i realise that this strategy only worked well in 2012, but before that, it performed badly.
Any thoughts as to why this is the case?
Market changes all the time shifts and adapts. This strategy potentially can work in many markets, but perhaps on different time frame. I would say that this strategy works very well in strong up trend, not so well in ranging market. I am waiting for Robopips finding on Wednesday to be published, should be a good read:)
what is the recommended time period that a strategy should have been proven to work on backtest to give sufficient confidence for the future?
Maybe we can add in a trend filter to the strategy.
I though I was the only one not able to get the EA to work properly in backtesting. I use defaults and backtest the last 2 years and get severe losses. Before I saw crisscrossâs EA I programmed one of my own which is much more simplified, without all the bells and whistles, just the essentials but it also ends up with a loss.
How is the EA performing in forward testing, is it significantly different from backtesting (i am asking for this particular case, i know that usually it is)?
Well so far its been profitable esp for Sep. Time will tell i guess. With back testing the best i managed is profit factor of 1.1 going back 2 years. It interesting to note that with BE value of less then 22 pips and tsp turned of the system is profitable in long back tests, but if you narrow the values down then it not profitable. Theoretically it should be other way around ? My only guess is when we have small values in a ranging market the system make small amount of pips with safe values however should we have a 100 pip day the system canât take full advantage of it as its got no room to breath sort of speak, hence S/L get triggered. In conclusion the losses outweigh gains, hence why small values are not optimal, we want to net max amount of pips, to balance out our losing day. Hope i am making sense cos i think i confused my self./. 
Hello Pipcompounder, KIllerNo & Mostwantedpip,
I have read all your comments and noted them. I am going to start checking all these errors and bugs and the logic thoroughly once again in the next rev and come back within 3 days. Thanks for your comments.
The EA is still a long way from being perfect but it will get there soon.
Safe trading.
Hi,may i have a try in detail?
I took the London Breakout strategy, modified it & coded an EA. I took away the BE & TS part. Seems to be more profitable that way as we are letting the trade run and closing all open orders that has not hit the initial SL or TP at a specific time later.
BTW, iâm not sure if entering multiple orders after getting stopped out is a good idea: Might get chopped around in a ranging market.
And multipe TPs doesnât seem to be as profitable as a high single TP; unless of course you are intending to adjust the TPs manually.
Hi KillerNo,
I understand your point. The EA opens pending orders DistBuy or Distsell pips away from the RefBar. After London Open hour the refbar size is increased but the EA still uses the same logic to open pending orders.
I can implement your logic to place pending order after Londonopen hour at the same price as Refbar+RefpipHigh or Refbar-Refpiplow but i worry this will fall within the range of Freeze level of broker and will prevent pending orders from being placed.
So, I still suggest to keep the Distbuy and Distsell and you can try to reduce the values of these two variables to say 1 pip each and test.
What is your opinion?
hi crisscross1983,
thanks for the answer. how do you mean,that will fall in the range of freeze level?EA can not place a new pending order if price is too close to this level?
Actually if I reduce the distsell and distbuy levels as you suggested then only the first pending orders will be placed close to the ref candle high/low after gonna be in refpip distance, so it is not a big issue.
could you please give your e-mail address, I modified the TSL and breakeven codes, check it maybe gonna be alright.
one more question 
If I am using the EA on more Pairs, do I have to use different magic numbers on each pairs or only on the same Pair e.g. same EA running on EUR/USD with different settings.
I was running today the EA on 3 EUR/USD charts with different settings and with different magicnumbers(I modified them before) an realised that I set up 20pips trailing for the first chart and 7pips for the second and the EA was using 7pips trailing
for the first one as well. What else need to change to get working properly.
EDIT: it looks like is an issue only if I am running the EA on the same Pair e.g on 2 or 3 EUR/USD chart. I think the EA does not check the magicnumbers in the ModifyOrders() function. Crisscross1983 could you confirm it?
thanks
regards
20 pips in the bag today so far.
I think some people that backtest for more than a month back wonât get accurate results because of the modeling qualityâŠ90% isnât even good enough, in my opinion. To me, what matters is making the EA do what we want, so, for example, what I want is if price moves to open three trades, and then retraces back a little negative to where the entry was, I want those orders closed and pending orders placed again, even though it will be a small loss. This way, if the price is outside the reference candle there is always trades opened until either BE or TP is hit, and if price is inside reference candle for that day, no trades are open, but the pending orders are placed, ready for breakout. I think if we can make this logic come to life, this idea will workâŠdoes everybody see what Iâm getting at?
hi,
I understand you. If we could move the Lockinpip to negative value, then you can do everything what you want with this EA, but trailing stop is not working I think if Lockinpip is negative.
But I think with those options what I mentioned before would be more effective the EA.
Hello KillerNo,
Btw I am going to change the logic of the TS routine.
Running the EA on three different pairs should not be a problem with the same magic number (I think) but with the same pair on the same MT4 platform will definitely be a problem. So better change magic numbers if you are using on the same pair.
Hi Pipcompounder,
The next revision of EA will be able to place as you want it to.
Then we will test it again.
Safe Trading.
hi crisscross,
thanks for the answer.I had a problem running the EA on the same pairs with differtn magic numbers as well on the trailingstop part.
I think in the ModifyOrders() function we have to check the magicnumbers as well otherwise the EA dosent know which settings to use for example for EUR/USD par.
Maybe here should add:
if(OrderSymbol()==Pair) && OrderMagicNumber() == âŠ
what do you think?
Hi Criscross.
The EA has been failing to place 3 buy orders from ref candle and also when it places sell orders it gives me a pop up window saying error setting trailing stop loss. So It places the sell orders fine, but fails to place buy orders⊠As mentioned before if there is anyway i can assist you in solving this I.E. error report or anything else like a screensot then by all mean i will be happy to help.
I am wondering if anyone else has experienced this problem 2?
After initial sell SL was hit, the buy has reached 40 TP, with the possiblilty of the last trade going to 60 TP, if price still goes up.
Originally Posted by killerno
EDIT: I found one more bug in the EA. e.g. If I set the DistBuy and DistSell values to 3pips and the RefPipHigh and RefpipLow values to 3pips ,the EA gonna place at LondonTime:59 minutes the first peding orders to ref. candle ± 3pips.
this is alright. If price hits SL and goes back to the extended ref. zone in this example ref. candle±3pips it is gonna open new pending orders again to ref. candle ±6pips (DistSell/DistBuy+RefPipLow/RefPipHigh) instead of ref. candle ±3pips.And thereafter every new pending orders will be placed to ref. candle ±6pips.
please crisscross1983 check this bug as well in the EA if you have time.
CRISSCROSS:
thanksHi KillerNo,
I understand your point. The EA opens pending orders DistBuy or Distsell pips away from the RefBar. After London Open hour the refbar size is increased but the EA still uses the same logic to open pending orders.I can implement your logic to place pending order after Londonopen hour at the same price as Refbar+RefpipHigh or Refbar-Refpiplow but i worry this will fall within the range of Freeze level of broker and will prevent pending orders from being placed.
So, I still suggest to keep the Distbuy and Distsell and you can try to reduce the values of these two variables to say 1 pip each and test.
What is your opinion?
Does this mean that in effect, if the BE SL is hit, and the RefpipHigh \Low is far enough away, pending orders are already going to be placed immediately? Even though the reference candle grows, there wouldnât have to be a loss? This may be what we want, also if price were to retrace back into reference candle zone and then just retest the entry point prior to going all the way opposite that entry, then the actual entry point would be moved just out of reach, which would keep a potentially bad trade from opening, right? Maybe I need to do some backtesting, just changing the RefpipHigh \Low. Please tell me if this makes sense.
Question about the EA settings vs the strategy. I have the âkeep placing ordersâ for 23 hours⊠what are you setting that too?
for instance this morning I hit both buy and sell stops, which stop lossed, then the EA created the orders again which the buy I rode for the pip gain.
Is this the strategy to keep the orders coming every time the order point is crossed? I question this as if the trend stays flat the system could potentially trigger over and over for a large loss.
Iâm still pretty new so forgive me if i donât have the terminology to explain this.
I have my settings at 4 hours after the reference candle ends, so it will catch reversals during first part of london session. I agree, if left all day, it could range out the account, so to speakâŠthere is a balance to this part. If we didnât have a reentry for a little time, today would have been all losses.
Does anyone care to share all of his settings for the EA? And also has anyone tested multiple settings, observing success rate etc?