Same EA, different terminals from same company, same history data, different results.
I cannot understand why is it so.
What was the Spread on each of the 3 runs ?
Maybe the spread is causing a difference for the results.
1. Check the spread of the currency pair being used to test strategy.
SubNote: Sometimes at the end of a trading week just before the close of the market, the spread jumps really high. Example: avg. spread of 3 jumps to 7. When the market week begins again, the spread resets. I believe all strategies are spread dependent. I believe you would agree, but it is worth clarifying.
2. Check the accuracy of the history downloaded.
SubNote: Again, the importance of history accuracy is strategy based. Some strategies require 99% accuracy and some, require just 90% accuracy. I strongly believe that if a strategy requires 99% accuracy for positive results, the strategy is not going to "hold water" when the ship sails for years to come. The market is very general and not specific. The goal should be to catch general movements of the market and not "nit picky" moments, unless ;) you have an out of the US broker using tight spreads under 1. Then you can be "nit picky" all you want. ;) Basically if your broker uses spread above 1, I would not advise the "nit picky" approach and strongly advise catching the general movements approach.
Hope this was of value,
Thank you for your post.
olympus999,
I am not sure, but i think spread should not affect if trade is opened or not. Third one has 14 more trades compared to second.
I am guessing there are errors in the OrderSend() functions OR ;), the EA is lacking code blocks that deal with errors. I could be wrong, but ;) I could also be right.
Thank you.
olympus999,
You may have downloaded 100% identical historic data for all three terminals, BUT ;) what is the accuracy of the data being tested? 90%? Anything below 90%, I would personally consider unreliable. I strictly prefer downloading history straight from the mt4 terminal (it seems you may not be able to download history with the terminal though...). I do not have experience using this source for downloading history, but if the source you are downloading history from is outside the terminal, I would be skeptical of its validity.
Thank you.
olympus999,
You may have downloaded 100% identical historic data for all three terminals, BUT ;) what is the accuracy of the data being tested? 90%? Anything below 90%, I would personally consider unreliable. I strictly prefer downloading history straight from the mt4 terminal (it seems you may not be able to download history with the terminal though...). I do not have experience using this source for downloading history, but if the source you are downloading history from is outside the terminal, I would be skeptical of its validity.
Thank you.
I don't remember very accurately, but i did do some verification. I compared my data with reliable data(it might have been bloomberg) randomly. And it checked out. I do not download from terminal because several gaps appeared and they were significant, i tried to remove them but no use. i googled this problem, i tried to delete historic data, download data again etc. Nothing seemed to work.
At some point i had very hard time getting good data. I still would like to use tick-data but i have not found site that offers it from year 2001, if price wouldn't be to high i would pay for it.
But yes, at the moment i have testing quality of 89.99%, i have not yet found better historic data.
Incorrect, unless you use tick data and a pre -prepared fxt file the Spread does not come from the history data . . . it comes from your Broker, if you were connected to your Broker when you ran these 3 tests it is very possible that they had different Spreads . . .
Something i will look into!
Still 4 digits. I did have one other site where i downloaded it, but this information i have in another computer. Do you know where you could get data for 5 digit broker and can you share it?
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
You agree to website policy and terms of use
picture:
http://www.upload.ee/image/2913828/Weird.png
Third terminal(third from up in picture) is installed some time ago.
Second terminal(second from up in picture) i installed recently.
And first one(first in the picture) is copy of second terminal.
All terminals are from same company and i downloaded identical historical data for all terminals in offline mode.
There is almost no difference between second and first, but the difference with third one is enormous. I also found one place where second one opens a trade, but third one wont. But there is no difference in price or indicators. I cannot understand why is it so.