I think we agree that timing helps, at least it does in the backtest. We are using different metrics, and so have different conclusions about the timing parameters. rayvt prefers 25th percentile CAGR cycle, maximum drawdown, Sortino Ratio, and Standard Deviation. I use average CAGR cycle, GSD, and LDDD3. Our trading rules are also differenct, with rayvt using a monthly screen, and me using a daily index.

Just looking at the results from 2006 to 2026, rayvt found that the 4 timing schemes had similar results: increasing CAGR from 24% to about 27%, and decreasing Stdev from 28% to 25%.

I found (using different timing schemes than rayvt), that three of the sma timings did not change CAGR much, and decreased GSD from 22 to 19. One of the sma timings I tested decreased CAGR from 16 to 12.

From this, I would conclude that a random timing system will probably decrease GSD and have an uncertain effect on CAGR. The reason for this rather inconclusive result is the short backtest. A longer backtest is needed to evaluate these types of timing systems that are trying to detect rare events. 20 years is not long enough. We can't chop up the time series or use rolling returns to evaluate these timing systems, because we don't have enough data. Some other method is needed to avoid overfitting.