That's a better setup than the one proposed by RoughTrader, but it still produces a $1860 loss over the last 12 months on 221 trades. Here is a 3-day sample:
Let's try this: 1) Start at 9:30 eastern 2) Enter long when a) market makes a new intraday high and b) price is above it's 20 period EMA on a 5 min chart. 3) Inverse for short entry. 4) Exit long when price moves below the EMA. Exit short when price moves above the EMA. 5) Diversify across different markets such as: S&P, T bonds, Crude, Coffee, etc. 6) Close all open positions prior to market's closing time.
@RoughTrader i have codet this strategy in easy language and performance is indeed loking promising, but i am not able to let the skript know that after the the takeprofit it should remain flat untill next trading-session, i am getting 2 orders at the same bar; the takeprofit and emediatly the same bar the next long position. tried things like if marketposition = marketpositionflat then.... any thoughts....? will post results later (if succeed)
@nonlinear5 check your code again, if you wrote the skript as roughtrader has explained it should be profitable maybe try a different underlying below doesnt incl tp's just plain buy at h sell at l a simple stopandreverse untill endtime (5pm) still working on it
I'm a little confused. Have you guys tested what I described in my last post where a new intraday high or low is used for entry and the EMA is used for exit?
this is what ive get after includig commision, i think the basic is ok although it does indeed have his drawdowns.roughtrader has explained he uses a different entry, does this reply to only the first entry of the session or to all? i have finally succeed integrating the 4-step tp but according to my backtest it doesnt make it more profitable the best i could get untill now is integrating a sl and let the system trigger new trades all session long by new 10bar highs and lows untill endtime (5pm) is achieved an average of 6,2 trades/session incl shorttrades over 36months
the basic rules allready give a profitable signal by just buying last bars high and selling last bars low from 9am till 5pm, ow have to find a way to maybe reduce the avg trades/session and the drawdown if barcounter > maxitems(N_High,N_Low) and time >= Start_T and time < End_T then begin if high > H_Level[1] then SetUp = 1 else if low < L_Level[1] then SetUp = -1; end; if marketposition = 0 or marketposition = -1 and SetUP = 1 then buy ("Long") next bar at H_Level[1] + 1 point stop else if marketposition = 0 or marketposition = 1 and SetUp = -1 then short ("Short") next bar at L_Level[1] - 1 point stop; if time >= End_T then exitposition ("GameOver") next bar at market;
dutch tradesignal will not give reliable backtest results for code that uses a "buy at price + something of the previous bar stop" instruction rather the plain buy at market instruction. they warn about this in their wiki but i think there is a huge flaw in their approach ats design anyway. my recommendation, for the purpose of backtesting this or any other strategy, is that you model it to execute orders at market based on a given event, say the crossing the support or resistance line, as this will prevent the behaviour above.