Diversified Trend Trading – Excel Projects

Hey C - I need some time for me to get my head around this, I haven't had much of it lately but when I do I'll chip in.

thanks for the time you've put into this already..

The Big C is a tad busy, but we should keep the dialogue moving, so in the interests of transparency, i should let you know that I assisted with the construction of this series of spreadsheets.

If you want to bounce any ideas around or are having issues with anything, please keep asking and if i can i will try to respond.

cheers
Fred

Guys

I apologise for my quietness of late on the forum. I have a pressing matter that I need to focus my attention on and as a result will need to put my contributions on this thread on hold for a short period.....but I will return once things settle down to continue. So please be patient and if any visitors to this thread would like to continue in my absence....then feel free to do so.

All the best and stay safe.

C

Hope everything is well.

Best,
B.R.

Cheers Bass

Will be returning soon mate. Thanks for the well wishes. :-)

C

Dear Sr
Thank you share the system.
I download system, the system is running. But i don't know, when i want trade, where do i see? and when can i open order?

Hi

Sorry about the delayed response. Just saw it.

The excel spreadsheets supplied relate to back-testing and not live trading.

Regards

C

I just thought I would throw this into the mix........

Nice ajs

Great to see you with the model up and running.

Hopefully I will be able to contribute more to this thread so it can be taken somewhere and also share some other models I am working on. Also if you have a handle on the spreadsheets and the mechanics of the models, you may be able to take it further yourself with some models of your own.

Once things settle down I will be back and we can expand the universe of instruments and walk through the basis of the spreadsheets.

C

To much risk for reward ?

To make this assessment you need to have an appreciation of what comparable measures are available over the time-frame concerned. You will hear anecdotal statements that Return (CAGR) should be at least equal to the level of max Drawdown, however it is best to check these statements against available audited performance statistics as it is not an appropriate statement to make over the period concerned.

Also you need to note that the back-test model does not compound returns (reinvest profits) which is the usual approach taken to evaluating performance. The reason for this is that simple compounding obfuscates performance metrics and it is difficult to isolate performance attribution to the model versus performance attribution due to compounding. To evaluate the performance after compounding you need to revert to your monthly performance statistics tab in your spreadsheet and apply compounding to these results to observe their effect.

Your back-test produces a CAGR of 9.28% over the period 1/1/2000 to 31/5/2016 and a max drawdown of approximately double the CAGR of 18.94%. Compare this against the best professional fund manager returns for a similar period. Refer to the statistics below of the CTA's.



Look at the max drawdowns incurred over back-tests between 1/1/2002 to 31/8/2016 and the CAGR results. The best performer (Dreiss Research) has incurred a max drawdown 51.44% in producing a CAGR of 16.95%. The best relative performer (FORT) when comparing CAGR against Max Drawdown produced a CAGR of 13.51% yet a max drawdown of 17.16%. The Fund with the lowest drawdown for the period (Transtrend - Standard Risk) produced a CAGR of 4.97% with a max drawdown of 10.59%. Clearly this is highlighting that the best of the best can't achieve the anecdotal equivalent drawdown to return relationship.....I think it is safe to assume that us humble retail traders are unlikely to outperform the professional fundies.

While it needs to recognised that CAGR results for these CTA's are Net of fees and not Gross returns like our tests, it can be clearly demonstrated over the period that a drawdown of twice the CAGR is perfectly acceptable.

Furthermore, if a max draw of 18.94% is beyond your risk appetite you can scale the results by reducing the trade risk% to turn down the volatility heat of returns.

I personally think that the return profile you have produced combined with the absence of curve fitting, makes your test a good one for the instruments concerned that should be stored for future consideration. Remember however that the most robust strategies works over any liquid instrument so I would then start increasing your investment universe to see how this strategy performs over a more diversified set of instruments across asset classes.

What you then need to do is trial new systems over the same test period that respond to different market conditions than your Donchian Breakout strategy, to produce a different set of equity curves that when combined as a diversified system strategy produces enhanced returns over the single strategy with a combined lower max drawdown and higher sharpe ratio.

For example, here is a system blend composite of 15 separate systems each with their own characteristic equity curves that comprise trend following systems with a dash of trend retracement, mean reverting and buy and hold equity strategies. The results assume profits are reinvested and this strategy is applied to 40 instruments across about 8 separate asset classes. Applied weighting are given to these discreet systems to then produce a consolidated result.




The conclusion reached by this treatment is that while it is unlikely that there is a single EA that can be applied to all market conditions, a composite approach to EA development allows you to overcome this hurdle.

You are going great guns ajs.

I am looking forward to getting more involved in this thread when I can.

C

Thanks Copernicus for the good explanation. When I have time I will add some more assets and retest.

Hello Copernicus,
I have had a go at adding some more assets to test but with no luck. I modified this part of the macro
Dim location As String
Dim x As Integer

x = 1

While x <= 5 changed it to 10 to add another 5 assets. It throws up an error and the assets added get deleted.

Please can you tell me how you can add more assets to test?

Thanks.

Hi AJS

You were almost there. Just a few extra things you need to do.

These are the steps you need to take.

Step 1

Let's say we want to increase our universe to include Brent Oil and Sugar which now increases the portfolio from 5 to 7.

So our datahistory folder should now include the data for these additional instruments and look like this.



Note the naming convention used in the datahistory files. The 1440 at the end of each file relates to the D1 timeframe. The actual name of the file we use for the instrument definition in the VB coding is therefore the preceding text prior to the 1440. For example SUGAR#111440 becomes SUGAR#11 and BRENT_OIL1440 becomes BRENT_OIL.

Now that we have added our new instruments to test, the next step is to adjust the PM file and the PC file using the abridged naming notation.

Step 2

Go to your System file directory and ensure that you only have the PC, PM and TM files included in the directory as we need to create the portfolio again. Your system file directory contents should look like this.



Now open the PM file and firstly we add the additional instrument names into the "Dump Sheet worksheet Row A1. This info is automatically transposed into the PC file when you run the code. After you add the additional instuments into row A1 it should look like this.



Now you also need to amend the code of the PM file to include these additional instruments.

In Developer mode in the PM file go to macros and select the 'CollectData" macro and select "edit'.

You need to copy the prior instrument (example NZDJPY) and paste this section of code for the 2 new instruments. Take note that you need to exactly specify the new instruments according to the syntax described above in this post. Also note that the 'pasteref' needs to change to reflect the new position that the information will be dumped into your dump sheet. For example BRENT_OIL information will be pasted into F1 and SUGAR#11 will be pasted into G1.



Save and then close the PM file once this is done.

Step 3

Now you need to simply amend the code in the PC file to reflect the greater number of instruments as you have already identified.
'createportfolio' and select edit.

Open the PC file and select Developer mode and then select macros and choose edit. Then simply adjust the 'While x <=5 to While x <=7 and then close and save.



....and that should be that.

Ready to Test

You are now ready to create the portfolio.

Select the PC file and then select the 'Create Portfolio button' and the amended portfolio files will then be created.

Once that is done you are then ready to open the PM file and run the test across the increased universe.

I hope this makes sense....let me know how it goes.

Once you have this technique down you can start really increasing your universe further. The max I test is about 45 instruments but you can really go as far as you want.

Cheers AJS

C


PS AJS....You may also run into problems regarding the instrument definitions from your broker which we need to include in the PM file in the worksheet titled "Instrument Details". If you want to use your own brokers instruments, you will need to ensure that your instruments and their critical details in collumns A-H are included such as Symbol, (use 1 for SWAP calc), tick size, base exchange, lot size, buy SWAP, Sell SWAP and Spread (PTS).

I have attached a wider universe below (Datahistory.rar) if you want to use the supplied instruments from Avatrade, if you run into any problems using your own instruments and are unsure how to apply your broker's instrument details. The attached instrument are already defined in the "Instrument Details' worksheet of the PM file.

If you want to use your own broker supplied data

Attached is a script "RLF InstrumentDump v3.ex4' kindly supplied by Fred (RedLineFred) that you can drag onto your MT4 charts to capture the instrument details applied by your broker that you will need to apply to the 'Instrument Details' worksheet of your PM file.

The script, when activated, creates a csv file in the following directory of your MT4 program titled 'DataX' that includes all the required details for each instrument that you need to include in your PM file.



Once you have activated your script have a look for the DataX file and open it up. It should look like this.

Cheers

C

Note to readers

Guys, things have settled down a bit so I will get back into progressing this thread for those that may be interested. Obviously the scale of this project is large, (in retrospect, much larger than I anticipated) but we need to take baby steps....I hope you can all appreciate this, as otherwise I will be talking to a fictitous audience if I move too fast......and hopefully there are a few that are keen to progress down this path and I don't wan't to lose you.

In reality, I really don't expect many to stick the course, but I will progress for a little way to hopefully resurrect interest and see how it goes. This project is really for those that want to be involved in building systems from ground up and want to know exactly what is under the hood.

Sorry for the delays to-date.........that is what protracted legal battles result in *groan*....and it's not over yet, so I may need to go offline at times for periods of time, but in the next few weekends I will get the ball rolling again. Hopefully there may be a few that can help out and move things along in these instances to prevent this thread from going cold.

Anyway....at this stage you have the basic module up and running and with the supplied data are running some tests and variants of the Donchian Breakout Strategy supplied.

In my next few posts I will get into the nitty gritty of the spreadsheets themselves to allow you all to get an idea of what is involved so you can possibly create some different strategies yourselves and run them using the basic module.

I will also provide all the necessary info you need to use your own data sources for testing as opposed to being reliant on what I supply to you.

Once this is achieved you will be able to develop and test your own strategies.

Speak soon.

Regards

C

Keep going if you have the time, i'm following with great interest .

Cheers Bass. Will do mate. Great to hear from you old friend

Sounds great C. Welcome back!

Nice to see you back C. I don't always have time to work on this, but I sure would like to learn how to do it.
Thanks