Report: HIGHLY RANKED STRATEGIES THAT CORRELATE NEGATIVELY TO THE S&P 500
After reading some comments on another thread, I was thinking that it might be time to highlight some strategies that have negative correlations to the S&P 500. I confess that part of my motivation is that I run two of these negatively correlated strategies, both involving Bitcoin and Blockchain. By comparing strategies, one can perhaps see some of the relative advantages and disadvantages of each strategy.
The point of this exercise is not to suggest safer individual strategies, but rather to take advantage of diversification to create safer combinations of strategies. As Modern Portfolio Theory and CAPM suggest, investors are not rewarded for taking idiosyncratic or firm risk, since that can be mostly diversified away by making investments that are less correlated. Theoretically, diversification allows about the same expected average return at significantly less downside risk.
METHOD:
Late on Friday evening, Feb. 2, I downloaded the top 100 strategies on the Grid listed by C2 score. Of these 100 strategies with C2 scores of at least 76.4, 15 had negative correlations to the S&P 500.
Two of these (CkNN Algo and VIX Club - Short Vix) were excluded from my analysis because they had negative both 30-day and 60-day performances. That, of course, does not make them bad strategies, just less attractive to those contemplating a switch this week. A third (Futures Proxy) was excluded for reasons that are explored on another thread ( Futures proxy - a proxy to vix arb? ).
I am not, however, endorsing any of these strategies, and I am not excluding at least one strategy with a worrisomely high win rate.
RESULTS:
In the first table, the resulting 12 strategies are ranked by correlation to the S&P500:
As you can see, among the top 100 strategies in C2 score, “forexland” has the largest negative correlation to the S&P 500 (-.125). If weekend days are included in C2’s stats (they probably are), these negative correlations are probably functionally a bit larger (which is a “good thing”).
While Table 2 ranks the same strategies by Total Return (or annual return for the 2 strategies over a year old), Table 3 ranks them by Maximum Drawdown. The most striking thing here is that Payoff Matrix has the lowest total return of all 15 strategies (22.5%), but it also has the lowest DD (only 5%). The strategy “Wizard EUR/USD Forex” has the highest total return (157.3%).
Table 4 shows the same 12 strategies with negative correlations to the S&P 500, ranked by annualized return (with trading costs) and includes the C2 Score, Subscription Fee, and the last 30, 60 and 90 day returns according to the Grid late Friday evening.
As you can see in Table 4, my two Bitcoin Related strategies have the highest annualized returns in the group, but like Payoff Matrix have been operating for only 1.5-2 months. Because I think that Total Return is more important than Annualized Return, Total Return is the one column that I added by hand to the data that I downloaded from the Grid.
The strategy with the HIGHEST subscription fee ($497) is Payoff Matrix, which also has the LOWEST total return (22.5%) in the group–and the LOWEST drawdown (5%). The strategy with the lowest subscription fee is “100 Percent Win,” which is free.
NOTES:
-
The analysis here was based on C2 ratings and S&P 500 correlations on the Grid late Friday evening, Feb. 2, 2018. Strangely, the correlations with the S&P 500 have changed on the Saturday, Feb. 3 Grid in a way not consistent with merely adding another day’s data to the mix. If I had done the analysis on Saturday, the strategy COREX would have joined the list and most of the other data would have changed somewhat.
-
I plan to add my personal comments on my Bitcoin Related strategies in a later post in this thread, rather than burden the relatively straightforward descriptive analyses in this post.
