Virtual AMM protocols disagree on the price of risks they internalize. SYMMIO AMFQ tech can help them out.
Why do protocols make assumptions about financial markets and internalize risk when they can hedge them out using SYMMIO ‘s AMFQ tech?
Abstract
Virtual AMM protocols can be defined as protocols that offer to trade at market (oracle) price without having deep liquidity, which also means risks are internalized or pushed back to the traders themselves.
In his trade with GMX, Andrew Kang’s $18 million profit was almost entirely eaten up by a whopping $16 million fee. This raises questions about the trader’s ability to use virtual AMM protocols for the cost they need to pass on to traders for the use of their shallow liquidity.
In the GMX model, 30% of traders’ costs accrue to GMX protocol token holders, while 70% accrue to investors in the GLP pool in exchange for bearing all traders’ risks. As a result, Andrew Kang’s trade led to a $4.8 million gain for GMX holders, but a $6.8 million loss for collateral providers.
Virtual AMMs internalize risks, which raises questions about the pricing of these risks for all parties involved in the protocol. Have virtual AMMs implemented reliable methods to price risks and ensure that both traders’ costs and counterparties’ rewards are fair and aligned with the market price of these risks?
In practice, vAMM protocols use varying, sometimes even opposite hypotheses for the price of risks they bear. Some assume that the value of traders’ information is very small, while others assume it’s very large.
In theory, it is impossible for any of these models to be always consistent: pricing risks require option prices in addition to spot pricing, but vAMM protocols solely rely on oracle spot prices.
SYMMIO tech permits access to deep financial markets, greatly benefiting on-chain traders, who will no longer need to pay significant costs for their gains, as well as crypto-protocols, who can now hedge their risks and facilitate trading rather than betting against their users.
The SYMMIO AMFQ tech, which unifies on-chain security and transparency to global liquidity, can ultimately be used for universal and safe access to liquidity for investors, and for universal liquidity provision for market makers.
Intro — disagreement on how to replicate a market without deep liquidity shows the need for SYMMIO tech and deep liquidity
Order book technology requires pre-committing liquidity (locked margin) since liquidity cannot be “forked” — that is, locked at the same time in the order books of main exchanges and in order books of a myriad of on-chain protocols. Rather than bridging abundant liquidity to build better products, the crypto industry is investigating other trading technologies to attempt to mimic the deep liquidity found in traditional financial markets. Let’s take a deeper dive into the assumptions that underlie virtual liquidity models.
Virtual AMM is the go-to crypto technology/experiment for trading at the market price (aka “oracle price”) without market-like liquidity to offset trades.
The virtual AMM tech assumes that market mechanisms and consensus can be replicated artificially, without the presence of market participants, their trades, and liquidity.
To the untrained eye, oracle-price trading may seem like a way to magically trade at true market prices with minimal costs for traders and fair rewards for counterparties. However, market behavior is complex and it is unlikely that an entirely separate logic will achieve the same consensus as the market as a whole. And without consensus, it’s also challenging to ensure fairness for both traders and the protocol investors that act as counterparties.
To provide basic proof that virtual AMM protocols do not achieve market-like consensus (and do not price risks market-consistently), we demonstrate the lack of agreement among these protocols on how to price risks. Specifically, there is no consensus on how to pass on costs to traders or on which party should bear the risks: the protocol, traders themselves, or counterparty providers.
Under the hood, the Oracle Price vAMM protocols make opposite assumptions regarding the trader’s information ratio, and traders’ “bets” can result in surprisingly different outcomes.
Oracle price virtual AMM tech contains vastly divergent beliefs on how markets operate, with vastly different outcomes for traders. To simplify things:
- at one end of the spectrum, some protocols, such as GNS, assume traders have no private information. They assume it’s safe to bet against traders for, in ideal conditions at least, a small fee.
- at the other end, GMX x4 vAMM (the “PeP” version) assumes that traders’ private information is potentially valuable, so traders bet between themselves rather than against the protocol. At this end of the spectrum, traders’ odds are defined as with a bookmaker. If all traders are long, they cannot win.
Risk must be priced, or better, externalised, to avoid direct costs to traders that prevent cashing in their gains. The price of risk is given by options prices, not just spot price.
vAMM protocols solely trade on the basis of oracle spot price, with traders’ costs that depend on internal assumptions and their own exposure. Yet spot prices do not provide a forward-looking indication of risks. Without hedging risks or even pricing risks correctly, the best that protocols can do is halt trading altogether (see Redefining on-chain FX trading).
Rather than experimenting ad-hoc models, we believe protocols should externalise risks, or at least adequately price it. Both methods require AMFQs to bridge either liquidity or information from traditional financial markets.
The SYMMIO AMFQ Tech can assist vAMM protocol and either:
- enable them to hedge their risk in deeper traditional financial markets
- enable them to price risks correctly, by AMFQing or RFQing option prices from traditional financial markets
Overall, SYMMIO’s AMFQ technology can assist both traders and protocols to trade at competitive prices, hedge their risk, and align incentives between the different parties in vAMM protocols.
A tale of antagonistic assumptions about traders’ information content
As explained in the GMX x4 proposal, oracle price tech approaches are experimental in nature. This suggests that they were designed to learn about virtual markets rather than being thought of as an optimal solution for on-chain trading.
There have been quite a few different models used for leverage trading (…) each have their own pros and cons and we [GMX] believe it is worth continuing to experiment (…) these different trade-offs.
(quoted from GMX x4 proposal)
The two most opposite assumptions in the space of virtual AMM experimental protocols are whether traders are on average random (so the protocol needs to bet against them) or whether traders’ information is valuable (and traders should bet against each other).
Protocol assumptions 1: random traders (no valuable information)
Virtual liquidity usually means that all trades settle at oracle prices rather than at an internal price, and all trades happen inside a closed system. Risk is thus internalized, and, in the standard vAMM model, borne by the protocol and its investors:
- In the Gains Network, the user-provided DAI vault is the counterparty common to all trades, and GNS holders are the counterparty to the DAI vault
- in GMX v3, all traders trade against the common general liquidity pool (GLP), which is subsidized but separate from GMX
These two set-ups assume, overall, that the value of the trader’s private information is lower than fees.
There is evidence that the average investor, even the average fund manager, does not beat the market after fees over a long period. This in a way is logical because since the market as a whole represents the average investor’s portfolio without fees, the average investor can hardly beat themselves after fees!
Nevertheless, some known investors have consistently outperformed the market over the years. And some strategies perform particularly well in some market conditions, so adjustments are necessary if the protocol users (eg the crypto-crowd) are directionally right at a given moment in time.
The standard crypto virtual liquidity model results in either discouraging fees, or trading being altogether halted when market conditions become risky for the protocol.
This model has raised some concerns, notably the much-advertised Andrew Kang’s position against GMX.
- this single position exhausted part of the GLP, showing that in fact the protocol or its liquidity providers could be the counterparty to a single, possibly non-random trader, rather than the counterparty to a large number of (more) random traders
- this trade also underlined how fees could eat up traders’ profits: Andrew Kang’s directional bet earned him $18m, from which he had to deduct $16m in fees. These fees accrue 70% to the GLP liquidity providers, meaning its investors were at a net loss. At the same time, 30% of the fees accrue to the GMX stakeholders, meaning in this particular case the protocol itself was at a gain
Source: https://twitter.com/alphanonceStaff/status/1628280014876868608
Protocol assumptions 2: informed traders (high information ratio)
The Andrew Kang story illustrates that protocols may not bet against the average non-informed trader at all times.
In its PvP AMM, GMX has proposed a so-called PvP vAMM with radically different assumptions, acknowledging that betting the house against traders may be risky. In the PvP vAMM logic, traders’ odds ratio is determined by the ratio of long to short traders, similar as to what is done with bookmakers.
To simplify things, the pot to share is made of both long and short traders’ deposits, shared between winners.
Shortening their own example:
“Alice (resp. Bob) opens an ETH long position of size $10k (resp $20k), she deposits $1k (resp $2k) USDC as collateral, this mints 1k (resp 2k) GD tokens kept within the pool. The pool now holds a total of 3k USDC and 3k GD tokens worth $1 each. If the price of ETH increases by 5%, Alice makes a profit of $800 vs $500 USDC if she was using a regular trading platform; while Bob has made a loss of $800 vs a loss of $1000
If the price of ETH decreases by 5%, Alice makes a loss of -$575 vs $500, while Bob makes a profit of $500 vs $1000 with a regular platform
This proposal makes it possible for the protocol and its investors to bear no risk against traders.
source: https://medium.com/@gmx.io/x4-protocol-controlled-exchange-c931cd9a1ae9
Reaching fair prices: bridging cheap liquidity vs relying on arbitrageurs
Protocols with thin liquidity sometimes rely on arbitrageurs to bring prices or funding costs to acceptable levels.
An arbitrageur can take advantage of large funding fees on a synthetic asset (such as WBTC), mobilizing their USDC to purchase the asset spot, while simultaneously shorting WBTC on the protocol to profit from the elevated funding fees.
Synthetix, a perpetual protocol (which we will review separately), advertises profitable arbitrage opportunities created by its high funding rates. At the same time, arbitrage is risky and costly, which implies that traders need to expect high funding rates to deploy their capital.
This suggests that arbitrageurs extract value from traders when protocol liquidity is insufficient but cannot bring down funding costs to competitive levels.
Rather, it is more cost-effective to systematically source liquidity from where cheap and abundant to where scarce and expensive, via the SYMMIO AMFQ system.
Overall, using the market and its liquidity seems cheaper than making assumptions about market consensus and relying on arbitrageurs to pocket money from imperfections created.
Is market liquidity available? Yes. Through SYMMIO.
Source: Synthetix advertises that funding rates get large enough to permit profitable arbitrage: https://blog.synthetix.io/synthetix-perps-futures-funding-arbitrage/
Bringing together market liquidity, on-demand AMFQs, blockchain security, and transparency to establish a universal and secure trading experience.
The blockchain was designed for universal secure access to financial services, before SYMMIO it was just missing deep liquidity.
The blockchain has been created by integrating cryptographic security and transparency of the public ledger, and it’s a one-of-a-kind technology that makes the world’s assets fully secure in a completely trustless way. But so far it only hosts a fraction of the world’s trading needs.
By contrast, five centuries of banking crises have made it clear that regulation and government intervention alone cannot guarantee the safety of everyone’s funds and transactions. Yet liquidity remains mostly created and hosted via traditional institutions.
Universal access to secure financial trading results from bridging deep liquidity, security, and transparency.
SYMMIO’s AMFQs are an additional tool that can be used universally
Any trader and protocol can benefit from the AMFQ technology, which designed to offer a reliable and scalable trading experience:
- The primary competitive advantage of virtual AMM protocols is their ability to execute quasi-instantaneous transactions, but they can only achieve an optimum user experience under all circumstances and thrive if they have access to deeper liquidity, which can be effectively provided by the SYMMIO technology.
- Conversely, any front-end utilizing SYMMIO contracts can leverage their innate advantage of providing true market impact for any asset and trade size, along with managed borrowing costs, to also offer “thin liquidity” instant settlement services by using a small percentage of their locked margin for market-making.
Technically:
- Existing vAMM protocols can continue to operate on one-block-ahead oracle prices, provided they use the SYMMIO tech to hedge their risks as soon as their books have significant imbalances.
- Front-ends and market makers lock in capital, some of which can be provided for instant settlement. For instance, a bid-ask “curve” on top of known previous block oracle prices would permit to offer instant execution at a known price, rather than fast execution at next block oracle prices as with standard vAMM tech.
The SYMMIO AMFQ tech is versatile and can ultimately be developed to fit user needs under all market conditions:
- thin orders could be optionally instantly confirmed provided a standard volume-price impact is embedded into instant market making. This means execution time can be quicker than with any vAMM technique.
- standard orders, typically require an AMFQ to validate price conditions, with settlements with delays of a few blocks. AMFQs act as a price quantity aggregator and permit delivering the best trading conditions.
- large orders, i.e. orders that would result in price impact if executed on order books, require liquidity formation, similar to market open or close.
Besides, the alignment of incentives between the SYMMIO protocol and its users also makes it possible to develop value-added services for traders, and it’s not unlikely some front-ends will integrate trading signals, managed funds, and more complex strategies than just leveraged longs.
Overall, SYMMIO’s agnostic, secure AMFQ technology can help both crypto protocols and traditional finance better serve worldwide demand for financial services.
