musselmant wrote: "You can just pick all Nasdaq100 stocks and see on gtr that 325 days beats 200 for a timing test on SPY."

I confirmed musselmant's finding. Aussi posted a gtr1 screen with sma timing in MI post 5735, with various buy and sell cutoffs. I used the cutoffs Aussi identified (1.0 0.9 and 0.95 0.89), and sorted the results by CAGR/LDDD3.

Results from 19721214 to 20251128:

CAGR/LDDD3        Screen        CAGR  SAWR  GSD   LDDD3  MDD  UI  Sharpe  AT
   1.52     N1T_sma325_100_090  14.7  9.0   21.0   9.7   -42  13   0.59   0.2
   1.44     N1T_sma325_095_089  14.8  9.0   21.7  10.3   -45  14   0.58   0.2
   1.37          N1T_BCC0       16.3  7.9   21.9  11.9   -72  25   0.65   0.6
   1.36     N1T_sma200_095_089  14.3  8.4   21.3  10.5   -50  14   0.57   0.3
   1.18     N1T_sma200_100_090  12.8  7.8   20.9  10.9   -49  15   0.50   0.3
   0.90             N1T         13.6  6.2   25.4  15.1   -81  31   0.49   0.0

In the longest backtest with N1T, timing using 325 day sma beats 200 day sma, with a higher CAGR and lower risk metrics. Timing using BCC=0 has the highest CAGR but also significantly higher risk metrics.

Results from 19991231 to 20251128:

CAGR/LDDD3  Screen              CAGR  SAWR  GSD   LDDD3  MDD  UI  Sharpe  AT
   1.36     N1T_sma325_100_090  12.5  7.8   18.6   9.2   -38  14   0.67   0.2
   1.28     N1T_sma325_095_089  12.6  7.7   19.2   9.9   -38  14   0.66   0.2
   1.06     N1T_sma200_095_089  11.4  6.7   19.5  10.8   -50  17   0.60   0.4
   0.77     N1T_sma200_100_090  8.9   5.6   19.3  11.6   -49  18   0.47   0.4
   0.65          N1T_BCC0       10.2  4.0   21.2  15.7   -72  36   0.51   0.7
   0.44             N1T         8.4   3.0   26.3  19.2   -81  44   0.39   0.0

A 26 year backtest that includes the 2000 and 2008 crashes also shows 325 day sma better than 200 day sma. Timing using BCC=0 does worse in this time period, with lower CAGR.

Results from 20051230 to 20251128:

CAGR/LDDD3  Screen              CAGR  SAWR  GSD   LDDD3  MDD  UI  Sharpe  AT
   2.22          N1T_BCC0       17.3  12.5  18.2   7.8   -29  9    0.95   0.7
   1.78     N1T_sma325_100_090  15.4  10.1  18.3   8.7   -34  12   0.85   0.2
   1.71     N1T_sma325_095_089  15.7  10.9  19.1   9.2   -38  11   0.84   0.2
   1.64     N1T_sma200_095_089  15.2  10.8  18.9   9.3   -37  11   0.82   0.4
   1.38             N1T         15.8  10.2  21.6  11.4   -54  12   0.77   0.0
   1.07     N1T_sma200_100_090  11.5  8.6   18.7  10.8   -38  13   0.64   0.4

A 21 year backtest that includes the 2008 crash also shows 325 day sma better than 200 day sma. Timing using BCC=0 does better in this time period, with higher CAGR.

gtr1.net: GTR1 Backtester - N1T
gtr1.net: GTR1 Backtester - N1T_sma200_095_089
gtr1.net: GTR1 Backtester - N1T_sma200_100_090
gtr1.net: GTR1 Backtester - N1T_sma325_095_089
gtr1.net: GTR1 Backtester - N1T_sma325_100_090
https://gtr1.net/2013/?~N1T_BCC0:h1::trp%281,1%29n...