This follows our previous piece, Building On-chain Marketplaces

Introduction

To recap, no DeFi options venue to date has managed to garner significant and long-lasting usage. Yet options are so useful an instrument and so pervasive in TradFi that we believe it is inevitable they will eventually gain mass adoption on DeFi rails.

To understand why, we must examine the advantages, shortcomings, and directions of potential improvement of the three primary offerings in the current DeFi options landscape —DeFi option vaults (DOVs), option AMMs, and order book-based exchanges.

DOVs introduced options flow to DeFi by wrapping the options writing process in an accessible and user-friendly yield-generating format. However, these vaults have seen a precipitous drop in users and TVL due to the financial inviability of their initial strategies, the significant underperformance of these strategies relative to alternative forms of yield generation, and the misaligned incentives between market makers, vault creators, and vault depositors.

AMMs originally innovated in the spot markets by enabling anyone to provide instant liquidity for long-tail assets. Options, though, are unsuited for AMMs because of the inherent complexity of options pricing and the significant technical design challenges of bringing high-fidelity pricing on-chain. Without efficient and accurate pricing, LPs risk losing significant money to toxic flow, a problem that often becomes so acute with option AMMs that LPs cannot make positive returns without charging prohibitively high fees. This toxic flow inhibits liquidity, which results in uncompetitive pricing for end users.

Order book-based exchanges work well given sufficient liquidity and demand, but the latency and throughput limitations of current blockchain technology make purely on-chain order books compare unfavorably to their CeFi counterparts. The idiosyncratic risk of using high-latency on-chain order books can cause material and uncontrollable adverse trading outcomes for market participants. This forces traders and market makers to limit their exposure to these venues, exacerbating the cold-start liquidity problem that all order book exchanges face. Hybrid order books have great potential but face significant challenges specific to options marketplaces—namely, the inherently fragmented nature of options liquidity and the need to build an integrated, liquid futures exchange to enable under-collateralized options writing.

There are reasons to be optimistic, though, about future improvements to these three options liquidity venues. Better designs are currently held back by the performance limitations of blockchains and still-developing oracle infrastructure. As blockchains improve their latency and throughput and as on-chain oracles become more robust, we can imagine teams experimenting with more sophisticated and financially viable designs that provide more value to their users.

Now, we will examine more closely the challenges facing the various attempts to bring options on-chain in order to help inform the next iteration of DeFi product design. It is important to note that while we may sometimes be critical in our analysis, we have the benefit of hindsight, and many of our conclusions would have been difficult to reach without the bold path paved by earlier projects.

DOVs

DOVs found early success in bringing options on-chain by wrapping systematic volatility selling in a straightforward UX, in which users deposit collateral into vaults that generate “yield” by selling options against the vault's collateral. These vaults inject options-selling flow into the ecosystem by letting retail users engage in options writing, an activity traditionally undertaken exclusively by sophisticated options traders, market makers, and corporate treasurers.

Despite their initial traction, DOVs have generated poor returns for depositors and have seen a significant drop in users and TVL since inception. This is due to (1) the financial inviability of the current strategies employed by DOVs and (2) misaligned incentives between vault depositors, vault creators, and market makers.

Heads I Win, Tails You Lose

There are two methods for profitably employing options-selling strategies. The first is volatility arbitrage, a complex strategy conducted by sophisticated options market makers. The second is yield enhancement on top of directional exposure, which is how DOVs typically run their strategies. The two most common DOV strategies are call overwriting—selling fully collateralized calls on an underlying token like ETH—and cash-secured put selling—selling fully collateralized puts using USDC as collateral.

Let's walk through the mechanics of call selling strategies—the most common vault strategy—to illustrate the economic pitfalls of current DOV strategies (spoiler—there is indeed “no free lunch”).

There are two types of risk when selling call options: downside risk and upside risk. With a call selling strategy, you are exposed to downside risk if the underlying asset falls in price, since that asset is the collateral you deposited to write the options. This risk is only partially mitigated by the option premium collected. You also have upside risk—because selling a call caps your upside, if the underlying increases dramatically in value, you risk losing out on an enormous amount of profit vs. simply holding the asset. Put another way, the max potential profit of selling a covered call is just the option premium, while the max potential loss is unlimited in USD terms.

Call selling DOVs often find themselves in the no-mans land of yield strategies. In raging bull markets, a naive covered call selling vault will significantly underperform a strategy of simply holding the spot asset, as the premiums gained from call selling will not offset the massive opportunity cost in upside. In secular bear markets, the decrease in the value of the underlying collateral will not be sufficiently offset by the small APR collected from premiums. This dynamic illustrates that the profitability of call selling strategies is highly conditional on either a choppy + bearish or mildly bullish market regime and they are therefore not suitable for use in the indiscriminate set-and-forget format employed by vaults.

Let's walk through the math behind this.

Typically, DOV call selling strategies sell ~10-delta calls on a weekly basis. The typical yield of such a vault is around ~18% APR in premium for selling ETH covered calls (0.35% weekly). Given the current 365-day volatility of ETH is around ~90%, assuming a normal distribution, the probability of an 18% drawdown on a weekly basis is 8%. Such a move would completely wipe out the annual profits of the vault, and over 52 consecutive weeks, the probability of such a move is around 98.7%. From this, we can see that it's actually quite likely for a single not-even-black-swan-level event to wipe out an entire year's worth of vault profits from call selling. This illustrates that when running this strategy in isolation, the premiums collected from call selling do not offset the risk of nakedly holding the volatile underlying asset. This naturally leads us to our next question: what's the cost of selling crypto upside for pennies on the dollar?

A 10-delta call has an approximately 10% chance of being in the money. This means a 10-delta ETH call selling vault has a 90% chance of keeping the depositors' collateral on a per-week basis, assuming weekly options selling. Over a year of weekly indiscriminate covered call selling, you have only a 0.4% chance of keeping the initial ETH principal deposited in the vault (a 99.6% of losing some of that initial ETH). Historical backtests provide evidence that in bullish market conditions, covered call strategies can result in severe loss of upside and principal and significantly underperform relative to simple buy-and-hold strategies, to the tune of underperforming buy-and-hold by 430% during the historic bull run from April 2019 to January 2022.

In addition to the negative expected value of a naive covered call selling strategy, DOV also suffer from front-running during their weekly auctions, which tends to line up with weekly expiries on major CEXes. These predictable selling patterns incentivize various hedge funds and market makers to front-run the weekly auctions by shorting short-dated options and buying them back after the predictable selling flow passes, resulting in below-market pricing for vault depositors. The price impact resulting from predictable auctions leads to alpha decay, causing reduced profitability as TVL grows.

The graph below shows the impact of DOV auctions on implied volatility: on average, implied volatility is 4 vols lower on Fridays (when DOV's typically conduct auctions) than on any other day of the week. For 1-week 10 delta calls, prices being 4 vols lower means vault depositors suffer from a decrease in APY of around 5.35% .

In conclusion, despite their claims and projected yields, simple DOV call or put selling strategies are far inferior to those employed by sophisticated traders and portfolio managers and often even compared to just buying and staking. Blindly selling volatility this way is equivalent to mining for gold specks as a volcano erupts above you.

Principal-Agent Misalignment

In addition to the dubious viability of indiscriminate options selling, DOVs suffer from severe incentive misalignment between market makers, vault creators, and vault depositors. In short, the incentives of the participants are as follows:

• Vault creators: want to extract maximal performance and management fees
• Vault depositors: want to invest in strategies that outperform the market on a risk-adjusted basis
• Market makers: want to cheaply purchase volatility from options vaults

Market makers and vault creators are adversarial in the way that participants in any two-sided marketplace are: the buyer—in this case, market maker—wants to purchase at a low price, and the seller—the DOV—want to sell at a high price. Normally, arbitrage brings these prices into line for efficient markets. No such mechanism exists for DOVs, though, since vault depositors are forced to sell to the highest bidder, and the universe of bidders is limited to the set of market makers allowed to participate in DOV auctions. Given DOVs do not provide any form of secondary liquidity for offloading inventory, market makers are only incentivized to purchase options at below “fair-value” prices, which is negative EV for vault depositors. Over time, though DOV pricing ought to improve as increased market maker participation makes the auctions more competitive and vaults adopt more non-standard expiries.

A greater issue is the misalignment between vault creators and vault depositors. The incentives of vault creators are at odds with depositors, given the primary incentive of vault creators is to maximize the fees they can capture. This misalignment is illustrated in the “performance” fee model employed by vaults.

DOVs typically charge their users a fee on the option premium, incentivizing vaults to sell as many options as possible, with frequent auto-rolling. For example, call selling vaults typically charge performance fees whenever the call option that was sold expires out of the money. However, if the price of the underlying asset falls drastically, the call option would still expire out-of-the-money, yet the vault would still be able to charge “performance fees” despite the user losing a majority of their principal in USD-terms. If the price of the underlying asset increases to the extent the call options expire way ITM—causing a significant loss of the underlying vault collateral—the vault could still collect fees next week by simply rolling the strategy and selling new calls at a higher strike price. This weekly rolling means that the vault can charge users a 10% performance fee on a strategy that, in the previous week, may have lost 20% of the vault collateral.

These fees significantly affect the return profile of options overwriting strategies, and in TradFi, collecting such high performance and management fees on these simplistic strategies would be seen as extreme profiteering. We find it hard to believe that they have a long-term place in DeFi.

Future of DOVs

While the directional exposure inherent to passive call selling or put selling options vaults is negative EV, selling volatility can be sustainably profitable with more sophisticated strategies. We see three promising avenues for DOV's to explore: delta-neutral vaults, user-managed strategies, and pool-delegate approaches.

Delta-neutral vaults can mitigate the negative impact of directional exposure inherent in traditional DOVs. Protocols such as Opyn and Friktion already offer “crab strategies” which take advantage of structural differentials between implied vol and realized vol. As the chart below shows, implied vol for crypto is typically systematically higher than realized vol. This difference can be used to turn a profit.

User-managed strategies would leverage the liquidity and connectivity to market makers that DOV's have already established in order to provide an easy-to-use interface that lets users define their own covered call and put selling strategies. Traders could create their own covered call strategies to implement custom sell limit orders with additional yield as well as put selling strategies to place buy limit orders. These are actively managed directional strategies and would not include an auto-rolling feature.

Finally, we can imagine the creation of actively-managed option selling “pods”, inspired by both the actively-managed volatility funds in TradFi as well as the success of the pool-delegate approach pioneered by Maple Finance. Professional options traders would strategically sell options to earn yield while managing portfolio-level risk characteristics in accordance with their market views. They would earn performance fees based on aggregate fund performance and would be managing a portion of their own capital, which resembles the incentive structure traditionally used for asset managers. The key here is designing a robust way to encode risk controls for pod managers. This could be either directly within the protocol or through Maple's approach of providing a generalized framework for legal recourse.

In conclusion, we believe DOVs are really competing against alternative forms of yield generation rather than full-blown options marketplaces, and we expect most DOVs to pivot to more sophisticated strategies and alternative verticals—such as structured credit products—since vanilla DOV options writing strategies have consistently produced subpar returns.

AMMs

Spot AMMs found early and lasting product market fit in DeFi by making it possible for anyone to provide instant liquidity for long-tail token swapping in a permissionless manner. Takers still use spot AMMs despite their pricing inefficiency and fees because they gain access to liquidity on long-tail assets that aren't listed on most or any exchanges, and liquidity providers (LPs) enjoy the benefits of what is essentially an automated portfolio re-balancer while having a chance to collect trading fees.

Option AMMs have not experienced the same degree of success due to the challenges of pricing options on-chain. Inefficient price discovery leads to increased vulnerability to toxic flow, making it difficult for LPs to make money without charging takers inordinate fees. This restricts the liquidity that can be deployed, resulting in uncompetitive pricing and low trading volumes.

Toxic Flow

To understand why it is so difficult for option AMM LPs to consistently earn a profit, first consider what it means to be a market maker or liquidity provider more broadly. Liquidity providers risk capital to have a chance at earning a profit—such as by market making on an exchange or providing capital to an AMM. Most liquidity providers are exposed to toxic arbitrage flow, which is where informed takers who know more about the ‘true price' of an asset can exploit stale pricing from makers to pocket a spread. While takers can choose when and where to trade, makers provide constant liquidity, e.g. honoring resting bids on order books).

LPs for AMMs are particularly exposed to toxic flow because they rely on a correctly calibrated pricing function encoded in the smart contract, unlike in an order book-based exchange where individual makers can dynamically reprice their quotes as they learn more about the true price. LPs receive trading fees from takers to offset the trading losses that result from consistent exposure to toxic flow, though even these fees often aren't enough. Market participants may still find value in LPing for spot AMMs, though, because spot AMMs function as an automated portfolio rebalancer, with additional yield from trading fees on top. These LPs are happy to distribute the dollar-value of their portfolio across certain assets equally, as long as they receive fees to compensate for the impermanent loss. This is of little use in option AMMs, since few, if any, are trying to maintain a balanced portfolio of a single option type and spot.

To recap, there are always sophisticated takers looking to arbitrage assets based on an asymmetric advantage in knowledge of prices, speed, or access to different venues, and AMM LPs are exposed to an immense amount of toxic flow. Now, we'll describe why this toxic flow is particularly damaging when it comes to options.

Options Pricing

There are two types of option AMMs in use today. The first can be likened to a DOV—in these AMMs, LPs deposit collateral into pools, and the collateral is used to sell options against various strikes and expiries. The second type is more sophisticated and tries to mimic a professional options trader by trading volatility and delta hedging with futures or perps. However, the only way either of these designs can reliably make money for their LPs is by charging exorbitantly high fees, which means no taker is going to want to trade.

Why is it particularly challenging for LPs to make money providing liquidity to option AMMs? Because toxic flow is even more damaging when trading options than when trading spot. Pricing an option is harder than pricing spot, and selling a mispriced option exposes you to exponentially greater downside risk than mispricing a spot asset. Let's talk about why.

Fundamentally, the price of an option is much more difficult to determine than the price of a spot asset. To update a high-fidelity spot price for an asset on-chain, you can construct an aggregated trade tape from major centralized exchanges and take the volume weighted average for a given period (for example: the past 1 minute). However, this same technique cannot be applied to options—for any given option, the underlying spot asset will trade much more frequently than that option itself, and since the price of spot is such an important input into the option pricing model, an algorithm that naively computes the volume-weighted average price for an option would always meaningfully lag the market. In addition to constantly changing spot price, another key input to an option's price is implied volatility, and each market participant will often have a different implied vol value for the exact same option depending on their proprietary, internally calibrated vol surfaces. In short, pricing an option quickly and accurately requires a complex model, high-frequency market data ingestion, and sometimes even manual intervention.

Additionally, mispricing an option as a seller exposes you to exponentially greater potential loss than mispricing a spot order. Spot has a linear payoff—a 20% move against your spot position just means you are down 20%. Options, though, are highly-levered instruments with convex payoffs. Selling an option means you are short gamma, i.e. you're exposed to negative convexity and hope the price doesn't move too much. If you are short an option, a 20% move against you within a short time frame could easily result in a significant drawdown or a flat-out blowup, depending on the implied volatility you sold it at and whether or not you appropriately delta-hedged the position.

The difficulties and risks inherent in pricing options are magnified by the lack of computational power, speed, and storage space available on the blockchain for on-chain AMMs. On-chain AMMs are significantly less effective at quickly pricing options compared to off-chain entities—most notably sophisticated options trading firms—who will take advantage of pricing inaccuracies and their superior access to high-fidelity real-time market data to arbitrage against the AMM.

Future of AMMs

In conclusion, it's easy to misprice an option, and you lose exponentially more money doing so when selling an option than if you were to misprice spot. Off-chain entities are far quicker and more likely to correctly price an option faster than any on-chain AMM, so option AMM LPs are quite likely to consistently lose money providing capital unless the AMM charges exorbitant fees that drive away takers. To date, no AMM has managed to solve these issues, though there is the potential for improvements in blockchain and oracle infrastructure to allow for innovative designs that can successfully compete in these environments. One opportunity here is an oracle approach based on incentive alignment, where a vol surface for an AMM can be calibrated off-chain from professional market-makers who each stake collateral, and who get slashed if their estimates diverge from the vol surface reached via an on-consensus mechanism.

Order Books

Central-limit order books (CLOBs) provide efficient price discovery and are the primary market microstructure for trading crypto options today. They feature options chains that list strike and expiry combinations for various underlying assets, and each permutation of strike, expiry, and asset has its own distinct order book. Deribit—a centralized exchange—commands the greatest market share, with over 85% of all BTC options trading volume. No one else comes close in terms of liquidity or product features.

In DeFi, there are two approaches for building CLOBs: On-chain order books, where order matching and trade settlement both occur on-chain, and hybrid order books, where order matching occurs off-chain, with trade settlement on-chain.

On-chain order books are constrained by the throughput and latency limitations inherent to blockchain infrastructure today. These limitations make it risky for market makers to provide deep liquidity, and make it difficult for protocol creators to design safe liquidation engines. These limitations require significant technological improvements in core blockchain infrastructure to overcome.

Hybrid order books seem promising in the medium term, though they still face problems with the liquidity fragmentation inherent to options and the need to offer portfolio margin for options writing, which necessitates a liquid perp exchange.

On-chain Order Books

Market makers are currently willing to deploy very little liquidity onto on-chain order books, partially due to the lack of institutional-grade infrastructure required to do this safely (such as wallet infra), but primarily because they are exposed to uncontrollable adverse outcomes caused by the high-latency and low-throughput of current blockchains.

On-chain order books are vulnerable to MEV and other forms of front-running and market manipulation. The high latency of on-chain order books can cause unacceptable trading outcomes, such as being unable to cancel orders during periods of network degradation. Furthermore these latency / throughput limitations also impossible for a market maker for an on-chain CLOB to be particularly tight or competitive when they are required to mass refresh their quotes. These performance limitations cause on-chain order book exchanges to be generally less liquid and suffer from worse pricing than their centralized counterparts.

In addition to latency & performance issues, the transparency of orders and positions tied to a unique wallet for on-chain order books can be daunting for both takers and makers. Neither a market maker nor a liquidity taker would want everyone to be able to see their orders and their inventory/balances.

While the ideal terminal state for trustless and censorship-resistant marketplaces might be on-chain, we believe the current performance and latency limitations that face on-chain order books prevent them from competing on equal footing with Deribit.

Hybrid Order Books

Hybrid order books seem technically feasible in the medium term, and there are notable projects building in this direction. There are two problems they must overcome, though. The first is the fragmented liquidity inherent to options and the difficulty of sourcing options market makers who can provide said liquidity. The second is that selling options is very capital intensive, so it's necessary to have a robust and liquid perp market to provide market makers sufficient capital efficiency for options writing. These challenges also afflict on-chain order books to an even greater degree.

Fragmented liquidity

CLOBs require deep liquidity aggregated on a single book. This is tricky with options because there are so many possible combinations of strikes and expiries for a given asset. This liquidity fragmentation makes option order books susceptible to cold-start issues and forces order book operators to confront a difficult tradeoff: they can either offer many different strikes and expiries at the cost of sacrificing liquidity, or they can concentrate liquidity by offering only a limited set of trading choices. Because CLOBs require deep liquidity on every book, order book-based exchanges need to onboard sophisticated market-makers who can provide 24/7 real-time pricing. However, options market-makers who can algorithmically quote on order books are a very specialized subset of an already small universe of crypto-native market makers, and they have little incentive to invest in the heavy technical integration required to onboard onto a new exchange unless there exists some proof of volume that can be profitably traded against. This proof, however, cannot be shown without a market maker already on board — a classic chicken-or-egg problem.

Portfolio Margin

It is extremely capital intensive to write options due to their convex payoff profile. A trader can purchase an OTM call option for a small premium while receiving uncapped exposure to the underlying asset. The option seller must collateralize this exposure, which can require a significant amount of capital. Imagine selling 1,000 ETH call options at a $2,000 strike price. If ETH shoots up to $3,000 at expiry, you are on the hook for $1 million less the initial premium received.

Options market makers will generally hedge this directional exposure, so any time they trade an option, they also add an offsetting futures position to delta-hedge. In crypto, they would typically hedge through a perp (the most liquid type of future). If they sold calls, they would simultaneously get long perps, so that if the calls expire ITM, the losses from the option position would be partially offset by gains in their perp position. This hedging allows them to remain delta neutral, and profit via the arbitrage of the other greeks which they have an informed view on.

While being short an option and being long a perp are individually capital-intensive positions, in correct proportions they have minimal directional exposure and can thus collateralize each-other. Sophisticated options exchanges such as Deribit allow market makers to utilize their offsetting options and perp positions to significantly reduce their capital requirements through a process called portfolio margining.

Any options exchange that offers portfolio margining must have liquid perp exchange and—as recent events have shown—perp exchanges are notoriously difficult to operate safely. This adds a daunting barrier to entry for teams in space: not only do they need to build a fully functional options marketplace, they also need bootstrap a secure, robust and liquid perp exchange. If the perp exchange does not have sufficient liquidity, large market moves can threaten the integrity of the options contracts, as the perps used as collateral will not be effectively liquidated, meaning that the options holders will not get paid out.

In these tail-scenarios, exchanges rely on carefully designed risk controls and a robust liquidation engine. This is particularly challenging to safely implement on-chain, given blockchain performance limitations. Without a robust liquidation engine, exchanges risk total insolvency and must cover losses with an insurance fund to make users whole or enforce clawbacks and socialized losses.

Future of DeFi Order Books

In conclusion, on-chain order books are not quite ready for prime time, and hybrid order books seem technologically feasible in the medium term, but face a long road ahead before they can reach feature and liquidity parity with Deribit.

That said, we remain optimistic about developments in DeFi CLOB design. On-chain order-books can take advantage of the latest developments in blockchain scaling solutions to mitigate their performance issues. They can also use privacy-preserving cryptographic primitives to prevent tracking of orders and positions on a per-user basis. Hybrid order books can take advantage of their native composability with on-chain primitives such as DOVs to bootstrap liquidity. As primitives advance, both architectures will become easier to implement, and there may be an inflection point where the trade-offs start comparing favorably against centralized alternatives.

Conclusion

The advent of smart contracts set off a Cambrian explosion of financial experimentation in DeFi and the possibilities of internet-native digital asset ownership, which is even now taking place at the speed of the internet. While there is tremendous excitement around these new primitives, for any financial product to last, regardless of whether it's TradFi, CeFi, DeFi, or Ce-DeFi, must be financially viable and economically sustainable. Oftentimes, the excitement generated by new projects in this space have masked fundamental problems, and we have seen countless examples of unsustainable token emissions driving unsustainable usage before crashing back down to earth when the emissions run out.

At the same time, the community has also been able to use DeFi primitives to democratize access to financial services once accessible only to a select few. DeFi's promise is in producing equitable and efficient financial markets that mitigate the systemic risks, reliance on trust, and lack of transparency common to TradFi . We see this transparency as essential to DeFi's ethos, both in how products are built and ideas are communicated, which was a significant reason we thought it useful to conduct this analysis of the DeFi options space and open it to the public.

Options are one of the most versatile and important risk transformation tools we have at our disposal, but their current implementation in DeFi can be vastly improved. While building products in DeFi faces structural challenges in performance and technical maturity relative to CeFi, the non-custodial, transparent, and permisionless nature of DeFi offers builders important dimensions on which they can compete. Our mission as a team is to drive further DeFi adoption, and in the third and final essay, we will introduce the first product in the Revv ecosystem, which we believe can drive this exact type of adoption.