthetic assets or pure speculating.

[braingenious]

> We are now in a world where wealth is being rapidly redistributed from people who don't read the instructions, to those of us who do.

Can you link me to these instructions?

[krapp]

The complete instructions for all cryptocurrencies are as follows:

1) "Caveat Emptor"

[derivagral]

I have no stake in crypto besides a bit of eth. 2 coinbase examples. First: your deposits are covered only if they're denominated in USD. Second example was the fine print on Coinbase's crypto-loan thing where they said they're giving custody to a 3rd party and giving you all the risk. Good luck on that APY!

A free third example is all of these "APY" and yield-farming things in defi (usually) project figures based on a couple of early whales moving into a shitcoin, getting random people into it, then swapping liquidity to the next one. The initial move isn't exactly a lie, but assuming the next 6-18 months will sustain the advertised number is (imo) quite deceitful. Now you get to "stake" and lock yourself in for months! Yikes. That's on top of liquidity vanishing in a crisis.

tl;dr the nuance on this stuff is detailed, and the details are important in times of crisis

[el-salvador]

> I have some friends who lost everything in this, and it sucks. I did my best to explain this exact scenario to people who would listen, and ended up getting some pretty high praise this week from people I saved, but you can't save everyone.

Same here.

I was asked about crypto many times since it became legal tender in my country.

The country definitely went through an 'investment' craze for some months. Maybe in a similar way to what Robin Hood / GME was in the US in late 2020/early 2021.

For some people the damage was limited to the $30 they received from the government in their wallets, as a consequence of buying high and selling low.

For others who went all in and bought crypto with credit cards, is a different story.

Like for many people here the best investment they can make is paying up their up to 49.99% APR credit cards, but that doesn't seem exciting enough for many.

I remember having conversations with friends and acquaintances about investment options. Researching which bank offered the highest APR for term deposits, (they can sometimes be up to 5.5% for USD time deposits here), which investment options were available, or how usually investing in yourself (education) or paying debts is a much better investment.

At least one of them decided to go full on crypto "deposits" with higher APR, I haven't asked if they are underwater after the recent news.

Another thing that surprised me is that the ones that spent the most weren't IT people. Since originally here, Bitcoin was mostly an IT niche topic. A few lucky people that mined when it was possible to do it with GPUs or bought when it was bought about a $100.

With all this, most of my friends in IT were like "oh, I've known about this for a long time, I wish I had bought them years ago" but didn't take major actions. On my case I remember the first Slashdot post about Bitcoin and even downloading one of the early Bitcoin Windows wallet but I never mined.

Others were mostly bit annoyed for having to work after hours to update their companies' payment systems to handle crypto in a short notice.

But non-IT people where like very excited about it. It was surprisingly popular with 60+ and older acquaintances. I'm sure it was heavily promoted on local TV and probably by social media posts targeting their demographics.

I've met some 75+ retirees that even wanted to buy ant miners and others that bought crypto with their pensions. Like their first online purchase with card was for buying coins really. And I'm talking about the demographics that actually goes every month to the bank to update their bank booklet after their pension is deposited.

Very weird days.

[josephg]

I've ridden a few boom and bust cycles in crypto at this point.

One pattern I've noticed is that when you start hearing regular people on the street talking about buying crypto currency, its time to sell.

[kristjansson]

Shoeshine boys giving out stocks tips are the ruin of all great bull runs.

[JumpCrisscross]

> Seems like something that should be so easy to catch in a simulation/test

The system has two stable equilibria, one and zero. If we assume zero is virtually permanent, the system always ends there. But it's highly sensitive to a volatility assumption. Ex ante there is no way to precisely estimate that.

[Tao332]

I get the feeling even that wouldn't have made much difference

https://twitter.com/stablekwon/status/1464897977793728514?re...

[TechBro8615]

That’s what this is. Nobody knows how, or if, any of this works. If people are putting money into it they can’t afford to lose, that’s their problem.

[phphphphp]

I disagree. Anchor was a dumb-money trap that increased the fallout of UST’s collapse by at least an order of magnitude, sure, but it was not responsible for the collapse: algorithmic stable coins are fundamentally flawed, they’re a perpetual motion machine, they’re based on the belief that you can artificially create and sustain a market within a set of very narrow parameters. Luna was doomed to fail regardless of Anchor — and the founders knew that, because they’d failed in the same way before, anchor was their attempt to shore up the ruse (which worked… until it didn’t).

[MuffinFlavored]

> Anchor was a dumb-money trap

What % of current crypto climate/culture is "dumb money trap"?

[phphphphp]

Most of it, but it’s not unique to the cryptocurrency industry, only the degree to which crypto is dumb money — many aspects of traditional finance are going down the same path, where it’s no longer professional against professional, but professional shooting a bucket full of laymen.

[vmception]

99%

I think the difference is what to do about that

Users should be more discerning, that’s my vote

[px43]

DAI works fine. Over-collateralization with a looming threat of liquidation has proven to be resilient against all imaginable forms of market turbulence. It still leans heavily on oracles that are basically run by humans, so there is risk, but the incentives seem to be about right to keep enough people honest.

UST was an obvious bad idea. This exact scenario was warned about over and over again, and Do Kwon did what he could to try to delay the inevitable, but here we are.

[phphphphp]

If the algorithm is “give us assets more valuable than the claim we give you in return” is it an algorithmic stable coin, or just an asset with an arbitrary price? I’m unsure I understand how DAI can be characterised as an algorithmic stable coin, given “algorithmic” is usually understood to mean “the thing from which it derives it’s value is an algorithm” whereas for DAI, like Tether (ostensibly), the value is derived from its collateral. I’m open to an argument to why it is a algorithmic stable coin though, perhaps my definition is too narrow? To my understanding, the point of a stable coin is stability of value for the holder, which isn’t true of DAI — because they’re still exposed to the volatility of their collateral.

[px43]

People who actually hold DAI don't experience liquidation risks. It's the people who create the Collateralized Loan Deposits (CDPs) who are taking that risk, and in exchange, are given DAI.

Also, technically the value of UST was fully backed by LUNA, until the total market cap of LUNA dipped below the market cap of UST, and then it wasn't.

IMO it's not the collateralization that makes something an algorithmic stablecoin, it's the currency is managed entirely by on-chain smart contracts, as opposed to Tether or USDC etc, which are managed manually by humans working in a room somewhere. Who knows though, I just made that definition up. The algorithm that manages DAI is pretty rad, so it seems unfair to deny them the status of "algorithmic", but that's fine.

[audleman]

> Also, technically the value of UST was fully backed by LUNA

And what was LUNA backed by? Unless somebody gives a better definition I'll say it was backed by two things. One was UST successfully working as a stablecoin. The other was the vague promises made by the charismatic founder that they were enabling a whole ecosystem with hundreds of developers, creating blah blah. In other words, people bought into the hype that Terra was to become the next big blockchain startup and holding LUNA was as good as owning early stock.

So that goes back to the first point, and as soon as Terra wasn't able to run an effective stablecoin what was left? "We're still here making noise" says Do Kwon. Face palm.

[Fargoan]

There are a lot of dapps on Terra. I'm guessing they'll all move to their own Tendermint chains or Juno

[55555]

Technically, sure. But now remember that market cap is simply last trade price * outstanding tokens and realize that if there’s not enough buyer demand to liquidate all tokens at that price or greater then UST wasn’t backed at that moment in any real sense. The true theoretical backing is buyer demand * buyer’s best bid for the sum of all buyers.

[ProjectArcturis]

One dollar of DAI is backed by $1.50 of a basket of crypto. What happens when that basket drops more than 33.3%?

[charcircuit]

If a vault has a liquidation ratio of 150% and the value of the vault drops below that then the crypto in the vault is auctioned off for DAI.

If it's not able to raise enough DAI to make up for the DAI that was issued when creating the vault the protocol can mint and sell MKR for DAI to get rid of that debt.

[kristjansson]

And who's going to buy MKR?

[charcircuit]

There's hundreds of millions of dollars worth of volume trading MKR. If you can get MKR at a discount you can just sell it on the market.

[wtracy]

That's why DAI isn't really an algorithmic stable coin. Overcollateralized stable coins really belong in their own category.

That's not too say that DAI is safe, but any sort of algorithmic stable coin faces massive systemic risks unique to their design.

[brtkdotse]

> people like decentralization

I submit that very few people give a shit about decentralization. They want their currency to work and not drop 90% of its value overnight.

[cam0]

Agreed. People like the idea of getting rich. And if they happen to get rich, they like the idea of getting richer.

[spopejoy]

Right, but that's like saying "people don't care about how a company treats workers", "people don't care about privacy" -- speculators/shareholders occupy the most craven beliefs in all markets generally.

Decentralization is an _ethos_ that _can be found_ in crypto but you have to look for it, which means you care about it. If you don't care, fine, don't bother with it.

Meanwhile decentralization is actually an excellent measurement of the honesty of a given crypto scheme, and surprise surprise, only a few projects are willing to put decentralization before profits. But they do exist, just like there are a few tech cos that aren't extractive ad machines (not many).

[kova12]

Problem with regulation is that the term itself is ambiguous. What does it mean, who will do it, what are their limits and on what authority do they act?

I am not familiar with this specific crypto, but when I am offered to join a business that pays 19% in USD, I immediately know it can not possibly not fail. Regulations or not, you can't offer such dividend without major risk.

Then the subject of jurisdiction comes up. Who issued the tokens? If it is a USA company or person, it is one story, and they are likely already subject to some regulations. If it is an international community like ethereum, its another story. USA can probably ban it's distribution in USA, but not regulate.

Then, why would the regulations have to be mandatory? Organizations can voluntarily approach SEC and ask them to regulate the crypto they are about to issue. Submit to the appropriate rules and be an investor grade business. No new laws are really required for that. This way you could have trusted that if Luna fails, someone would go to jail. It's just wouldn't have been 19% dividend in that case, following rules comes with a price tag.

[codedokode]

But do you need regulation? Promises to produce 20% interest out of thin air when banks don't offer a single percent are suspicious enough in my opinion.

It doesn't make sense. If there was a risk-free way to get 20% income without doing anything then people could stop working and live just from their savings.

[vmception]

Agreed! I would emphasize (and empathize) that there are muuuuuuch higher yields in the crypto space, so, many of these people felt like they were being responsible in a happy medium.

(The other yields are not fixed yields and very temporary, but it is easy to come out ahead. Not for passive investment chasers)

Regardless it was still up to them to be more discerning. Amazing it got that big.

[Fargoan]

About half the yield came from loan interest. The other half was from Luna Foundation dumping money into Anchor's reserves. Users could take out collateralized loans against Luna, Ethereum, Atom, Avalanche, and Solana tokens. Anchor isn't what brought down Terra. The way UST was supposed to remain stable is what killed the whole thing when it backfired.

[codedokode]

As I understood, people were taking loans at exorbitant interest rates because they could invest it back into some Anchor or UST related tokens and get even more profit. This is not a realistic business model.

Here is a quote from a random article:

> A popular strategy that many have used was the recursive lending strategy, this involved:

> 1. Depositing bLUNA as collateral on Anchor

> 2. Borrowing UST to buy LUNA

> 3. Swap LUNA to bLUNA to repeat Step 1

Regarding this:

> The other half was from Luna Foundation dumping money into Anchor's reserves.

And where do those money ultimately come from? I guess from unlucky investors who bought UST/LUNA and were left with useless tokens? Then this resembles a classic Ponzi scheme.

[1] https://medium.com/qi-capital/lessons-learned-from-the-may-c...

[charcircuit]

>This is not a realistic business model.

I disagree. The return could just go down over time. The idea of taking more loans out against your borrowed loans is risky and can be done with any other loan provider. It opens you to a lot of risk because if you get liquidated you will lose a lot.

>Then this resembles a classic Ponzi scheme.

Anchor itself isn't. A Ponzi scheme is where earlier investors are paid with the funds of later investors and is not sustainable without growth. In a Ponzi scheme not everyone can withdraw their money as the total reported money is more than the money deposited. In the case of Anchor the money reported is actually the same as the money deposited. Everyone is able to withdraw their UST. You can still get 18.06% APY on UST from Anchor right now.

The actual issue is with how Terra handles contraction of UST demand. As it turns Luna into an inflationary currency instead of a deflationary one. This is the danger of having your collaterals value tied to the growth of the stable coin itself. I can see how Terra + Anchor can look like a Ponzi scheme, but I personally don't see it that way. I personally just see another algorithmic stable coin deegging as demand for it goes down. The same thing would have eventually happened even without anchor. People could just get board of the project and decided to pull out.

[wtracy]

Also those borrowers were receiving inflationary Anchor tokens as rewards. When the value of Anchor would spike high enough, borrowers were actually paying negative interest rates.

Anchor did a decent job of obfuscating the nature of the scheme.

[Fargoan]

That strategy is just a round about way to increase leverage. The interest rates weren't even that crazy. I think the highest I ever saw was 10%. Most of it gets offset from staking rewards from the collateral assets. I thought it was a good way to take out collateralized loans against staked ETH. It was great until it wasn't.

[lampshades]

> But without some form of regulation, either by the broader crypto community itself or by governments around the world, situations like the LUNA/UST collapse will keep happening.

I don't understand what's so bad about just letting it happen continuously. People will eventually learn and if they don't, they don't.

[mensetmanusman]

Because the winners gain real money/weapons/power to continue to exploit the low-IQ and eventually the crime spills over into your safe world.

[ProjectArcturis]

Same reason we have laws against Ponzi schemes and other fraud.

[tptacek]

As I understand it: Anchor is the only reason UST was (briefly) workable to begin with. There wasn't enough liquidity in the UST/Luna system to support a major stablecoin; Anchor was how they dragged that liquidity in: by giving people free money to help prop it up.

Take away Anchor and you're stuck at Step 7 of Matt Levine's Algorithmic Stablecoin Analysis: "if you do a good enough job marketing Luna, its price will not be zero. If the price is not zero, you're in business". As I understand it, the price of this whole ecosystem was driven largely by Anchor.

https://twitter.com/tqbf/status/1524509243587575808/photo/1

[javert]

> But without some form of regulation, either by the broader crypto community itself or by governments around the world, situations like the LUNA/UST collapse will keep happening.

That's right, and that's OK. Capitalism allows failure. A free society allows failure. This is healthy. Luna was not a systemic risk. There has been no contagion. Nobody has lost their house or their job. When you don't allow failure, you get Lehman and the 2008 crisis. You get the Fed inflating assets to the point that it's actually dangerous, because they won't let the stock market fall.

[jeromegv]

People definitely have lost their house on crypto.

There’s allowing failure and there’s allowing Ponzi schemes. I don’t want to live in a society where those are legal.

[javert]

Then you don't want to live in a society, you want to live in the Soviet Union. Because nobody can tell you if a business is a fraud in advance. There is no all-seeing, all-knowing bureaucrat that can detect Ponzi schemes. You just have to take away people's permission to do stuff. Which is ridiculous---Ponzi schemes only hurt people who are reckless. Why punish and impede all of us for the sake of reckless people? Your viewpoint is literally uncivilized.

[I-M-S]

The totality of my life experience has shown me it's precisely the other way around - it's the reckless people who punish and impede us all for their sake. Just like there are no all-knowing bureaucrats who can detect Ponzi schemes, there are none who can detect systemic risks either.

[_Algernon_]

Whatever the soviet union was or wasn't, it was still a society. Making such exaggerated claims makes it difficult to lend the rest of your arguments any credence.

[javert]

It was a slave pen. But that's neither here nor there.

[_Algernon_]

Lot's of societies through history had slavery. Doesn't stop them being societies.

[stale2002]

> Then you don't want to live in a society, you want to live in the Soviet Union

I think that there are many in between states for society, between the two extreme points of "allow ponzi scheme investments to be sold to uninformed investors" and "Communist dictatorship".

Instead of either of those two extremes, we could choose a middle ground of "Liberal democracy, where people running ponzi schemes are prosecuted".

> There is no all-seeing, all-knowing bureaucrat that can detect Ponzi schemes

Nothing is perfect. But it doesn't have to be. The government can just go after the obvious ones, at a very minimum, and get it mostly correct.