Binary Options Primer

Introduction
SIP-53 proposes a mechanism for Parimutuel Binary Options on Synthetix. This post will provide an overview of the proposed SIP along with examples of markets it can support. It will provide an overview of binary options, parimutuel markets and other information useful to assess the value of implementing this SIP into the Synthetix protocol. It will also discuss alternative approaches to binary options markets that could be explored in the future.

Binary options
A binary option is a contract that allows someone to make a trade on a yes/no outcome. For example a market could be created on the statement; BTC will be above $10k USD on 31st December 2020. This statement will be either true or false. True is represented by long options and false is represented by short options. When true the long side of the market resolves to $1 and the short side resolves to $0, when false the long side resolves to $0 and the short side to $1. Binary options are useful for hedging and placing simple trades on specific outcomes rather than holding an asset.

Parimutuel markets
Parimutuel markets are a way of defining prices and payouts based on a pool of funds. In the case of parimutuel binary options the payout is defined by the total pool of funds on either side of the market. A simple example is a market with $500 long and $500 short, there is $1000 total in the pool. This market is pricing in equal probability of both outcomes. If the long side is true the entire $1000 will be paid to longs, who put $500 in so the payout is 2:1 or more simply they paid 50c per option and the options paid $1. Now take another example where the breakdown is $200 long and $800 short, again the market has $1000 in the pool. So the long price is 20c and the short price is 80c. This second market is assigning a probability of 20% to the long side and 80% to the short side.

Binary options phases
The mechanism defined in SIP-53 uses a parimutuel system for pricing binary options, one trade-off is that a trader does not know what the final price of the option will be. There will be an indicative price offered at the time the option is opened, but this price will shift as the market moves. In order to counter this trade-off each market is split into phases.

1. Establishment - where parameters are set and initial capital is deposited into the contract
2. Bidding - where traders can place bids in a floating price auction
3. Trading - where the number of options is finalised are issued as erc-20 tokens
4. Maturity - where the outcome is defined and the options can be exercised
   

The establishment phase requires a market realtor to choose specific parameters for the market including the maturity date, the strike price and the price feed. Once specified initial capital is seeded and the bidding period begins.

During the bidding period the price floats, anyone can place or withdraw a bid in the market up until the trading period begins.

Once the trading period begins the number of options are fixed and issued as erc-20 tokens. They can then be traded OTC or held to maturity.

At maturity the options on the correct side of the market can be exercised for $1 and the option on the wrong side expire worthless at $0.

Market structure and timing
One of the consequences of this market design is that the timing of the phases of each options market are critical to ensure participation. While some experimentation is needed on the part of market creators, there are some likely patterns we can expect to see based on experience with other parimutuel markets.

One of these patterns is that most of the volume is likely to occur as the bidding phase closes, this is due to the fact that market participants will be trying to optimise the amount of information about the likely outcome. One way to offset this is constraint the length of the bidding period relative to the trading period. For example if the bidding period is 24h and the trading period is 6 months, there is very little difference between the information a trader placing a bid at the start of the bidding period has access to vs a trader placing a last minute bid. This is especially true given the floating nature of the parimutuel pricing. However, if the reverse was true and the bidding period was set to 6 months and the trading period 24h it is unlikely that any activity would occur until very late in the bidding period because the vast majority of information about the market remains unknown relative to the time before maturity. There is of course nothing stopping a market creator from constructing a market in this fashion, but it will be capital inefficient as the initial capital will be locked during a long period of low activity in the market.

Pricing Volatility
One of the interesting considerations of these markets is pricing volatility. Let’s consider the market similar to that mentioned earlier in the post, BTC above $10k USD. But let’s pick a maturity date of 31st of May 2020 and set a bidding period of one week. In this market as the price of BTC moves on a daily basis we are likely to see fairly large swings between the probability assigned by the market to each outcome. For example yesterday the price of BTC was under $7k and is now above $9k, it is likely that the market would have assigned a low probability maybe below 25% to BTC being above $10k 24h ago, but now it may be even be over 50% given the price momentum. This kind of price action is good for both liquidity and volume so one of the keys is likely to be to keep the strike price close to the current spot price on market creation.

Payout volatility
Another consequence of the parimutuel system is that you can get an estimate of the payout on your options at any time but this will not be locked in until the bidding period closes. Take an example of a market with a one week bidding period on BTC over $10k USD, the market starts with initial probabilities of 20% long and 80% short. A trader enters the market with a $1 trade going long and the probability does not shift due to the small size of the trade. However, over the course of the week the price of BTC keeps rising such that at the end of the week the price is well above $10k and now the odds have flipped to 80% long and 20% short. One way to look at this is that the trader who placed her bet early in the bidding period has lost money because the price was not locked in, however, this is a misunderstanding of the mechanism as the initial price of a bid is only reflected at the close of the bidding period. So anyone placing a bid should expect that new information could enter the market the longer the time remaining in the bidding period. In this case the trader may choose to withdraw her bid as the bidding period gets closer to closing, as the expected profit and risk adjusted return may no longer be positive based on her current beliefs. However, if the bid had been placed at the close of the bidding period and the price of BTC had risen subsequently then the trader would have locked in her odds and the price movement would be beneficial. This is again the reason why most volume will happen closer to the end of the bidding period based on the expectations market participants as the bidding closes.

Alternative approaches
The approach to binary options as defined in SIP-53 is certainly not the only approach possible, but it does meet the requirements of providing a pooled liquidity model that does not expose the debt pool to undue risk while leveraging Synth liquidity. There are some alternative approaches possible, such as removing the trading period and only allowing bids but not allowing withdrawals. Or issuing options only as long/short balanced pairs and allowing holders to sell the unneeded side into the market. One of the challenges with some alternative approaches is the exposure to parasitic profit cycles, there may be ways to reduce or eliminate these cycles in future models and additional R&D will likely be conducted in the future to determine the viability of these alternatives. That said it is very likely that the model as described will be more than sufficient for traders and provide a simple mechanism to hedge and speculate on that generates significant demand. In addition to this is it possible to extend the current mechanism to a multimodal model that accommodates multiple outcomes with the same constraints as the binary mechanism. Based on feedback and demand such alternatives may be enabled in the future.

Conclusion
This approach to binary options is simple and low risk to SNX stakers while extending the utility of the Synthetix platform. Post launch it will be up to the community to provide feedback on the mechanism to improve and hone it over time. One critical element to adoption will be to ensure a well designed user interface exists for interacting with these markets. This will be a separate piece of work to the protocol design specified by SIP-53 but is just as critical to adoption. Overall the prospects for these binary options and the potential volume they can generate is extremely high given their unique position in the DeFi market.