Thank you for sharing that, Paul. Incredible work given the time within which you have responded. Please do continue at your leisure, as Nickola suggests.

If the emphasis here is on testing strategies that work during holiday periods, I think the focus should not be on the whole of December - as about the first half of it certainly does not exhibit holiday tendencies, in my opinion, as volume holds up okay in the earlier to middle part of the month and bullish seasonality does not take effect until later in the month. In fact, seasonally speaking, the SPX performs poorly for the first half of December (serving as a tailwind for shorts) and strongly in the second half (and this December has once again followed this pattern, almost to the letter).

In my opinion, the real holiday period in December is basically the last 5 trading days of the month, and maybe a few more trading days than that (so perhaps the last 5-7 or even 5-10 trading days of the month, for instance). This year, at first glance, does not seem to be a good example of normal year-end bullish seasonality because the last 5 trading days of the month actually resulted in a rare negative return. However, taking a step back, one would've been well served sitting long for the last 9 or 10 trading days of the month. As an aside, from what I can recall, Jeffrey Hirsch defines the Santa Clause Rally as the last 5 trading days of Dec and first 2 trading days of Jan.

Regardless, to me, your analysis demonstrates a case where the edge of the strategy may not affected by the peculiar features of holiday trading. In other words, certain mean reversion based short strategies may be unaffected performance-wise from the low volume and bullish seasonality of holiday periods like late December, generally speaking, and thereby may be the best way to approach shorting during such periods. It is not surprising if this is true especially if the mean reversion strategy being employed is one that takes profits quickly and does not sit on a position for very long at all.

What about momentum-based shorting? I would be curious how shorting based on momentum rather than mean reversion works during a holiday period. My hypothesis would be that it would more often than not be less effective than during a non-holiday period.