DislikedHere's a cleaned up version of Blessing with the input names matching the manual once again. This should be helpful for those that do a lot of testing, and to generally reduce confusion. I hope to add a couple of useful features to Blessing soon, for small accounts, etc., but thought it would be good to get this base version out first. Trying hard to maintain continuity with earlier versions of Blessing ... {file} Best of luck!Ignored
- Joined Nov 2008 | Status: Trader | 47,736 Posts
Honesty is a very expensive gift. You wont find it in cheap people.WBuffett
- Joined Nov 2008 | Status: Trader | 47,736 Posts
Honesty is a very expensive gift. You wont find it in cheap people.WBuffett
- Joined Nov 2008 | Status: Trader | 47,736 Posts
Honesty is a very expensive gift. You wont find it in cheap people.WBuffett
- Joined Nov 2008 | Status: Trader | 47,736 Posts
{quote} Thank you very much, Sir. I will definitely have a look at the manual.
the manual posted by richard is the same last available version I got from jta.see richard' post#1324: https://www.forexfactory.com/showthr...9#post12570209
what you people missed is their jta forum discussions and updates notes website which unfortunately has been taken down.
but the v3.9.6.09 and this manual will be of good use to get started with and share our experience and discuss problems on day-to-day basis.
trade well folks, this is our blessing.
if anyone interested I still have the predecessor version Blessing2 more than 10 years ago. Its has only 500 lines code and easy for beginner coder to study and learn from a time tested and proven working expert advisor.
but let's not flood this thread with other stuff.
Honesty is a very expensive gift. You wont find it in cheap people.WBuffett
Guys,
I keep seeing recommendations to use broker's data, and I must protest, you may contradict me, and you may also verify what I say here.
If you rely on your broker's data, at least make sure it is accurate and without holes.
If you backtest report shows less than 90% tick modeling, that means there are "holes" in the data, and that can be verified and even corrected (explanation below).
Example of an excerpt from backtesting report (it doesn't preview in the post because of the resolution, click on it):
So, if you made it this far and indeed your modeling percentile is less than 90%, how do you correct the data? (For less work and better performance, skip to the last paragraph).
1. First, make sure your MT4 can hold as much history as your broker allows, and for that you got to Tools -> Options (Ctrl+O), than Chart tab, and increase the number of bar in history. Something like that:
2. You go to the history center (F2), choose your asset and timeframe, and click on Download once again (assuming you already did that to get recent data in the first place)
3. After downloading is finished, you may backtest again, and look at the percentage again. If it is less than 90%, proceed to step 4.
4. For more advanced Excel users - which you must be if you want to professionalize in this - go to the history center (F2) once again, this time instead of Download, click on Export, and choose a folder to save your .csv file. This file can be opened in Excel.
Now what you need to do is find holes and fix them.
- Holes are price gaps between previous close price and current open price (could be verified in currencies, less so in stocks and commodities).
- Holes could also be close price equals to open price (in any timeframe in a tradable asset this could seldom happen, but having holes could show you this happens too frequently).
Both of those could be verified by subtracting close price from open price.
- Holes are also skipped bars where there should be bars, which means when the market is open you could have a bar or more simply missing, could be verified by subtracting the previous row's date from the current row's date.
5. If you did encounter holes, you can fix them one by one by completing what is missing - simulating close price, adding missing bar data, etc.
Last paragraph -
The best method to improve modeling quality is to acquire accurate data from a reliable source, which can also give you tick data and not only 1 minute data as MT4 gives you.
What's the problem with 1 minute data?
That MT4 simulates the tick movement intrabar, it doesn't really have information besides OHLC of the 1 minute bar, no matter how much you try.
That could give you many faulty open/close trades and differences between two consecutive backtests with the same parameters.
With actual tick data, each backtest equals exactly the previous one.
I previously recommended Tickstory as the data provider, but you may use any other reliable source.
I keep seeing recommendations to use broker's data, and I must protest, you may contradict me, and you may also verify what I say here.
If you rely on your broker's data, at least make sure it is accurate and without holes.
If you backtest report shows less than 90% tick modeling, that means there are "holes" in the data, and that can be verified and even corrected (explanation below).
Example of an excerpt from backtesting report (it doesn't preview in the post because of the resolution, click on it):
So, if you made it this far and indeed your modeling percentile is less than 90%, how do you correct the data? (For less work and better performance, skip to the last paragraph).
1. First, make sure your MT4 can hold as much history as your broker allows, and for that you got to Tools -> Options (Ctrl+O), than Chart tab, and increase the number of bar in history. Something like that:
2. You go to the history center (F2), choose your asset and timeframe, and click on Download once again (assuming you already did that to get recent data in the first place)
3. After downloading is finished, you may backtest again, and look at the percentage again. If it is less than 90%, proceed to step 4.
4. For more advanced Excel users - which you must be if you want to professionalize in this - go to the history center (F2) once again, this time instead of Download, click on Export, and choose a folder to save your .csv file. This file can be opened in Excel.
Now what you need to do is find holes and fix them.
- Holes are price gaps between previous close price and current open price (could be verified in currencies, less so in stocks and commodities).
- Holes could also be close price equals to open price (in any timeframe in a tradable asset this could seldom happen, but having holes could show you this happens too frequently).
Both of those could be verified by subtracting close price from open price.
- Holes are also skipped bars where there should be bars, which means when the market is open you could have a bar or more simply missing, could be verified by subtracting the previous row's date from the current row's date.
5. If you did encounter holes, you can fix them one by one by completing what is missing - simulating close price, adding missing bar data, etc.
Last paragraph -
The best method to improve modeling quality is to acquire accurate data from a reliable source, which can also give you tick data and not only 1 minute data as MT4 gives you.
What's the problem with 1 minute data?
That MT4 simulates the tick movement intrabar, it doesn't really have information besides OHLC of the 1 minute bar, no matter how much you try.
That could give you many faulty open/close trades and differences between two consecutive backtests with the same parameters.
With actual tick data, each backtest equals exactly the previous one.
I previously recommended Tickstory as the data provider, but you may use any other reliable source.
Thunder OM Flex Secondary All Time Return:
-18.2%
- Joined Nov 2008 | Status: Trader | 47,736 Posts
Honesty is a very expensive gift. You wont find it in cheap people.WBuffett
- Joined Aug 2008 | Status: Money never sleeps | 5,878 Posts
Markets are not Random, they are designed!
- Joined Nov 2018 | Status: Trader | 564 Posts
Here are some results from some of my more recent Blessing setups ...
When you use your broker's data you don't need to forward test more than a few days to feel confident about going live.
Forex is not the stock market. Every broker provides their own version of market data. Check your contracts.
When you use your broker's data you don't need to forward test more than a few days to feel confident about going live.
Forex is not the stock market. Every broker provides their own version of market data. Check your contracts.
- Joined Nov 2018 | Status: Trader | 564 Posts
Here are some recently optimized set files ...
No guarantees! These still need to be forward tested by you ...
The file names look something like this:
201911032-gbpaud-h4-OL.set
The OL means it was optimized with Oanda live data. Brokers represented are Oanda (OL), Forex.com (FL) and FXChoice (CL).
A lower-case O, F, C, denotes optimization on a demo account. If you find that one of these broker's tests well with your broker's data then please share that information. Don't expect too much from different brokers.
Most or all of these files were created using Blessing_3_v3.9.6.11.mq4 or later, though should be compatible with earlier versions.
Best of luck!
No guarantees! These still need to be forward tested by you ...
The file names look something like this:
201911032-gbpaud-h4-OL.set
The OL means it was optimized with Oanda live data. Brokers represented are Oanda (OL), Forex.com (FL) and FXChoice (CL).
A lower-case O, F, C, denotes optimization on a demo account. If you find that one of these broker's tests well with your broker's data then please share that information. Don't expect too much from different brokers.
Most or all of these files were created using Blessing_3_v3.9.6.11.mq4 or later, though should be compatible with earlier versions.
Best of luck!
- Joined Nov 2018 | Status: Trader | 564 Posts
Here's a slightly updated version of Blessing with a new MinMarginPercent feature.
If you turn it on then Blessing will stop opening new trades when the Margin Level Percent drops below the level you set.
This is nice for testing, to help get an idea how much room a setup has within the account. I'm thinking it should be nice when populating an account with multiple Blessing instances, hopefully utilizing your equity more fully. Or just for an extra level of safety.
It doesn't currently remove existing pending trades. I may add that in the future.
You turn it on by setting UseMinMarginPercent to true. And set whatever margin level percent limit you want in MinMarginPercent.
Enjoy!
If you turn it on then Blessing will stop opening new trades when the Margin Level Percent drops below the level you set.
This is nice for testing, to help get an idea how much room a setup has within the account. I'm thinking it should be nice when populating an account with multiple Blessing instances, hopefully utilizing your equity more fully. Or just for an extra level of safety.
It doesn't currently remove existing pending trades. I may add that in the future.
You turn it on by setting UseMinMarginPercent to true. And set whatever margin level percent limit you want in MinMarginPercent.
Enjoy!