Reply

How Hard is It to find HighTimeFrame Strategies

35 replies

gentmat

Customer, bbp_participant, community, 234 replies.

Visit profile

8 years ago #114043

I just wonder who finds Strategies for 1H , 4H + … How do they find something reliable . 

 

Each time i use H4 and Daily TimeFrame (DATA of 10 – 12 years) , I ll get something around 150-250 Trades for 10 years . 

 

But the problem with 200 Samples is that it can be achieved by chance nothing more or less . I think everything less than 1500 trades is a bit useless Sample to work with but with H4 and daily im never exceeding 300 Trades with good equity . 

 

 

 

1- How do you guys work with high timeframe and how much trade you accept to achieve per year . 

2- Which Number of sample is considered  bit reliable for H4 or D1 

 

Best regards,
Maroun 

0

geektrader

Customer, bbp_participant, community, 524 replies.

Visit profile

8 years ago #131983

1) Sure it´s a intrabar backtest, but still, the trailing only adjust itselfs on bar open, regardless if you run a intrabar backtest or not. That´s how SQ and the resulting MT4 EAs do it (just check the SQ / MT4 backtests and you´ll see). So trailing stop is only adjusted on each new bar, not intrabar. However, once activated, the intrabar backtest checks if the new SL of the trailing stop is hit intrabar of course.

2) Yes, I don´t know, but I have written my own EA that runs aside to all other EAs that controls the spread and in case of a to high spread, deletes all stop orders if they are about to being triggered.


🚀 Unlock Your Edge in Automated Forex Strategy Development 🚀

Historical Forex Data Starting From 1987, 28 Pairs, M1, 99% Error-Free, Lifetime Free Updates

0

gentmat

Customer, bbp_participant, community, 234 replies.

Visit profile

8 years ago #131984

Wow thats smth new i have to investigate !!!
By coding i check close[0] which is price of this tick… if price higher than x then modify order . trailing is just modify TP ! So by theory and coding it should be instantly modified. But ill check it anyway im just chocked about that.

2- nice idea to delete pending orders.

3- i want to ask if the strategy you showed was like that before optimizing parameters (cause i
never had such great curve before. and do you change parameters value for the same strategy but for different symbols ?

0

geektrader

Customer, bbp_participant, community, 524 replies.

Visit profile

8 years ago #131995

Not much to investigate, any stop adjustment is done on bar open only in SQ, regardless if backtested with ticks or not or whatever;) You can already see that in the MT4 code it exports:
 

—-

void manageStop() {
 
    if(!sqIsBarOpen) return;
 
And sqIsBarOpen gets determined like that:
 
   if (tmp!= Time[0]) { // } || open != Open[0]) { – this doesn’t work with renko charts
      tmp =  Time[0];
      open = Open[0];
      sqIsBarOpen = true;
   } else {
      sqIsBarOpen = false;
   }
 
^ This hence cancels the stop adjustment function which is responsible for ANY stop adjustment (stop trailing, profit trailing, setting BE, etc etc) if the current tick isn´t the open of a new bar. So it get´s only executed on the start of each new bar, depending on your timeframe. It´s no problem though as otherwise backtests would be slow and would look TOTALLY different between bar open mode and every tick mode, which wouldn´t make much sense.
 
In this relation I also have to say that the “Profit Trailing” SQ3 does, is no real profit trailing. It trails ALWAYS at the start of the next bar, regardless if the order is in profit or not, they don´t check for “Profit > 0”. I found that strange at first, but backtests between SQ MT4 match in that relation, so it´s wanted like this by Mark it seems. See:
 
         // profit trailing on close
         trailingStopValue = getProfitTrailingByTick();
         if(trailingStopValue > 0 && OrderSelect(currentTicket, SELECT_BY_TICKET)==true) {
            if(OrderType() == OP_BUY) {
               tsLevel = close – trailingStopValue;
            } else {
             tsLevel = close + trailingStopValue;
            }
            orderSL = OrderStopLoss();
            normalOrderSL = getNormalSL(orderSL);
            newSL = getSpecialSL(tsLevel);
 
            if(OrderType() == OP_BUY) {
               if(isSLCorrect(tsLevel) && (orderSL == 0 || normalOrderSL < newSL) && !doublesAreEqual(orderSL, newSL)) {
                 rettmp = OrderModify(OrderTicket(), OrderOpenPrice(), newSL, OrderTakeProfit(), 0);
              }
            } else {
              if (isSLCorrect(tsLevel) && (orderSL == 0 || normalOrderSL > newSL)  && !doublesAreEqual(orderSL, newSL)) {
                 rettmp = OrderModify(OrderTicket(), OrderOpenPrice(), newSL, OrderTakeProfit(), 0);
              }
            }
         }
 
 
^ There is absolutely not check for the profit trailing if the order is in profit or not, it will just start the profit trailing ALWAYS next bar after your trade was opened. Also seen that live several times, even if the order was at -20 pips at the open of the next bar after it got opened, SQ starts to trail with the profit trailing distance.
 
Quite a few strange things, but the strategies that are generated this way are all good, so I won´t complain:)
 
 
Yes, the strategy was almost like that when it was found by SQ, but this just happened ONE time, never again did it find that strategy again. Was very big luck:) And it also never ever generated a similar strategy afterwards too in the last months of steady generating. So seems I got very luck having found this “almost” holy grail;) I am currently adapting it to M5 on 14 years of data, looks totally amazing as well, over 8000 trades it does and optimization is not even half through yet;)


🚀 Unlock Your Edge in Automated Forex Strategy Development 🚀

Historical Forex Data Starting From 1987, 28 Pairs, M1, 99% Error-Free, Lifetime Free Updates

0

gentmat

Customer, bbp_participant, community, 234 replies.

Visit profile

8 years ago #132001

Oh i see thanks for the explanation . Well your right its a sword with double side but now if u move to 15 mins it will be the good side (as i like on bar open trailing idea more than tick , it gives more edge for the mood ) .

You missed one quest that i would like to know that will help me a lot with researching strategies wich is:
when your using this strategy on lets say usdcad or any other symbol that u said it works on multiple
* is it giving same great curve or u are optimising the parameters of indis and st and tp . so in summary all symbol use same or different values of parameters.?

Thank you for all the help your help so far helped me a lot and ill be updating here in case i succeed

0

geektrader

Customer, bbp_participant, community, 524 replies.

Visit profile

8 years ago #132013

Yes, it´s very interesting that it trails on bar open only, because M30 and M15 and M5 all work with that scalper, all deliver good results, but trade at different times and with low correlation, so that will be interesting for my new portfolio of scalpers. Currently working on that.

 

I am using a custom optimization for each pair on 14 years of data, seems to work best so far live forward. Although for that system, the ideal parameters for each pair don´t really look a lot different between pairs, which underlines the stability of that system.


🚀 Unlock Your Edge in Automated Forex Strategy Development 🚀

Historical Forex Data Starting From 1987, 28 Pairs, M1, 99% Error-Free, Lifetime Free Updates

0

Viewing 5 replies - 31 through 35 (of 35 total)

1 2 3