Quote from frutza on May 19, 2022, 4:09 pm

i think it is because sp500 is not mean reverting. will try it on some mean reverting stock, if i can find any.

Quote from TraderJoe on May 19, 2022, 4:28 pm

I tried again, this time from 2005 to 2010. At the end of 2010, the SP500 reverted to the price level of 2005. It kind of worked this time, although the annual return is only 6%. This strategy may have potential, but I think it is too simple, and it does not trade enough. At least not with these parameters. Any ideas?

Quote from frutza on May 23, 2022, 11:54 am

here is a backtest of the strategy on S&P 500 VIX Cash (VI.C) on historical data from stooq. full code:

function run()
{ set(PARAMETERS + LOGFILE + PLOTNOW);
BarPeriod = 1440;
MaxLong = MaxShort = 1;

StartDate = 19970101;
EndDate = 20220620;

var opt = optimize(2,2,50,1);

var overbought_level = 70;
var oversold_level = 30;
var rsi_period = 4;
var adx_period = 20;
var adx_level = 20;

vars price = series(priceClose());
vars rsi = series(RSI(price, rsi_period));
vars adx = series(ADX(adx_period));

if(adx[0] < adx_level) {
if(crossOver(rsi, oversold_level)) enterLong();
if(crossUnder(rsi, overbought_level)) enterShort();
}

plot("RSI", rsi, LINE|NEW, BLUE);
plot("Overbought", overbought_level, LINE, BLACK);
plot("Oversold", oversold_level, LINE, BLACK);
plot("ADX", adx, LINE|NEW, GREEN);
plot("ADX Level", adx_level, LINE, BLACK);
}

i changed the ADX formation period from 7 days to 20. here's the report.

Annual return 8%
Profit factor 1.49 (PRR 1.27)
Sharpe ratio 0.20 (Sortino 0.19)
Kelly criterion 0.48
Annualized StdDev 41.18%
R2 coefficient 0.414
Ulcer index 17.3%

this strategy is VERY dependent on the ADX formation period, so handle with care!

Quote from Algo Mike on July 16, 2022, 1:46 pm

frutza, you are correct. The strategy depends heavily on a single parameter, which makes it a good teaching tool, but not a good strategy for live trading.