On thursday we had the second “big” down day in a row in SPY, with two days in a row with more than 1% each day. This does not happen very often. Knowing that SPY tends to revert to the mean: Can we make money by going long on the close of the second day? Let’s look at the statistics on this incredible simple strategy (buy when there is blood in the streets):
- I calculate an absolute value of the 25 day average of close to close, whether SPY is up or down. If it’s a down day of .35 and the next day is up .25%, the average of these two days is 0.3%. (On average SPY seems to fall much more on down days compared to the rise on up days. But I ignore that in this study.)
- If SPY falls (close to close) more than this 25 day average today, and it also fell more yesterday than yesterday’s 25 day average, then go long at the close. My experience tells me the velocity of the fall is much more important than the number of down days.
- Exit on the close after two days. I have measured other exit days, but two days seems to offer the best risk/reward. Also, I like to use “symmetrical” values. If entry is based on a two day pattern, exit should also be based on two days patterns.
This simple strategy have the following stats since January 2005 until present:
|P/L in %||#fills||#wins||Avg|
The nastiest day was an entry on the 7th of October 2008 with a loss of 9.5%. That one is a pretty hard one to swallow. Not sure how I would react with such a loser, but volatility was dramatic in this period. Here is the equity chart in %:
This strategy can be combined with the 5 day low strategy. The overlap between the two strategies is “only” 23 trades. If the 5 day low strategy trumps the two down days in a row strategy, we get the following equity curve:
What happens if you enter on the close after two big down days and SPY opens down the next day? Here is the stats those days from open to close when SPY opens 0.1% or more down:
There are 19 fills and 14 winners.
What happens if we turn this strategy upside down and go short? It does not work nearly as good, unfortunately.