Perpetual DEX Q&A: Background knowledges and supplemental explanations
Background knowledges and supplemental explanations
We have covered everything from the significance of a perpetual futures contract and the existence of a perpetual DEX in three parts, to details on the current Perp DEX industry and individual projects.
Although it is a relatively minor and difficult topic in this market, we thank our readers for their overflowing interest and attention.
As a result of this interest, we have received some questions related to the content of the report. Based on the Q&A we have received, we are sharing some information that has been omitted or that can be easily taken as basic background knowledge.
Q1.
- Do both long and short positions require a counterparty in futures trading? What happens if there isn’t one? For example, when I open a long position, do I need someone to take the opposite position (short)?
- Do exchanges usually only handle settlement of trades? Don’t they become counterparty themselves?
- It is said that GMX is not a perpetual exchange but rather margin exchange. Then, does a trader actually borrow the real spot assets from LP?
A1.
Firstly, it should be noted that they depend on the price discovery model. Currently, there are three main models: AMM, Orderbook, and Oracle.
A counterparty is always required for both long and short positions to be executed. It’s important to note who the counterparty is, but in the case of an orderbook, the exchange provides only the orderbook and acts as a broker, opening a marketplace where traders and market makers can meet.
In some cases, exchanges may conduct market making to directly execute trades to provide liquidity.
However, with some exceptions, some centralized crypto exchanges adopt a fully dealer model where they take the opposite position of all traders. It is important to be cautious of centralized exchanges’ dealer model because it can prevent users from withdrawing their profits.
AMM platforms directly take the opposite position of all traders, but this risk can be mitigated if there are external liquidity providers. Perpetual Protocol successfully transitioned to a brokerage model by driving almost all liquidity with external LPs, and Drift Protocol mitigates house risk by relying on some external LPs while also participating directly in JIT auctions.
In the Oracle model, if the platform’s treasury is not part of the LP pool, it currently follows a fully brokerage model where liquidity providers act as market makers and compete with traders. Specifically, GMX uses a leverage model that borrows actual liquidity, resulting in margin trades, not futures contracts. In contrast, GNS and Uniwhale realize futures contracts that exchanges only the P&L instead of borrowing actual liquidity.
Q2.
- How is the futures price determined? There is a theoretical price, but why doesn’t the actual price of futures follow the theoretical price?
- Can we understand long in futures as buying demand and short as selling pressure? And is the futures price determined by the correlation between the two?
A2.
The futures price is determined by the market. If the trading volume of futures is greater than that of the spot market, the market price follows the futures price. As futures have leverage and are in the form of “contracts,” they are capital-efficient, so the futures price always forms the market price. As all market participants do not refuse to execute a trade unless it is the theoretical price, the market price can always deviate from the theoretical price depending on supply and demand.
In the futures market, if there is a lot of buying demand, the futures price naturally rises. In the case of the orderbook, if the market depth is low, the price moves exponentially, but as the depth increases, the movement of the price slows down.
In the case of an AMM, as the trade is executed according to x * y = k, the price moves very intuitively along the curve according to the trade executed. Therefore, various attempts are being made to adjust the k value so that the AMM futures exchange can reflect the actual market depth.
As the oracle exchange borrows external prices, the price does not move no matter how many trades are executed on the exchange.
Q3.
- What does “Entry price” mean for perpetual futures?
A3.
The entry price in the orderbook refers to the price at which a trade is matched with the counterparty. If a limit order is executed, the price at which it is executed becomes the entry price, while for a market order, the average price becomes the entry price. In the case of AMM, market orders are typically used and the price moves along a curve, so the difference between the AMM mark price and the entry price widens as the order size increases. Ultimately, this is referred to as slippage due to market impact.
Q4.
- Perpetual futures are futures without a set expiration date, but when it comes to position closing, when and with whom is the trade settled?
A4.
Unless it is a physical delivery contract, in most cases where cash settlement contracts are supported by exchanges, closing a position in perpetual futures is nothing more than entering a new position in the opposite direction. If one wants to close a 50 ETH long position, they enter a 50 ETH short position, regardless of the order book, AMM, or oracle. The counterparty then enters a long position of 50 ETH.
Q5.
- What is the concept of funding rates in perpetual futures, and why is it necessary to match the futures and spot prices in perpetual futures contracts?
A5.
The concept of funding rates was introduced in perpetual futures contracts to ensure that the perpetual futures price stays in line with the spot price. This is necessary because perpetual futures contracts have no expiration date, unlike traditional futures contracts that settle in cash or through physical delivery. In cash settlement, the futures price converge to the spot price at expiration, while physical delivery requires an actual exchange of the underlying asset. As a result, the difference between the spot and futures prices approaches zero at expiration.
However, because perpetual futures contracts have no expiration date, there is no mechanism for the difference between the spot and futures prices to converge to zero over time. Therefore, funding rates are in charge of the convergence mechanism driven by market demand, incentivizing traders to maintain the futures price close to the spot price.
Since the mechanism relies on the market principle, there have been instances where the funding rates remained high for extended periods, despite significant trading activity. For example, when the Oasis Network launched a new DeFi protocol and offered high APRs, traders bought the underlying asset, deposited it, and sold perpetual futures contracts, resulting in a large difference between the spot and futures prices of the native token on Binance. Despite the high funding rates, the difference persisted for an extended period.
The theoretical price of a dated futures contract is set higher by the risk-free rate of return. From the seller’s perspective, they could receive cash immediately by trading in the spot market and earn interest on that money by depositing it in a bank. However, by entering into a futures contract, they give up that interest rate, so they must reflect that in the futures price.
The difference between the spot price and the futures price is called the basis, and theoretically, the futures price is always higher. However, the basis can vary depending on market conditions. The reason why the funding rate for perpetual contracts defaults to a positive value, even when the spot and futures prices are equal, is to reflect the risk-free rate of return that is partly incorporated into the theoretical price. Through cash-and-carry arbitrage that exploits the basis, the crypto market can achieve a self-referential risk-free rate of return that could set the default funding rate. However, this assumption requires an efficient market.
Q6.
- Most people seem to use futures to take advantage of high leverage. Why is it possible to have high leverage with futures? If we simply want to engage in leverage trading, couldn’t we just engage in a margin trading instead of a futures contract?
A6.
While both futures contracts and margin tradings are used to achieve leverage, they differ in the way leverage is realized. Futures contracts have evolved to settle large P&L with minimal collateral by entering into a contract to buy or sell a certain number of assets in the future. In contrast, margin tradings use borrowing & lending to realize leverage.
For a long position in dated futures, you would pay the cost of entering into a contract at an agreed-upon theoretical price that is higher than the current market price, pay funding fees for perpetual futures, or pay borrowing interest for margin tradings.
Q7.
- If GMX uses a model of margin trading, there seems to be a clear limit to the hard cap set by the liquidity provided by LP. Why has it received such a great response in the current market? Could it be because GMX was the first to implement an Oracle-based perpetual DEX?
- How is it possible to exchange only PnL as in the case of GNS and Uniwhale? Is the PnL model superior to the model of borrowing liquidity like GMX? Even the PnL model is ultimately based on the stable coin liquidity provided by LP, so doesn’t it also have a hard cap on the supplied liquidity like GMX?
A7.
The reason for the popularity of GMX can be attributed to several factors, such as the first-mover advantage of the Oracle model, strong initial community cohesion, and acceleration of layer 2 narratives and the advent of Arbitrum Odyssey since 2022. As for the current situation, it can be broken down as follows:
- From an investor’s perspective, the value acquisition model of GMX is clear as it has a low risk of future equity dilution and a simple and transparent revenue structure and value accrual model that distributes all platform profits.
- From a user (taker) perspective, GMX is currently the only platform with literally zero market impact.
- From a liquidity provider’s perspective, the GMX model offers an index like position in $GLP, market-making P&L, and a distribution of 70% of trading fees.
In the GNS model, traders and liquidity providers exchange only PNL, allowing for executions beyond the nominal amount of liquidity actually deposited. Although capital efficiency is high, there is a risk of payment default when there is a flaw in the system.
GMX mitigates this risk by using only actual liquidity and putting pressure on makers towards zero slippage. However, GMX was able to quickly resolve the chicken-and-egg problem and enabled the platform to reach its current scale. Therefore, GMX is a model for retail investors, while GNS imposes some market impact to compensate for the risk and can adopt a strategy to onboard whale traders.
Q8.
- In the dYdX app chain narrative, isn’t it true that the validator network managing the order book off-chain ultimately results in the same off-chain orderbook as before, and simply transfers the intermediary risk to the validator network?
A8.
dYdX is planning to transition from a CLOB where one entity manages the orderbook to a decentralized and permissionless validator network managing the order book. This is in line with the core purpose of blockchain networks that encourage correct choices based on economic incentives.
Even in like Ethereum and Aptos networks, while nodes are decentralized, individuals manage servers off-chain, much like AWS. And it seems that dYdX used the term “decentralized off-chain order book” in line with the circumstances of other blockchain networks.