First time I've posted on ET, and as such, I tip my hat to you all. I've been trying to power my way through a self-guided forex education for the past half-year or so. Anyway, if anyone wants to take a look at this trading plan and offer any suggestions/improvements, I'd be grateful. I'm not sure how many traders here use MT4, but I've included a strategy report for this EA-based strategy with an ECN broker. As follows: 1. Buy the EURUSD when it crosses above its 100 EMA and PSAR (max = 0.05, instead of the default 0.2) confirms the signal. 2. Sell at the opposite signal. 3. Close when PSAR reverses its signal. I end up with a kind-of-smooth equity curve, a respectable profit factor between 1.5 and 1.6, depending on how low the spread is at the time of backtesting, and no parameters to optimize since the strategy uses an indicator to close rather than a fixed SL/TP. Anyone have any ideas on how to increase the %winning trades, to limit the sizable drawdown, or to improve the strategy in general? Thanks for your help.
Quick look, The modality is less than 90%. Trades number is too few in my opinion ( i prefer to be more than 1000 trades) DD is huge in my opinion. I think if you come up with modality 90%, trade number more than 1000 trades, DD is less than 10%. profit factor more than 2. Then, you will attract more pro to get into your strategy and share their opinions. You can get many strategies giving similar results as yours.. You need to work more on your backtesting. Good luck My 2 Cents
I'm just curious how you guys can trust MT4 as a backtesting platform? Just data quality issues alone makes one really pause and think.
Thanks for all the replies! That was fast. Time frame is H1. So, the reason the modeling quality is so low is that I don't know where to get current (recent) historical data for MT4. If I want to backtest on 1990-2009, I do it on a specific version of MT4 that has that quality data on it. And I get 90% modeling quality. However, I prefer to test on the more recent past. Do you know where I can get good historical data for the past 6 months or so? Why is more trades better? Do you just want to see a larger sample, or do you think that EAs that trade more times a day are better? Lastly, I'm not hard-set on using MT4 for backtesting -- can you recommend any other platform I can use? Like I said, I'm relatively new in this area. And again, I appreciate these responses from such a qualified group of people.
AB, Tools/History Center/Highlight Pair/Download then... Charts/Max Bars in History (increase this number if you need to) ES
ES, Thanks for your reply. I've seen your name all over pretty much every forex forum I've come across. Anyway, I know how to download historical data in MetaTrader, but I've heard that the data provided with the software isn't reliable. Would you agree? Or do you think it's good enough... -Alex
AB, Believe it or not...I am not a backtester. I am a forward tester. I do not know how to backtest in MT4. I have read about it. When I forward test an EA, I usually load up about 10 MT4 platforms and test incremental units of inputs and note the performance change from platform to platform. I try to note if the market is trending or choppy during the testing period...if I need more information on how the velocity of DD changes I can manually apply the 10 year ranges to my pip values that I have discovered to see the extremes... There is a coder named wolfe in another forum who understands what I do and you can ask him how effective it is...However, The system that I was testing for him did not work....oh well... ES
I do not have a bias regarding EAs trading frequency but i prefer to look at a larger sample. If you look at MT4/forex specialized forums, you will find easily too many free EAs and Systems that give similar results.. However, backtesting is not telling everything about the future. Also, live trading has also many restrictions and limitations that makes it somehow different than backtesting. I hope you all success.. and good luck
So in your case ES, testing an EA that trades only once or twice a day would be very difficult, because it would take a long time to get a body of results? Personally, I forward test everything that shows a reasonably good backtest... but in these situations it can be pretty hard with a low trade volume. I assume you use MT4 for this sort of thing? Which broker do you use, may I ask?