Today, I’ll build on some thoughts I recently wrote about prediction arbitrage.
In prediction arb, you take both sides of a trade i.e., bet on opposite outcomes, in separate prediction markets. By taking advantage of pricing gaps between the markets, you can get a nominally guaranteed profit.
In the last post, I give a simple example of arbing two markets predicting a Dwayne Johnson presidency. If you haven’t, you may want to check out that post for context. Otherwise, some of the following may not make sense.
Since prediction arb typcially requires locking up funds until the markets resolve, we need to be sure the expected return exceeds what you can get from fixed-income assets. As a benchmark, we can use the annual “risk-free” rate of return from Treasury yields i.e., interest on U.S. government debt, which is currently ~3%.
So in order to surpass this ROI, a prediction arb for a market resolving in a month would need at least a .25% spread (3% divided by 12) while a market resolving in 2 years would need at least a 6% spread.
For simplicity, I’m assuming the end return will equal the pricing gap between the markets i.e., the spread. This is putting aside transaction fees, risks of invalid resolution, risk of prolonged market reporting etc. That’s why I say that profits from prediction arb are nominally guaranteed.
Let’s take a concrete example.
As I write this, you can sell shares, i.e., bet against, Trump getting re-elected in 2020 on Augur at 37% odds and buy shares on PredictIt at ~35%. That’s a 2% “guaranteed” return, but given that the election won’t happen for over two years, that comes out lower than the “risk-free” rate of return via Treasuries.
This prompts the million dollar (or Ether?) question: in what scenarios would the price spread between prediction markets be so dramatic as to create truly profitable arb trades?
In the last post, I looked at why prediction arbitrage opportunities would exist in the first place. Understanding why they arise is essential because only then can we know where to look for them. Of the 6 reasons I listed, I think the most promising at this point in time is:
“One market reacting to news and other developments at a slower speed than other markets. For example, due to gas prices and low liquidity, Augur may lag behind a market like PredictIt.”
For example, let’s say the star QB of a top-tier NFL team gets a torn ACL during gameplay. A mature prediction market forecasting whether the team will win the Superbowl would price in this event in a matter of minutes, if not seconds. The price would plummet as holders of YES shares panic sold and buy support evaporated.
But on Augur, the price could take far longer to correct. This market lag effect would be further amplified if gas costs were higher than usual and/or crypto prices are falling: since Augur traders hold crypto assets they are less likely to trade when crypto prices, especially Ether, are plummeting.
In such a case, it’s worth asking if it makes sense to run arb by taking both sides of the trade or whether it’s best to just trade in the market that is lagging?
The latter offers a potentially greater and quicker return but at a higher risk, so this would depend on your risk appetite. If you were absolutely certain that the lagging market would correct and that you’d be able to close your position (or wait till resolution) then taking one side may be the best option.
Just remember that low liquidity is a double-edged sword: easier to find mispriced shares but also harder to find someone else to trade them to i.e., close your position.
An in-between option would be hedging your risk, by say, selling 10 shares against the team winning the Superbowl on Augur while buying only 5 shares on the other market.
The other perk of an arb opportunity caused by market lag is that you would not need to wait for the market to resolve to get out. You could potentially get out as soon as the market corrected.
So rather than waiting say months, you could get out and collect your returns in the span of say, hours. If you sold shares on Augur, you could immediately post a buy offer at a lower price. By looking at where prices corrected to on the faster market, you could know how to price these shares.
So putting this all together, to identify a good arb trade look for an Augur market that…
- Has an equivalent market on another prediction platform like PredictIt speculating on the same outcome.
- Is prone to sudden price swings based on news and other developments. For example, whether Kamala Harris will be elected president will swing if/when she announces her candidacy, a scandal surfaces, or new polling numbers come out.
- Doesn’t expire in too long or has a significant spread that exceeds the “risk-free” rate of return from other asset classes.
Keep in mind there are many other types of arb opportunities with different causes or with different market types i.e., scalar and categorical markets. This is just one type that may have unique potential at this point in time.
Thanks for reading. To stay ahead with fresh insights on the future of prediction markets, join The Augur Edge.