How to Test Forex Strategy

Testing and optimization strategy on the historical dates (Back testing) is an important step for the developer and for the strategy buyer as well.

If you have bought Expert Advisor, make back test first and make sure that your beck test close to back test provided by the Expert Advisor’s seller. I am not misspoke saying “close”, because usually tests on different brokers will be little bit differ. All brokers have different servers, quotes, etc.
For this reason, absolutely wrong to use quotes from one broker for back testing for other broker.
The highest quality of testing can be obtained using tick data, but MT4 brokers usually do not provide tick history for their customers. For this reason, some developers are testing expert Advisors on the ticks history from other recourses. For example on DukasCopy ticks. But even a superficial examination of tick data from DukasCopy JForex Platform and for example AlpariUK, we can see a strong difference between ticks.

Data time                                             Bid DukasCopy                      Bid AlpariUK

2011.10.25  03:11:21                           1.39141                                     1.3912

2011.10.25 03:11:21                            1.3914                                       1.3912

2011.10.25 03:11:21                            1.39138                                     1.3912

2011.10.25 03:11:22                            1.39138                                     1.3912

2011.10.25 03:11:23                            1.3914                                        no ticks

2011.10.25 03:11:23                            1.3914                                        no ticks

2011.10.25 03:11:28                            1.39131                                     1.3911

2011.10.25 03:11:30                            1.39129                                     1.3912

2011.10.25 03:11:31                            1.39127                                     1.3912

Such kind of back testing or optimization can be disastrous for scalping Expert Advisors, where very important every pips.

I.e. as a result of this back testing you will receive 99% testing quality in the strategy report, but in fact it is the quality will be lower than 20 – 30%.

If you unable to use tick data from your broker, test it on the minute data and for loading data, do not use

History center.

Just follow instruction, step by step:

1. Open M1 chart
2. Switch off “Auto Scrool”
3. Press Ctrl + PgUp buttons on your keyboard and hold them. Chart will scroll back.
4. Repeat it for all time frames (M5, M15, M30…)

Back test on these data will be with 80-90% quality, but it will be real quality. We recommend to test it on real account, because data on real can be differ with data on demo.

The testing period 6 months – 1 year will be sufficient for the scalping strategy.

For Non scalping strategies, testing period more important. For this reason for non scalping expert advisors, you can make back test the method described above. And then make back test for longer period using metaquotes data from history center.

1. Press F2 button (Tools -> History center)
2. Select currency timeframe M1
3. Press Download
4. Repeat it for all time frames (M5, M15, M30…)

Back test on these data will be with 90% quality. But Metaquotes quotes are different with your broker.
The second and very important factor that may affect the performance of your expert advisor is spread. For this reason, we recommend to control the spread using our e-“SaveSpread” free expert advisor.

And make the back test with Spread Editor Software. For example if your scalping strategy trades since 1am till 7am on EURUSD and -“SaveSpread” Expert Advisor shows spread during this period from 1pips to 1.5 pips, you should make back test with 1 pips and then 1.5 pips.

If the spread 1.5 pips has a major impact on the expert advisor, you should use other broker for your EA with lower spread.

Leave A Reply (No comments So Far)

*

Current day month ye@r *

No comments yet