From dYdX to GMX: What Derivatives DEXs Have and Have Not Achieved

Acknowledgments: Special thanks to,

• The article discusses the state of the DEX derivatives market in the post-FTX era.
• Most DEXs have struggled to attract organic growth as a large chunk of trading volume has been generated by bots and traders who exploit token economics or liquidity providers.
• GMX emerged as the largest positive surprise in the vertical. Its zero-price impact design makes it an attractive venue for whale traders.
• GMX’s product is still far from perfect, with limitations including constrained asset selection, liquidity scaling issues, a lack of adequate protocol-level risk management, a large dependency on centralized inputs, and sub-optimal UX.
• Nevertheless, some of these shortcomings might have been intentional trade-offs that helped GMX find the product-market fit and become the derivatives DEX with the most organic demand.
• Still, the market has proven that to win the crypto derivatives market share, DEXs would have to beat CEXs at their own game.

and for providing valuable feedback and Darko Bosnjak for the help with data analysis!

TL;DR:

Crypto derivatives volumes have been outpacing the spot market, and the trend shows no signs of slowing down. In 2022, the top 10 derivatives centralized exchanges (CEXs) saw a daily average volume of roughly $95 billion, while the top 10 spot CEXs accounted for around $31 billion. As the industry continues to mature, we can expect the ratio of derivatives to spot volume to reach levels similar to those of traditional asset classes.

However, when we turn our attention to decentralized exchanges (DEXs), the story is a bit different. Derivatives volumes were only 56% of the spot market volumes, indicating a relative immaturity in the derivatives DEXs. This is especially apparent when we compare the DEX to CEX ratio within the specific vertical: derivatives DEXs account for a mere 1.5% of CEX volume, while the share is closer to 8% in the spot market.

It’s important to note that just because derivatives generate much larger notional volume in traditional finance and centralized crypto exchanges, it does not necessarily mean we will see the same trend among decentralized players. In fact, there is only one crypto derivatives market, and it’s possible that it will continue to outgrow the crypto spot market without any contribution from DEXs.

GMX FDV/Protocol Revenue (Annualized) vs Market Proxy

If there was ever a moment that could serve as a strong catalyst for the growth of derivatives exchanges, it was surely the bankruptcy filing of FTX in 2022 after a series of controversial events. This unprecedented shock served as a stark reminder not to trust centralized players. But the question remains: have we truly learned the lessons this time?

dYdX FDV/Protocol Revenue (Annualized) vs Market Proxy

Despite the unprecedented shock of FTX’s bankruptcy, recent data indicates that we may not have fully learned our lesson, or alternatively, that existing DEX solutions may not be equipped to meet the demand. The contribution of decentralized derivatives and spot exchanges to the total crypto exchange volume has even decreased, with these platforms accounting for only around 1.3% and 6% of CEX volume, respectively. This serves as a sobering reminder that while the industry has made progress, there is still much work to be done to build a truly decentralized and trustworthy system.

Taking a closer look at the market performance of leading players like GMX and dYdX, there doesn’t seem to be a significant increase in enthusiasm for the future of these projects. Their respective tokens are still priced at similar to lower multiples (FDV/Annualized Revenue) to those of the pre-FTX crash period and the recent increase in multiples is reflective of the general market increase, suggesting a lack of confidence in their ability to outperform the markets by exponentially growing their volumes.

*** Market proxy is taking into account protocols with revenue generating capabilities and taking mean of their respective FDV/Revenue ratios. It includes the following projects: Lido, AAVE, Maker, Pancake Swap, Compound, Balancer and Sushiswap

So, can DEXs really take over the crypto trading game? We’ve got some important considerations to keep in mind, including the key ways that DEXs stand out from their centralized counterparts. But beyond that, we also need to delve into the practical effects of differentiation points and take a look at the various events that could either bolster or undermine their importance.

For example, if regulators were to clamp down on centralized exchanges, this could create more short-term demand for DEXs as users seek alternative trading options. On the other hand, if the market experiences a prolonged period of stability, users may be less concerned about the potential risks associated with centralized exchanges and may opt for the convenience and efficiency they offer.

Dozens of DEXs have attempted to gain a foothold in the crypto derivatives market over the past two years, but most have struggled to attract organic growth as a large chunk of trading volume has been generated by bots and traders who exploit token economics or liquidity providers.

For instance, according to TokenTerminal, dYdX and Synthetix are among the top 5 dApps with the largest discrepancy between paid incentives and earned fees. Specifically, dYdX and Synthetix have a negative balance of $750 million and $650 million, respectively, indicating that they have paid more incentives than they have earned in fees.

Is the token price increasing due to the rising exchange trading volume i.e. better fundamentals or vice versa?

While Synthetix has chosen to use its rewards to incentivize the supply side, i.e., the stakers who form the backbone of the network, dYdX has opted for direct incentives to trading activity via its trade mining program. Such incentives can lead to impracticality in determining the true organic activity.

1. users start mining tokens pre-token launch (could do so with delta-neutral strategies);
2. tokens get priced by the market relative to the activity on the platform, thus FOMO in the first step directly reflects on the token FDV
3. high token price incentivized even more volume generation

For instance, some incentives indirectly incentivize trading activity as projects could inflate naïve liquidity via token mining rewards i.e. compensate LPs for the adversarial trading volume being generated at their expense.

Source: IOSG Ventures dYdX Fees/Incentives Ratio

In cases where incentives are explicitly targeted at trading activity, like with dYdX, it remains difficult to determine the extent to which the trading volume would exist without such rewards.

Essentially, trade mining has the potential to create the fly-wheel effect, where:

As a result, dYdX has been generating billions of dollars in volume daily. However, to what extent is this sustainable activity vs purely trade mining? Eventually, dYdX will run out of gas to fuel trading activity and its success would depend solely on the organic demand.

One Whale is Worth Thousands of Turtles Past 30 days volume source; Top 10 addresses vs others; Source:

GMX is a decentralized cryptocurrency derivative & spot exchange that operates on Arbitrum and Avalanche. The platform allows users to trade cryptocurrencies in a peer-to-pool manner without the need for intermediaries. GMX offers a variety of advanced trading features, including limit orders, stop-loss orders, and margin trading with up to 50x leverage.

By design, GMX is built for large traders. This is due to its zero-price impact trading. It achieves zero price impact trading by utilizing a liquidity pool model (GLP) and somewhat centralized oracle price feed. It allows traders to ‘rent’ all of the liquidity from the pool at the current market prices in exchange for 10 basis points trading fee and hourly borrowing cost.

Using $100k trading volume daily as a threshold for defining whale traders, we observe that roughly less than 10% of GMX traders could be classified as whales, yet this group of traders generates consistently more than 90% of the trading volume on the platform. This is consistent with the success stories of the majority of Tier 1 DeFi protocols, where time and time again finding PMF requires trade-offs that favor whales.

GMX has also been experimenting with innovative ways of utilizing token to bootstrap activity, however, its incentives are not focused on creating a false sense of traction due to the following:

On the importance of GLP

a) GMX doesn’t have direct incentives for traders such as the trade mining initiative

b) it is mostly focused on growing its TVL, but at the same implementing design choices that protect its LPs from adversarial volume i.e. it is not paying LPs to sit in the pool and tolerate arbitrage volume

Annualized borrowing rates; Source: Can GLP maintain sufficient TVL levels post-liquidity mining?

Hence, most of the GMX volume is, indeed, organic. The only exception could be the possibility of some traders speculating on the potential Arbitrum airdrop to the users of ecosystem dApps.

GLP Rolling APY

In comparison with typical perpetual swap contracts, GMX derivatives have a stricter ceiling as the open interest is limited by the depth of the liquidity pool to ensure the solvency of the protocol. Thus, scaling the liquidity pool has been the priority of GMX and these objectives have been reflected in GMX token economics.

1. hedge GLP basket of assets e.g. in case LP is not comfortable with the exposure to the volatility of certain tokens in the pool
2. hedge skewed market demand

The liquidity of ETH and BTC has been the most critical for GMX as traders are mostly interested in renting out exposure to these assets, which is illustrated by the historically largest borrowing fees for these assets.

Even when we ignore token mining rewards, GLP holders were able to harvest double-digit APY which in recent periods fluctuates between 15% and 30%.

Critics may argue that the yield offered by GLP is not commensurate with the significant directional risk exposure that GLP holders face. However, this criticism overlooks the fact that GLP holders are not necessarily passive market makers. While GLP may not have built-in risk management techniques, individual GLP holders have the ability to implement active hedging strategies. It is reasonable to assume that many GLP holders are doing just that:

Front-running: Tale as old as DeFi

Since hedging needs are typically met off-chain, it’s difficult to gauge the exact profit and loss of individual GLP holders. But, taking into account the yield and assuming reasonable hedging costs, it’s plausible that sophisticated market players can earn high single-digit to low double-digit APY (without factoring in token rewards) even after completely hedging their exposure.

At the moment, GLP is targeting a 50:50 ratio of volatile and stable assets, however, considering that stablecoins are largely underutilized there might be room to explore different target weights. Alternatively, GMX should find a way to increase the utilization of stablecoin assets.

Namely, one challenge facing GMX is the fact that the liquidity pool i.e. GLP is the counterparty to each trade, which means that borrowing fees are always paid to the pool regardless of the demand skew. This leaves room for imbalances and larger directional risk for GLP holders. In contrast, in regular perpetual swaps, if the market is dominantly bullish, it implies a larger cost of maintaining long positions but also the ability to earn funding payments by taking the other side of the trade. As such, assuming the market is efficient, no rational trader would pay for a short position on GMX if they could earn for having a short position on other venues. Similarly, when the overall market is bearish, there would be a lack of incentives to open long positions on GMX. However, since markets tend to take more leverage in bullish sentiments, this is less of a concern.

Arbitrage front-running is a common occurrence in the DeFi space and a major activity driver on blockchains. Despite this, protocols have been able to operate successfully with front-running risks. However, zero price impact designs present even greater risks as traders can exploit any information advantage to drain liquidity providers’ resources.

Front-running could happen as a result of:

a) oracle related issues: for instance, if oracle is missing an update for whatever reason or faces increased latency there is a risk of an arbitrageur front-running the price feed and exploiting any differences with stale oracle price

b) MEV: even if oracle is functioning as designed, on Ethereum mainnet arbitrageurs could read mempool data and attempt to sandwich the price feed update by paying extra gas cost

The optimal trade-off?

Moreover, on Ethereum L1 arbitrageurs could potentially read mempool data and attempt to push their transactions in the block prior to the oracle price update, thus exploiting SNX stakers.

While GMX operates on Arbitrum, which has centralized sequencing and avoids problem b), they had to devise innovative solutions to circumvent the danger of a).

💡 Slippage ≠ Price Impact

The way GMX resolves oracle issues is by accepting certain trade-offs on the UX part. Namely, users are subject to potentially large slippage, as their order would be executed only several seconds after the request has been sent to GMX i.e. the price is known to the users with significant latency. Users are able to control the slippage-related risks by defining an acceptable slippage range, yet, in volatile markets tight slippage range could mean a high number of order cancelations.

This is, of course, not the only trade-off coming from the zero price impact & oracle-centric solution. To protect liquidity providers from various attack vectors GMX has to limit its offering to only a small number of highly liquid tokens.

For instance, if GMX was to support some longer-tail assets, a trader could use substantial stablecoin liquidity on GMX to short such tokens with no price impact at all, even though a trade of similar size might move markets significantly on the most liquid venues such as Binance.

Not only that this would imply that LPs are selling liquidity below its true cost, but also presents a risk of a potential (costly) exploit where a sophisticated entity with substantial capital could open a position on GMX, move the market price and subsequently close the position on GMX with profit.

GMX might be able to circumvent these issues by implementing dynamic open interest caps and fees that could take the liquidity of the underlying as one of the factors in the calculation.

Nevertheless, even with these fixes, the GLP design would present an obstacle for the more diverse asset offering. Adding new risky assets to the GLP could substantially change its risk profile which may limit its ability to attract LPs as it would imply increased hedging complexity & cost and long-tail risk that LPs might not be willing to accept. Thus, it would be more prudent to consider creating independent new pools with different risk profiles.

And finally, GLP requires assets to already exist on-chain to be able to issue its derivatives whereas some other players are able to support derivatives on any off-chain assets with reliable price oracle.

Relative performance of $1000 invested in GMX and dYdX on the day of FTX crash

The decentralized exchange (DEX) derivatives market might still be in its infancy, yet GMX has shown impressive strength in the post-FTX crash landscape. Since then, GMX has managed to attract a significant number of new derivatives traders, numbering more than 40,000, who are contributing on average almost $100M in derivatives volumes daily.

The strong fundamentals are reflected in GMX’s relative outperformance compared to its main competitors since the FTX crash. A hypothetical investment of $1000 in GMX and dYdX on the day of the FTX crash would have become roughly $1400 and $800, respectively, as of today. While these results do not necessarily predict future performance, they suggest that GMX has managed to carve out a strong position in the competitive DEX derivatives market.

Source: Dune

GMX’s product is far from perfect, with limitations including a small asset selection, liquidity scaling issues, and a lack of adequate risk management. Furthermore, the platform’s trade-offs, such as sacrificing decentralization principles and introducing friction in the user experience, are not to be overlooked.

Nevertheless, some of these shortcomings might have been intentional trade-offs that helped GMX find the product-market fit and become the derivatives DEX with the most organic demand.

“ First to market seldom matters. Rather, first to product/market fit is almost always the long-term winner.”

This product-market fit is evident not only in GMX’s strong fundamentals but also in the number of forks it has received, making it one of the most forked DeFi protocols ever.

Assuming healthy internal dynamics and no black-swan events, this puts GMX in a prime position to gradually resolve all of its shortcomings and potentially evolve into a dominant player in the derivatives vertical.

But winning over users from centralized exchanges is no easy feat. The market has shown that being non-custodial is not enough. Instead, DEXs will have to beat CEXs at their own game by offering an equally good user experience that includes factors such as onboarding ease, trading costs, latency, price impact, asset offerings, advanced trading features, availability, reliability, and even recovery of lost funds.

Unfortunately, nowadays on the spectrum between “don’t be evil” and “can’t be evil” exchanges the optimal user experience requires exchanges to build closer to the left side of the spectrum.

However, as the technology matures it would become possible to move towards the right-hand side without sacrificing UX. Thus, having a long-term perspective in mind DEXs are positioned on the winning side of the spectrum.

But, CEXs are not without recourse in this developing landscape. We shall expect them to gradually improve their standards and try to defend their existing market position by adopting cryptographic solutions, firstly offering a hybrid model that lessens custody & transparency-related concerns, and ultimately, once primitive evolve enough not to require UX sacrifices, even fully becoming decentralized.

IOSG Ventures is a community-friendly and research-driven early-stage venture firm. Our portfolio covers L1/L2(Polkadot, NEAR, Arbitrum, Starkware), DeFi(1inch, 0x, Synthetix, DODO, UMA), GameFi(Bigtime, Illuvium) and SocialFi(Roll, Galaxy，CyberConnect). We are crypto-native BUIDLer and long-term HODLer to our early-stage developers & founders.

If you would like IOSG Ventures to consider your project, please send a summary of your project along with a pitch deck and/or white paper to hello@iosg.vc

Follow us on Twitter.

Originally published at https://medium.com on March 27, 2023.