I had lost internet connectivity that’s why the global variables were not loading. It’s loading properly now after restarting the connection. Thanks a lot. Do you think it can still be as good as it used to be back then? Looks like it’s still compatible and can work just fine. Your thoughts?
Your guidance and patience is incredible.
Thanks so much for this. I did some testing but got terrible results. I had set the parameter for time zone as -1. I will try to change the parameters as advised above.
Are you back testing with real tick data or just using your broker’s data feed?
If you are using your broker’s data feed (or even data from MetaQuotes), then you will not get good results.
[ol]
[li]That data only has M1 granularity which is too coarse for reliable back-tests of this strategy (especially at a 5pip trailing stop).
[/li][li]The broker’s history data downloads may be very unreliable and have many gaps in it.
[/li][/ol]
So, in essence learn to collect real tick data and generate good quality HST & FXT files for the back-tests. Use for example the “TickStory” application.
[ul]
[li]Tickstory – HIstorical Data & Resources for Traders
[/li][/ul]
When generating the FXT files with “TickStory”, make sure you use the correct parameters and time zone settings and that you set both your FXT and HST files to “read-only” to prevent them from been overwritten by broker data.
Setup, a separate MetaTrader installation, specifically for back testing, so that it does not interfere with your real DEMO or LIVE trading installations.
As for the “Open Hour”, according to your broker’s details, its time zone is EET/EEST (UTC+2/3 because of daylight savings), However, the best value for this parameter is not just dependant on time-zone but also on the sessions times that most affect the currency pair in question.
[ul]
[li]For EUR/USD and EET/EEST, the best open hour is either “23” or “0” for best results, but “8” can also work well.
[/li][li]For GBP/USD and EET/EEST, an open hour of “12” is best but “22” also does well.
[/li][/ul]
Ideally, once you are back-testing with real tick data, you should run an optimisation run on the “Open Hour” and see which ones work best for the Currency Pair in question.
I am backtesting using metaquotes historical data. I will try to get the real tick data and test again. Also,my broker has a stops level of 5 for eur/usd,gbp/usd and usf/jpy. Does that mean a take profit of 5, stop loss of 5 and trail stop of 5(default) will not work properly with my broker? The spread varies from 0.5-1.4 so I am testing using a spread of 1.4. It’s a 5 digits broker.
When running the Strategy Tester, make sure you set the testing Model to “Every Tick”. Don’t use the “Control points” or “Open prices only” models.
With a stop-level of 5pips, theoretically you could continue to use a stop-loss and trailing stop of 5 pips. However, the spread and slippage will have a real-life affect on the strategy so you will have to compensate by increasing the stop size to probably 7 pips or so. Run it live on a DEMO account for a week or so and see how it is affected!
I am using ‘every tick’ model when backtesting. Will forward test on demo and see how it goes. Last question please…(will sound dumb but it’s confusing me!). The stop loss,take profit and trail can remain as 5 pips but it’s recommended that i set the stop minimum pips to 7? Is that what you mean? Or should i set all the parameters in the stops management setting to 7? There is a stop minimum pips and soft stop loss pips. Should the two have the same setting,7pips in this case for example or should stop minimum be set to 7 then stop loss remain as 5?
I am recommending that you set “Stop-Loss” and “Trailing Stop” to 7 pips. Also, please read the PDF documentation carefully for explanation on each parameter.
Hallo Mr.Carnino. I have a question please. In the report above the modelling quality is indicated as n/a. I thought that if you use ‘every tick’ model to backtest you should get a modelling quality of 99.9%? Any idea why yours shows n/a? Regards.
Hi Carnino,
I wanted to risk 4% per trade so i changed the risk maximum to 4 but the marginal maximum remained as 5. However,the lots allocated per trade remained the same as when the risk maximum was 1. This is happening when backtesting. In that case,if i want to risk 4% per trade, ideally i should change the risk maximum to 4. But what about the maximum margin? Should i change it to 20 in that case?
Yes, there are two limiters, the [B][I]Risk%[/I][/B] and the [B][I]Margin%[/I][/B]. Even if you increase your [B][I]Risk%[/I][/B] to 4%, depending on your leverage and capital, the [B][I]Margin%[/I][/B] may be limiting your position size. You would have to increase that accordingly.
See Post #735 for explanation about Leverage and Margin calculations and Post #736 for an explanation about the two parameters on the EA. Also, make sure to follow BabyPip’s own School of Pipsology on these matters.
If you don’t understand these concepts fully, then don’t use any EA, but rather practice manual trading on a DEMO account until you fully understand it.
Hallo Mr. Carnino,
If one wanted to work with a fixed TP and SL without using the trailing stop and trail step,would it be possible with this EA? If possible then would one set the trail stop and trail step values as ‘0’?
Yes, you can use it, but you will have to set the [I]Trailing Stop & Step[/I] to a value larger than the sizes for the [I]Stop-Loss[/I] or [I]Take-Profit[/I].
Casting back your mind a couple of years or so,you may remember I was occasionally involved in this thread. For personal family reasons I had to drop out for a fair while.
I have now been able to re-kindle my interest in the Forex market, and am looking for a viable ‘set and forget’ system. Do you know if the original Parmar 3P system is still working for any of the original posters, or even you yourself. I am obviously giving it a try for myself, using the setup from the early posts.
No system is [I]Set-and-Forget[/I] if you want to be successful at trading. There is no “holy grail”!
As for the EAs found on this thread, they are quite old and I have no idea if they are still profitable or not. You will just have to run a few back-tests and see for yourself, but I suspect that the GBP/USD definitely would have been profitable due to the large moves during 2016.
As for my own EAs, those have evolved with me and changed over time to suit my needs. Their functionality no longer resembles the earlier versions posted here anymore, so they cannot serve as a comparison for judging the current profitability of the older versions.