Traders tend to pay a lot of attention to expiration week, and for good reason. There are also some post-expiration phenomena worth noticing: in particular, unsophisticated options sellers often allow their front-month contracts to expire before selling the next series. There’s an apocryphal story about a big SPX condor seller who always used to wait until the Monday morning after expiration to route his single monster trade for the month. But we’ll leave discussion of the disadvantages of that approach for another time. Here, we want to find out whether the price range established during the week after expiration is of special significance. To be more specific, we’re asking:
Do mean-reversion bets on breakouts/breakdowns beyond the post-expiration weekly range provide any discernible edge?
Member Tom D. raised the question that got us interested, and he put it this way:
At the end of the first week of an option period, mark the high and low of that first week’s trading. Wait for price to exit that range either up or down. Then trade for price to return to, or at least make a large move toward, that first week’s range before the end of the option period. (...) My fantasy about all of this was the following. As the market moves further from the range, idiot momentum players are buying calls if the move is above the range, and fearful hedgers are buying puts if the move is below the range. And the sellers of options are happy to oblige because those sellers, being the stronger hands, know that they and their friends can move the market back toward the range prior to expiration.
Does it work every time? No. And when it works, it sometimes produces a small return. But I recall concluding back in the early 1990’s that there was a 100% probability that the price range would be exceeded one way or the other sometime during the option period and that there was an 80% probability that price would return to the range prior to expiration.
Anything to this?
Parameters
Just to clarify, the rules of the strategy we’re testing are: mark the high and low of the range for the first week after options expiration. On the first close above (below) that range, sell (buy) the underlying in the expectation that it will revert to the mean, or at least back to within that range. Exit two days before expiration to avoid gamma-fueled surprises.
Since the intent here is to sell the OTM options that get a little more expensive on the breakout, you would sell call spreads or put spreads when the system signals a trade and ignore subsequent trades, allowing those spreads to decay. But there’s nothing to prevent you from playing both sides, so we decided to allow multiple entries in a single month, meaning that if, for example, the price broke below the range and then traded back into and then above the range, we’d have two trades there, a long entry and then a short, or a short put vertical and then a short call vertical if you’re playing this with options. In our test, most of the time only one entry was signaled, and a quick look suggests that any second signals were basically a wash.
Results
As you can see, this post-expiration range reversion strategy has outperformed a buy-and-hold S&P 500 approach over the past 15 years. It missed out on the late-90s tech bubble, tracked the early-00s real estate/commodity bubble pretty closely, and has dodged the worst of the worst in 2007-8, recovering promptly after a major drawdown last month. With a profit factor (gross profit divided by gross loss) of 1.45, this system looks pretty attractive given that it is in the market less than half the time. (All results reported do not include transaction costs, dividends, or interest earned.)
General Mean Reversion
But let’s dig a little deeper: one possible objection would be that mean reversion is typical of the market in general, and that any edge obtained above is due to this general characteristic, rather than to any specific feature of the week after expiration.
Initially, that objection seems misplaced: we tested two similar systems to see whether that first week after expiration offers any special edge. In the chart below, “Exp Wk Reversion” is the primary strategy as described above, “Any Wk Reversion” trades the same strategy using a one-week lookback that is agnostic about the last expiration but still holds to the next one, and “Non-exp Wk Reversion” trades the same strategy using a one-week lookback but excludes signals generated based on the first week after expiration.
Results are initially encouraging: the non-expiration system seriously underperformed our primary strategy, with a lower profit factor and an inferior equity curve.
Conclusion
However, the combined “any week” approach performed best of all - again in terms of both relative profitability and absolute return - which suggests to us that the most important variable in mean reversion trades of this type is likely not the entry condition itself, but the amount of time that is allowed for mean reversion tendencies to kick in. If there really was any edge to a first week-only approach, we would expect the any-week equity curve to drift somewhere between the other two, instead of outperforming significantly.
As to the hypothesis about momentum traders overreacting to post-expiration breakouts/breakdowns and providing nice opportunities for these contrarian plays: that hypothesis could be a consistent occurrence in the market and yet be easily overwhelmed by other more potent forces. And in the absence of any discernible edge from trading breakouts of the week after expiration versus other weeks, we don’t see much justification for accepting the hypothesis.
The silver lining, though, is that all three contrarian systems easily outperformed the benchmark over the past 15 years, which leaves us wondering why anyone still has money parked in any long-only index funds.
