Here is an indicative backtesting research:
Over the whole Nasdaq 100 database, from Jan2000 till now, settings buy/sell/short/cover at Open, delay +1, commission 0.5% the total % Net profit was a flat +5%.
The Morning/Evening star are impressive buy not that profitable in the long term.
Some stocks [like FLEX or CIEN] were interesting [>+400%] but the general outlook was not important.
Here is the AFL code
// MORNING STAR
v29= CdWhite(0.1) AND Ref(CdDoji(0.1),-1)
AND Ref(CdBlack(0.1),-2)
AND Ref(O,-1)<Ref(C,-2)
AND O > Ref(C,-1);
// EVENING STAR
v28 = CdBlack(0.1) AND Ref(CdDoji(0.1),-1)
AND Ref(CdWhite(0.1),-2)
AND Ref(O,-1)>Ref(C,-2)
AND O < Ref(C,-1);
// EXIT DELAYS
x=Optimize("x",19,1,20,2);
y=Optimize("y",19,1,20,2);
// TRADING RULES
Buy=V29;// BUY WITH A MORNING STAR
Sell=Ref(Buy,-x);// SELL X BARS LATER
Short=V28;// SHORT WITH AN EVENING STAR
Cover=Ref(Short,-y);// COVER Y BARS LATER
As you probably agree, it is meaningless to wait more than 20 trading bars. Both above formations are much advertised, they are attractive to watch but the only happy person is my broker.