Subject: Re: Timing Using Moving Averages
It can be confusing to try to compare various combinations of timing parameters to decide which you would want to run a real-world portfolio with.
The timing signal is the SMA (simple moving average) of the S&P500 index.
I looked at two ways to compute the SMA over the same calendar period. One was to take the moving average of the daily prices, the other was to take the moving average of just the end-of-week prices.
The time period was the same, just the price granularity was different. The length of the SMA calculation was 325 market days (15 month, 65 weeks) and 252 market days (12 months, 52 weeks).
There was not a consistant significant difference in the results of computing the SMA daily or weekly.
The trading method was using the weekly SMA and trading on the first of the month according to the IN/OUT of the most recent signal. No trades are done until the next month; there is no "early sell" or "early buy" in a month.
The next thing of interest was seeing the differences in the monthly signals between 325 day and 252 day SMA, at three different percentages below the SMA for the sell signal. The buy signal is always the price at or above the SMA.
Summary:
* The lower the SMA sell lthreshold is, the fewer the OUT months are.
* The lower the SMA sell lthreshold is, the fewer the differences in IN/OUT months between the 325 day and 252 day SMA.
* The lower the SMA sell lthreshold is, the more time is spent in the market.
* There is little difference in the CAGR return, for any of the 6 combinations of SMA period and sell threshold.
The return data is for one particular cycle of the 21 overlapped cycles. There is no reason to believe that the relative CAGR rankings would be the same for other cycles.
Conclusion:
* Since every combination of parameters has approximately the same CAGR, might as well use the combination that is simplest to compute and has the fewest trades.
---- TABLES ----
Timing signals are SMA of end-of-week values. Trade on the first of the month.
Sell at 0% below SMA
325 day 252 day 325 day 252 day Date 325 day 252 day
SMA % 0% 0% Not same signal 8/1/1988 out IN
# in 8158 8025 Days Months 10/3/1988 out IN
% in 79% 78% 425 19 12/1/1988 out IN
# sells 41 55 2/1/1990 IN out
CAGR 26.6% 25.3% 5/1/1990 IN out
6/1/1994 IN out
8/1/1994 IN out
9/1/1998 IN out
3/1/2000 IN out
6/1/2000 IN out
8/3/2009 out IN Since 2006
7/1/2010 IN out Since 2006
1/3/2012 IN out Since 2006
4/1/2016 out IN Since 2006
7/1/2016 out IN Since 2006
3/2/2020 IN out Since 2006
2/1/2023 out IN Since 2006
11/1/2023 IN out Since 2006
4/1/2025 IN out Since 2006
Sell -1% below SMA
325 day 252 day 325 day 252 day Date 325 day 252 day
SMA % -1% -1% Not same signal 8/1/1988 out IN
# in 8259 8128 Days Months 10/3/1988 out IN
% in 80% 79% 341 16 12/1/1988 out IN
# sells 45 45 3/1/1990 IN out
CAGR 27.4% 26.6% 5/1/1990 IN out
6/1/1994 IN out
8/1/1994 IN out
3/1/2000 IN out
8/3/2009 out IN Since 2006
9/1/2010 IN out Since 2006
1/3/2012 IN out Since 2006
1/4/2016 IN out Since 2006
3/2/2020 IN out Since 2006
2/1/2023 out IN Since 2006
11/1/2023 IN out Since 2006
4/1/2025 IN out Since 2006
Sell -4% below SMA
325 day 252 day 325 day 252 day Date 325 day 252 day
SMA % -4% -4% Not same signal 8/1/1988 out IN
# in 8445 8533 Days Months 11/1/2000 IN out
% in 82% 83% 212 7 4/1/2002 out IN
# sells 25 27 8/3/2009 out IN Since 2006
CAGR 26.6% 26.9% 1/3/2012 IN out Since 2006
9/1/2015 IN out Since 2006
2/1/2023 out IN Since 2006