With volumes that surpass spot trading, perpetual futures are among the most traded commodities in crypto. In comparison to futures, these instruments have higher parity with market prices and enable capital efficiency via leverage. However, the majority of the perp volume is concentrated on CEXs. Despite a slew of designs aimed to bring perpetual to DeFi, most remain unpopular. dYdX is an oddity, however, they employ a type of cheat code in the form of a centralized order book and pairing engine instead of an on-chain one.

GMX, an Oracle-based spot and perps DEX, changed the game by introducing perpetuals' simplicity to leveraged trading. GMX has recently discovered a great product-market fit with both traders and liquidity providers.

Because of GMX's popularity, rivals have attempted to recreate it using designs inspired by GMX. Mycelium (formerly TracerDAO) and MUX (formerly MCDEX) are two protocols that are quite similar to GMX. Mycelium is a modified straight fork of the GMX codebase, but MUX is not a fork; it uses an updated codebase from a prior iteration of its product. Before we can compare different protocols, we must first grasp how the Oracle DEX model works as well as why protocols are gravitating toward it.

The oracle-based DEX

GMX's success is due to its novel liquidity pool index, which is made up of 50% stablecoins and 50% risk assets (ETH, BTC, and altcoins). Each token has a target weight assigned to it. The pool provides liquidity for all spot and leveraged trades. Oracle pricing feeds are then utilized to calculate trade execution prices. Each protocol manages its oracles in a unique way. In the next part, we'll go through the specific distinctions.

Spot swaps on these platforms function similarly to an RFQ exchange between LP assets. The price is determined by the oracle, and spot swaps operate with no slippage - independent of trade size. Swap costs are constantly adjusted depending on how the trade will influence the asset composition of the liquidity pool.

Similar to this, leveraged trading uses oracle price feeds to execute prices and charges an additional open/close fee dependent on position size. The liquidity pool lends a trader a certain amount of an asset in order for them to reach the necessary degree of leverage. Traders borrow the underlying asset from LPs when they leverage up to the long side. A trader borrows stablecoins from LPs when they are leveraging short. For the asset borrowed, the trader gives LPs an hourly borrowing fee. Therefore, in all positions involving leveraged trading, the counterparty is the cumulative liquidity pool. As a result, LPs profit when a trader loses, and vice versa.

A comparison among protocols

Supported Chains

Mycelium is exclusive to Arbitrum. Although GMX is available on Arbitrum and Avalanche, each deployment's liquidity is limited to that chain. On the other side, MUX is a multichain protocol that is supported by Fantom, BSC, Avalanche, and Arbitrum. Through a bot that tracks liquidity across all active networks, MUX aggregates its liquidity and offers "Universal Liquidity." This addresses the problem of dispersed liquidity across chains by enabling MUX traders to access cumulative protocol liquidity.

The "broker module" of MUX is the source of this invention. The bot that keeps track of liquidity across all of MUX's supported networks is called the broker module. The broker module will search for available liquidity on each chain when a trade request is made on-chain. The broker module will fill the transaction using its oracle price if the amount of liquidity available exceeds the amount needed for the size of the contract. If a user wishes to withdraw 15 ETH from a pool but there are only 10 ETH available on Arbitrum, the user will receive 10 ETH + 5 muxETH. To exchange 5 muxETH for 5 ETH, the muxETH can be bridged to another chain with liquidity, such as Fantom.

Oracles

In order for this architecture to function as intended, a strong oracle and pricing structure is crucial. Each protocol has a unique method for including price feeds and for preventing harmful arbitrage.

GMX: A customized oracle that enables quick price updates by using CEX pricing. To prevent front-running, the oracle updates at the beginning of the block, and all transactions in the block receive a single price based on that oracle update.
Mycelium: Binance and FTX median prices are utilized as their own price keepers, which are used in conjunction with external oracles. Additionally, Mycelium hired a proprietary trading company to derive value from latent oracle feeds and reward LPs with 80% of the earnings.
MUX: Using Binance and FTX prices, the protocol's "Dark Oracle" oracle module performs high-frequency price updates at a rate of about 100 ms for each update. The harmful arbitrage caused by unequal pricing data and delayed oracle updates is lessened because of the high-frequency updates.

Max leverage and Fee

On GMX and Mycelium, the maximum leverage that one may take is 30x. MUX offers up to 100x leverage, which is probably a result of its capacity to disseminate price changes more quickly and prevent prices from becoming stale.

A 0.1% of the notional size opening and closing fee is charged by GMX and MUX. Mycelium charges a somewhat lower cost of 0.09% in an effort to compete on fees. However, keep in mind that Mycelium's long-tail assets have higher spreads than those of GMX or MUX, which could lead to inferior execution pricing.

Spot swaps

While MUX does not support spot swaps, GMX and Mycelium do. Zero slippage swaps are available via GMX and Mycelium for a 0.2% charge. However, if the overall cost (trade charge plus spread) is taken into account, switching an asset on Mycelium is probably going to cost more than doing so on GMX. Recently, Mycelium raised its basic price from 0.15% to 0.2%.

It's interesting that MUX has incorporated "Liquidity Multiplexing," which enables the LP token to be used for yield farming and margin trading on DEXs at the same time. To guarantee that the pooled assets can completely cover traders' gains, the overlap limit has been set at 80%. This reduces the ability to support spot swaps, but it enables MUX LPs to use their inactive assets to produce more yield.

Value distribution

The most generous platforms are GMX and Mycelium, which provide LPs 70% of fees received. With 30% and 50% of fees earned, respectively, GMX and MUX provide the best terms for token holders or stakers. About 15% of the revenues are used by Mycelium to benefit the token.

The restriction for MUX is that in order to obtain veMUX and receive a percentage of protocol fees, MCB, the MUX token, must be locked for two weeks to four years. Stakers have the freedom to unstake at any time because GMX and MYC do not enforce any lockups on principals.

While MUX and GMX share a high-level design, they also set themselves apart. Instead of facilitating spot swaps, MUX is totally committed to giving traders access to a leveraged trading environment. Searching for yields that eventually benefit LPs is done with resources that could otherwise be employed for spot swaps (but aren't).

The similarity between MUX and MCDEX v3 suggests that it is more of natural evolution. However, Mycelium, a direct fork with minor modifications, is not all that dissimilar from GMX. Given that they both target the same characteristics and employ comparable risk limitations, the two have a significantly closer overlap than MUX. As the current leader, GMX has a clear advantage. Mycelium also appears to be attempting to compete by reducing its prices.

However, has this tactic been effective, or is GMX still clearly in the lead? Let’s take a look at the current landscape.

Current landscape

GMX Current landscape

Source: Dune analytics

GMX has performed successfully in recent months, with volumes reaching a high of $818m on July 29, 2022. Mycelium and MUX reportedly acquired tremendous traction, garnering 18% and 19% of volume share, respectively. Mycelium volumes peaked at $60M on September 15, 2022, while MUX volumes peaked at $86M on August 26, 2022. However, their rise did not last, and both protocols presently fluctuate between $1-2M in daily volumes, with occasional increases. This allowed GMX to reclaim the lead over the duo.

Token incentives are used by most protocols to jumpstart their liquidity and users. This has proved unsustainable, as the tokens are often sold upon farming, and mercenary liquidity relocates to better pastures as it finds more attractive locations to produce farms. GMX has demonstrated that its methodology can create significant actual dividends for token holders and LPs. These fees are created naturally through spot swaps and leveraged trading fees. As a result, seeking to repeat this achievement is desirable.

LPs - what makes GMX different

Since they serve as the liquidity source, LPs are crucial to this DEX architecture. LPs get reimbursement in the form of fees for the use of their money. In the beginning, MLP and MUXLP produced higher yields than GLP, with MLP's yields rising as high as 160% annually.

GMX outperforms

Source: Delphi digital

However, this impact vanished as volumes decreased. Now, LPs only earn 50% of protocol fees for MUXLP, compared to 70% for GLP and MLP, which puts it constantly at the lower end of yields. GLP's yield has been far more reliable, averaging between 20 and 40% annually.

Liquidity usage is the appropriate statistic to assess the consistency of returns because the majority of the protocols' fees come from leveraged trading. Utilization of MLP and MUXLP has been erratic, resulting in variable yields produced for LPs. Furthermore, from a peak of about 35% to a range of 5-20%, their use has fallen.

GLP use on Arbitrum and Avalanche has been more steady, leading to higher LP yields. GLP (Arbitrum) has continuously remained between 20 and 40%, whereas GLP (Avalanche) has remained between 10 and 20%. GLP's yields are more constant and higher, which is more appealing to LPs because of their consistent use.

Although leverage is a helpful trading technique, it is also quite risky because you are assuming greater risk. Notably, regardless of the protocol they employ, traders on these platforms typically lose money over the long term. Since LPs are the counterparty and move in the opposite direction from traders, this is advantageous to them. MLP has a significant advantage thanks to its 6% profit margin since August 2022 per dollar of AUM. A pricing manipulation exploits caused GLP (Avalanche) to lose the majority of its gains after it first turned a profit.

As realized losses from traders accumulate back to them, LPs' profits are reflected in the rising value of their tokens.

Lower-risk LPs are expected to work a lot better than higher-risk LPs in a weak market. GLP (Arbitrum) lost the least, as was predicted. Being that the competition between traders and LPs is zero-sum, it should be noted that trader P&L also influences pool performance.

It's interesting to note that up until recently, GLP (Arbitrum) closely followed the index's performance. The exchange's losing traders are most likely to be blamed for this outperformance.

Comparative valuations

GMX price

Source: token terminal

To keep things in context, even if GMX has a greater value when we consider the earnings it produces, it still has a superior fundamental position over its rivals. The initial traction of Mycelium and MUX propelled them to multiples that were equivalent to GMX, with MYC peaking at 10x P/E on September 11, 2022, and MCB peaking at 16x P/E on August 23, 2022.

As was already noted, revenue to token holders drastically fell once Mycelium and MUX lost their volume market share. In comparison to MYC's 61x P/E and MCB's 226x P/E, GMX now offers a 19x P/E.

GMX has made more money than it has distributed in token payouts. This shows that the revenue is sustainable and organic and that they are not covering the full cost of their output. GMX is a perfect illustration of how tokens with strong foundations may better weather storms while tokens have been falling throughout crypto, regardless of the product's traction. In spite of the market slump, GMX did better than other tokens in 2022.

Risk and downsides of exposure rental models

The exposure renting approach has been successful in enabling leveraged trading and spot swaps. It offers traders a seamless user experience, gives LPs excellent profits, and even results in value accumulation for token holders.

We've covered a lot of territories so far, but the majority of it relates to the operation of this model and its effects on traders and LPs. But this paradigm has certain drawbacks as well. Let's examine the key risk associated with this DEX design.

Utilization-Based Funding

The GMX borrowing rate operates similarly to a money market. Long-position traders that use leverage pay a borrowing charge on the surplus assets needed to reach the desired degree of leverage.

On CEXs, financing rates are determined by subtracting the perpetual contract price from the current price of the underlying asset.

According to this DEX model, funding rates grow linearly according to asset usage (loan assets divided by total assets in the LP). This indicates that the leveraged trade open interest cap is placed at the pool's maximum asset usage. As a result, unlike with perps, the system's need for an increase in collateral is constrained.

The borrowing rate that results from this can be very different from CEXs. For instance, during the Ethereum Merge, funding rates for ETH on FTX fell below -0.04% hourly, while the fee to short on GMX was -0.006% hourly. However, traders frequently pay higher rates on GMX compared to CEXs under typical circumstances. The balanced funding method used by perpetual also allows longs to compensate shorts or vice versa. This approach requires all traders, whether long or short, to fund LPs.

Limited Tradeable Assets

The assets that can be traded in each protocol depend on the liquidity pool's asset composition. For instance, the only trading pairs for GLP on Arbitrum are ETH, BTC, UNI, and LINK. Since this liquidity pool is an index, protocols can only add a certain number of assets before debilitating OI limitations are introduced for each asset. Remember that the maximum usage rate and the quantity of each asset in the pool determine the OI limit.

A design issue exists in how these pools were built as an index. DEXs utilizing this concept must create many indexes to meet demand in order to reach a bigger TAM. However, because LPs hold a variety of connected crypto assets (such as BTC, ETH, and others) with a significant allocation to cash, it only slightly benefits them. As a result, the index is kept somewhat stable and LPs are able to generate fees from a wide range of assets.

Oracle Manipulation, Front-Running, and Arbitrage

The key to this DEX architecture is oracle price feeds, therefore there is always a chance of oracle mispricing (unlikely) or manipulating. Opportunistic arbitrageurs running ahead of oracle updates are a possibility during tumultuous times. For instance, ETH on CEXs increases from $1,400 to $1,600 in just 10 seconds. ETH remains at $1,400 since the protocol's oracle price feed has not been refreshed to reflect the price change. Arbitrageurs can purchase ETH on DEXs for $1,400 and sell it on CEXs for $1,600, earning a profit of $200. This $200 profit is the result of an arbitrage that the LPs must pay for.

By utilizing CEX price feeds and delivering price changes at the beginning of each block before deals are conducted, GMX in particular has discovered a solution to this problem. By doing this, you can be guaranteed that each trade in a block will have the same updated execution price. However, in a low liquidity situation, when altering an asset's price is simple and inexpensive, it does not safeguard the integrity of the price feeds.

Due to limited liquidity on CEXs, GMX on Avalanche was exploited on September 18, 2022. A clever trader was able to profit from AVAX's low liquidity. This allowed the exploiter to profit by almost $565,000.

GLP on Avalanche was the losing party since the LP was used to get the exploiter's profit. The current remedy used by GMX is to limit OI for AVAX. By dynamically changing OI restrictions based on the CEX's liquidity circumstances, this problem may be resolved.

In response, Mycelium decreased the OI limit and raised spreads on its long-tail assets.

Impossible Price discovery

On Oracle-based DEXs, price discovery is not feasible because transactions are executed based on outside pricing data. Because of this, no DEX utilizing this strategy will ever be the biggest or most liquid exchange, which limits their capacity to develop from the start.

Price discovery is a crucial component of every market, whether it is financial or not. While this is acceptable for liquid assets like ETH, this methodology may provide better trade execution for illiquid long-tail assets as it does not impact prices on major marketplaces. Spread and slippage variables depending on demand and liquidity might assist to address the issue with illiquid assets.

Smart contract exploit

The possibility of smart contract exploitation grows when LP assets are used to provide yields on other DEXs. It's vital to remember that the increased yield for LPs carries its own risks because investing user money in other protocols could have undesirable consequences.

These dangers might be considered as a compromise for a bigger goal of offering an on-chain leveraged trading experience similar to CEX. The danger of oracle price manipulation, which has impacted GMX and Mango Markets lately, is the most obvious one in this situation. To prevent price manipulation that can harm LPs, there has to be more effective risk management for illiquidity assets.

Conclusion

Success for GMX hasn't gone unnoticed. Despite the competition, the exchange has been able to maintain its strong lead while new protocols struggle to gain acceptance. Volume and LP yields are directly related in this model. As a result, protocols that struggle to draw substantial, steady volumes will eventually have their LPs leave for better places. In other words, Mycelium and MUX need to continually raise their volume levels to match those of GMX.

The modifications made by Mycelium represent its most recent effort to gain market share. While competing on fees helped Uniswap improve stablecoin volume relative to Curve, luring consumers away from the incumbent is a difficult undertaking in and of itself. Would volume abruptly shift away from Uniswap if SushiSwap dropped their swap fees by half? Most likely not. It's not simple to unseat the dominant protocol in a given industry, and a slight reduction in fees might not be sufficient.

It's not like GMX doesn't have a few tricks under its sleeve, after all. GMX intends to roll out more features in the future, including X4 and a synthetics PvP AMM. With these planned releases, Mycelium and MUX may struggle to catch up, especially because they are essentially competing on the same chains (EVM networks). However, as previously said, it is far too early to declare victors and losers. However, if Mycelium and MUX do not devise a targeted plan to bridge the volume disparity, the tides appear to be turning in GMX's favor. GMX's glory is likely to last for a very long time.

Disclaimer:
This article is intended for and only to be used for reference purposes only. No such information provided constitutes advice or a recommendation that any investment or trading strategy is suitable for any specific name.

GMX evaluation: Will its success last?