![MQL5 - Language of trade strategies built-in the MetaTrader 5 client terminal](https://c.mql5.com/i/registerlandings/logo-2.png)
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
VB. to link microsoft products
The Alpiri M1 data has many gaps in it. I have downloaded the 2 year history into MT with the History Center....
3 and 4 minute gaps are very common.....
i recommend testing on the M5 bar will be more accurate.... the reason for this is if your code collects 5 M1 bars to simulate a 5M bar, you are often creating an inaccurate 10M bar....
you are missing bars that have potential tops and bottoms in them.
So can we get an update here? This thread is all over the place! What is decided to be the best version of this EA, on what time frame, on what pair to be the most profitable?
Lets establish this from here and move on...
Thanks
~Billy
you want the right recipe? us too so test and tell us.![](https://c.mql5.com/forextsd/smiles/biggrin.png)
So can we get an update here? This thread is all over the place! What is decided to be the best version of this EA, on what time frame, on what pair to be the most profitable?
Lets establish this from here and move on...
Thanks
~BillyHow about you? have you tested the EA on the different settings that holygrail propose to us for forward testing? what are your results?
To test this kind of EA, we will need months before knowing for sure which pairs/settings and timeframe are the best.
Right now there are few choices to do: there is the old version of the EA (v2) and the new ones (v3). I am testing them both. Why don't you give it a try and let us know what you think is the best setting and maybe propose to use a different way to set it up then and only then we could "move on...."
Sada
To test this kind of EA, we will need months before knowing for sure which pairs/settings and timeframe are the best.
To be honest with you, Sadaloma, this EA needs to be optimized once a month over the past 4 months of data.... to accomodate market changes.
Shouldn't we develop an optimizer that can be run once a month to find the best parameters/pair? Backtests with MT are questionable in my opinion due to missing M1 bars.... M5 bars cover all prices moves in each M5 period and so my vote is to have someone code the EA in standalone application that imports 4 months of M5 bars.....
This application would cycle through all possible combinations of parameters to keep us fresh each month....
To be honest with you, Sadaloma, this EA needs to be optimized once a month over the past 4 months of data.... to accomodate market changes.
Shouldn't we develop an optimizer that can be run once a month to find the best parameters/pair? Backtests with MT are questionable in my opinion due to missing M1 bars.... M5 bars cover all prices moves in each M5 period and so my vote is to have someone code the EA in standalone application that imports 4 months of M5 bars.....
This application would cycle through all possible combinations of parameters to keep us fresh each month....Yes that would be great but I fear that this would not happen in the near future. Many have proposed to do just that with different types of EA in other forums but unfortunately nobody with the skills is willing to do it for the "greater good of mankind"...
We could always hope that someone might be interested but I wouldn't hold my breath for it to happen anytime soon.
There is also another way that have been succesfully used. There are other platforms (metastock or tradestation or something else...I can't remember) that have a good even great "tick"backtest capabilty... I think the original version of this EA was simple enough to code it in another language for another platform.
What do you think?
Sada
backtest / Forward Test
Hi everyone,
I noticed forward test doesn't confirm with backtest. Forward test did sell at 04.16.2006 6:14 and was loss, whereas backtest at the same time opened buy order and was profit, using same EA.
Also I noticed if you run backtest very quickly sometimes its gives wrong results, may be data is still in computer memory, if you give few miuntes gaps then different results.
In short MT4 is not good for backtesting. Even forward testing with demo account is not accurate, demo accounts receive fewer ticks then real accounts, and this EA depends upon tick movement.
Ouch!
Ouch! That hurts.
Most of my profit blown away because the brokers took their profits today (and yesterday!). Raised my SL from 30 to 70. Lets wait and see what next week brings. After all I'm still above my initial account of 5K! I still trust this EA!!
Tick backtester
Hi!
I hope to have finished my tick backtester this WE.
I have just translated all code in VB Except indicators used (EMA and Stochs) but this should be easy. Thanks to my boss to have let me in peace today![](https://c.mql5.com/forextsd/smiles/teeth_smile.png)
I just need to code the environment (Check data quality, correct gaps, reading data, manage trades and reporting).
On EURUSD one year tick history I got from visualchart I have an average of 5 ticks per minute (I took 250 open days / year).
Data integrity will be a constant problem.![](https://c.mql5.com/forextsd/smiles/brick1.png)
And while I am running only in demo I will not implement a tick recorder if visualcharts tick data are same quality than MT4 demo servers.
DOES SOMEONE HAVE GOOD TICK DATA?Is there a good provider not selling this too expensive?
After have coding all of this it would be a good investment .
After I'll be able to analyse data with multidimensional cubes and make data mining (as example, to know in which conditions (Trend, etc.) this EA doesn't work) to adapt strategy. it will also allow to make autobacktests and setting new parameters automatically.
For example, on low TF you can adapt hour by hour (with limits to changes).