The RSMA42 timing system goes to cash when the recent price is more than 10% below the price a year ago. The prices are smoothed using a 2-month moving average. Results from 19731231 to 20251128 when applied to the daily index N1T:

      Screen        CAGR  SAWR  GSD   LDDD3  MDD   UI   Sharpe  AT
    N1T_RSMA42      16.4  10.3  21.1   8.5   -46  12.6   0.67   0.2
N1T_sma325_100_090  15.4  9.1   21.1   9.4   -42  13.1   0.63   0.2
     N1T_BCC0       17.1  7.8   21.9  11.7   -72  25.5   0.68   0.5
        N1T         14.4  6.2   25.5  14.8   -81  30.9   0.52   0.0

{N1T_RSMA42} has the best LDDD3, UI, and SAWR over this timeframe. CAGR and Sharpe are slightly lower than {N1T_BCC0}. {N1T_RSMA42} has lower turnover (fewer signals) than {N1T_BCC0}.

Annual returns in years with a loss:

Year  N1T  RSMA  BCC0  sma325  rank N1T  rank RSMA  rank BCC0  rank sma325
1974  -23    0    13     8         4         3          1           2
1975  17    -1    16     -1        1         4          2           3
1984  -11   -8    -5     -9        4         2          1           3
1987  13    12    13     -7        1         3          1           4
1990  -7   -14    -9     -17       1         3          2           4
2000  -40  -28   -40     -15       3         2          3           1
2001  -26    4   -22     4         4         2          3           1
2002  -35    2   -22     2         4         2          3           1
2008  -41   -9   -17     -27       4         1          2           3
2011   2     2    -3     2         2         1          4           2
2022  -33  -21   -27     -22       4         1          3           2
 
                         avg     2.91      2.18       2.27         2.36

RSMA is N1T_RSMA42
BCC0 is N1T_BCC0
sma325 is N1T_sma325_100_090

BCC0 was the best between 1974 and 1990. sma325 was the best in 2000. RSMA was the best 2008 to 2022.

N1T gains when {N1T_RSMA42} was in cash:

  Date      N1T     Date2     N1T   days  Gain
19731231   0.82   19740411   0.87    71    7%
19740523   0.82   19750418   0.76   228   -7%
19820401   1.95   19820826   2.16   102   11%
19840521   3.00   19840917   3.02    82    1%
19871218   4.36   19880311   5.01    57   15%
19900926   5.45   19901217   5.97    57   10%
20001122  76.00   20020214  44.83   304   -41%
20020319  44.90   20030430  34.30   281   -24%
20080819  53.77   20090817  44.68   250   -17%
20220608  399.11  20230327  399.33  200    0%

{N1T_RSMA42} had a few misses 1982 to 1990, but the timing has done well since 2000.

The {N1T_RSMA42} timing system goes to cash when the recent price is more than 10% below the price a year ago. The prices are smoothed using a 2 month moving average. {N1T_RSMA42} uses three parameters: the 10% limit, a 1-year lookback, and a 2-month moving average.

I chose the 10% limit knowing that the system needed to stay in the market most of the time, and wanting a round number that had human appeal. Some traders might want to close a position after a 10% loss. I didn't test any other limits.

Many momentum systems seem to work with about a 1 year lookback, and people often work on a yearly calender. I didn't test other lookbacks.

I did a mount-of-toast test to find the 2-month averaging window. The GSD is stable for windows from 1 to 90 days, and is higher for windows longer than 90 days. The CAGR is roughly level for windows from 30 to 80 days, and slightly lower outside this range. 42 days is near the middle of this 30 to 80 range and is an even 2 months. The GTR1 code for the parameter study is:

Mo: ratio(sma(1,param0),sma(param1,param0))
where param1 = 253 - param0

gtr1.net: GTR1 Backtester - N1T_RSMA42
gtr1.net: GTR1 Backtester - N1T_sma325_100_090
https://gtr1.net/2013/?~N1T_BCC0:h1f0.1::trp%281,1...
gtr1.net: GTR1 Backtester - N1T