Backtest results! | Forex Factory

Options

Permalinks
Print Thread

Similar Threads

Free Expert Advisor with positive backtest results 5 replies

Visual backtest result different from non-visual backtest 0 replies

Backtest results 5 replies

MT4 EA - strategy tester - backtest results inaccuracies 15 replies

Unhappy Weird results in backtest build 200.. need help! 7 replies

Attachments: Backtest results!
Exit Attachments

Backtest results!

  • #1
  • First Post: Jul 29, 2006 7:21pm Jul 29, 2006 7:21pm
I am a beginner trader that only started trading last April. I started trading without having a clue of what it is all about and how much is involved in it. After loosing a considerable amount of money I decide that I need to learn. For a few months I was reading everything I could get my hands on. I found many different strategies and approaches to trading (most of which on FF), but none of them would seem to fit me. So I decided to try and come up with a strategy of my own. For the past month I was working on developing a mechanical strategy that I have finally finished a few days ago. I wrote a script for it that helped me back test it. I used about 3,5 years of data and the results are in the spreadsheet attached. Since this is the first time that I do something like this I was wondering how good/bad the results actually are. I am hoping that the more experienced people in this field could tell me their opinions or comments regarding the back test results. Any replies would be greatly appreciated.
Attached File(s)
File Type: zip Backtest.zip   282 KB | 906 downloads
Stubbornly persistent
  • #2
  • Jul 30, 2006 3:42am Jul 30, 2006 3:42am
Hi,

Your Test 2 eurjpy and usdcad look interesting.

First, have you accounted for spread, and if the pip count is correct for real life entry and exit. I see you have 582 trades for eurjpy. Every pip spread means -582 off your total pip result.

Second, have you converted your pips into lots and $. Due to rounding and minimum lot size (unless you are using oanda), it will impact your profit and loss. Although pips are an indication if the system will be profitable, I've found I much prefer the results converted into cash and observing the percentage drawdown in cash rather than pips. I've sometimes found some surprising differences.

Third, make sure your backtesting results reflect your trading method exactly. Pick a few random or questionable trades and make sure manually that they agree with how you would have traded. Also check if any stoploss/profit orders are executed correctly and you are not 'cheating' by seeing the future in backtesting. Make sure your testing is done at 1 minute intervals.

Fourth, are you going to be awake for all the trades? I noticed that your intervals are four hourly.

I'm sure there are others but those are common mistakes that I have made.

Best regards
Alan
  • #3
  • Jul 30, 2006 4:13am Jul 30, 2006 4:13am
Hi firehorse!
1. I have accounted for the spread. I used FXCM’s spread even though I am using oanda for my trading (the spread is even tighter). Since I am using oanda the rounding and minimum lot size should not be an issue.
2. I have not converted the results in $. I have a spreadsheet with specific risc management rules that I plan to use to convert the pips in $ to see what an impact the drawdown would have on my account. The problem is that since the data used for the backtest goes back 3,5 years it would take forever to try and do this manually since there are so many trades. I am working on getting a formula in excel that would alow me to do that automatically. I am also hoping that the drawdown for different pairs would happen in different periods so that there would be some overlapping making it easier for my account to tolerate it. I have not checked this yet. Still working on it in excel.
3. My backtest results do reflect my trading method exactly. I have created a script for the method used, and my charting software generated the entries and exits and provided me with the results. I have not optimized the system in any way other then finding the stop loss and take profit that seemed to have the best results. I have not used any filters either but plan to do that in the future and see if the results would improve. I am using the 4h chart and a trade can only be placed after the candle has closed so there isn’t really a need for backtesting on tic by tic data (or 1 minute).
I hope this answers your questions. Thank’s a lot for the reply.
Regards,
Sergiu.
Stubbornly persistent
  • #4
  • Jul 30, 2006 4:22am Jul 30, 2006 4:22am
sergiu,


Good luck to you. I want you to know though. I have heard alot of people say that backtesting results suck. The best way to do it is manually go back and write down every trade you would take and do it that way.
JOHN 3:16
  • #5
  • Jul 30, 2006 4:44am Jul 30, 2006 4:44am
Hi Sergiu,

You've done well to account for the spread
Quoting sergiu
2. I have not converted the results in $. I have a spreadsheet with specific risc management rules that I plan to use to convert the pips in $ to see what an impact the drawdown would have on my account. The problem is that since the data used for the backtest goes back 3,5 years it would take forever to try and do this manually since there are so many trades. I am working on getting a formula in excel that would alow me to do that automatically. I am also hoping that the drawdown for different pairs would happen in different periods so that there would be some overlapping making it easier for my account to tolerate it. I have not checked this yet. Still working on it in excel.
I have a running total in the columns next to each trade result to convert each trade into risk/stoploss size, trade size, profit/loss in cash.
Quoting sergiu
I have not optimized the system in any way other then finding the stop loss and take profit that seemed to have the best results.
This is a common mistake I have done myself. To do backtesting correctly you need to find the best result for 2.5years and then apply that to the last year of untested data to get a true result for 1 years worth of trading. Otherwise you are seeing into the future

Or you can test the first year of data, find the best stoploss and profit for that year and then apply it to the next 2.5 years.

Many of my own systems have broken down when this was done I hope yours is more robust. Good luck

Best regards
Alan
  • #6
  • Jul 30, 2006 4:49am Jul 30, 2006 4:49am
Quoting renasdad
sergiu,


Good luck to you. I want you to know though. I have heard alot of people say that backtesting results suck. The best way to do it is manually go back and write down every trade you would take and do it that way.
I thought about it. The back test was just to give me a general idea as to whether the system would have any potential. I also think that manual testing has its pitfalls. If you see that a trade that one might have taken would have been a looser one might start looking for reasons why one would have not taken it, making additional rules in the process which would be irrelevant to the strategy in the long run.
Stubbornly persistent
  • #7
  • Jul 30, 2006 4:53am Jul 30, 2006 4:53am
Quoting sergiu
I thought about it. The back test was just to give me a general idea as to whether the system would have any potential. I also think that manual testing has its pitfalls. If you see that a trade that one might have taken would have been a looser one might start looking for reasons why one would have not taken it, making additional rules in the process which would be irrelevant to the strategy in the long run.

Think about it this way though. Do a manual backtest, stick to the rules, take every trade that is signaled, when you see one that might have been a loser, take a note of it. Do that for each trade taken that fails. Maybe not each trade but, what could have prevented it. Manual backtesting can help a person tweak a system and understand a certain currency pair. When I manually backtest, I feel I am more "ONE" with my chart.
JOHN 3:16
  • #8
  • Last Post: Jul 30, 2006 4:59am Jul 30, 2006 4:59am
Quoting renasdad
Think about it this way though. Do a manual backtest, stick to the rules, take every trade that is signaled, when you see one that might have been a loser, take a note of it. Do that for each trade taken that fails. Maybe not each trade but, what could have prevented it. Manual backtesting can help a person tweak a system and understand a certain currency pair. When I manually backtest, I feel I am more "ONE" with my chart.
Makes sence!
Stubbornly persistent
0 traders viewing now