Naked Put Study 2 (Part 6)

Last time I studied longer periods of the equity moving average (MA) filter and settled on 75.

My biggest worry is still a significant slippage setback. Depending on period, 103-159 crossovers were seen in 16 years. Each crossover represents a trade. This would require either closing/reopening the entire inventory of naked puts (NPs) or buying long puts to neutralize. The former is likely to be ravaged by slippage. The latter could be done with significantly fewer contracts but these expensive puts will have a much wider bid/ask spread.

Ideally I want something more akin to the 10-month equity MA. Dating back to 2000, this indicator has supposedly triggered a handful of times and avoided some large market corrections. As previously discussed, something triggering a handful of times is exactly what I need. The problem with the 10-month SMA is that it only triggers at end of month. The NP strategy can lose huge sums of money in days; it demands a more reactive indicator.

Perhaps I could use a MA of underlying prices rather than account value. At the least, this would prevent longer periods from cannibalizing the beginning of the data set since I have access to underlying prices dating back to last century. At most, underlying prices are easier to download than account values are to backtest. Underlying prices are also less volatile than account value due to the price/IV (implied volatility) double whammy. Price declines and IV expansion can both hurt NPs.

In the case of underlying price MA, perhaps I increase the period as much as necessary to halve (at least) the number of crossovers. I could then study efficacy of the filter. This does, however, remove optimization from the process because I am putting an external constraint on period. I believe optimization should be done to prevent fluke. What’s worse is that external constraint is forcing a small sample size, which may not be reliable.

My next steps are to study number of underlying price crossovers by period and to think through a maximum adverse excursion (histogram) study.