Hi Mike

I don't frequent FF anymore as I was just not getting much valuable info from content related to what I do...and there was too much bickering on the forum.....so I let the Data Mining Thread of mine lapse. My workflow has changed a bit since then....but hopefully some of it was helpful.

I now post on my website at https://atstradingsolutions.com/ and involve myself with some of the professional fund managers on Linked-In plus lately Twitter.

I plan to write up some good content on particular approaches to data-mining using EA studio and I have already written some content that may be helpful to those interested in trend following/momentum approaches.

I like your approach you have nutted out Mike. You obviously have spent considerable time with it. Since you are trading on M30 I totally understand where you are coming from in some of your solutions that reduce slippage and execution issues. :-)

I am currently benchmarking my research efforts against the top CTA's in the trend following/momentum space. By closely monitoring their performance, you can see what are reasonable versus unreasonable expectations. Interestingly the approach I now use pretty well mimics their performance.

Below are the long performance results since 1 Jan 2000 for the Trend Following Index that comprises about 60 of the top trend following CTA's in the world. Note the CAGR and Max Drawdown for the 19 year period. This provides an important benchmark that defines whether my test results are pie in the sky or not.

https://i1.wp.com/atstradingsolutions.c … .png?ssl=1

Here are the results of my currently compiled portfolio which spans 13 instruments (FX, Commodities, Indexes, Metals). The equity curve very closely mimics that produced by the trend followers that gives me confidence in it's future robustness. The equity highs and drawdown are almost perfectly aligned with that of the professional CTA's...but still need a bit more diversification to more precisely mimic them.

I am adding to the portfolio as I go and plan to settle on about 30  instruments across asset classes.

https://i0.wp.com/atstradingsolutions.c … .png?ssl=1

You also need to note that the CTA's performance results are net of management and performance fees so the CAGR pre fees would be slightly higher and the Max Draw of the TF Index slightly lower if we were comparing apples for apples.

I plan to scope out the workflow of the processes I currently undertake on my blog so hopefully will get a bit of interest in the methodology and hopefully will be able to share notes with those on a similar path.

Cheers

Rich

Yep Mike. Diversification and risk weighting are the key priorities for me as well. Fortunately EA studio makes life very easy with regards to system diversification. Then you just need to deal with instrument and timeframe diversification.

I find it is very important to diversify across different markets when adopting trend following techniques as there can be long periods in any single market where trends rarely present themselves. If you are restricted to a few instruments you can find yourself spinning your wheels uselessly with building drawdowns.

Risk weighting is incredibly important especially since EA studio operates off standard lot sizes for market selections. Like you I use historic max drawdown as a basis to adjust position sizes for each strategy to risk weight them.

The chart below provides an example of a compiled portfolio that does not use a risk weighting approach (a raw portfolio). Notice how 3 instruments significantly bias the results of the overall collection of 14 instruments.

Now here is a risk weighted solution applied to the same 14 instruments using the drawdown as a basis to adjust position sizes.

Risk weighting produces a significant boost in performance result for the total portfolio.

Here is the raw portfolio result.

...and here is the risk weighted performance result.

The difference is significant.

Mike I can now pretty well manage my drawdown profile through adjusting position sizes at the global portfolio level. My most recent blog on the website gives an idea how to do this.

I actually have no long term profit target as I avoid prediction in it's entirety. By focusing on the risk management side of the portfolio, profits are just a symptomatic outcome associated with the fact that markets do tend to trend at points in time. I never know when that will occur....so I just let the portfolios do their thing...as I know that when markets decide to trend I will be riding them. My worst fear is the fear of missing out on a trend....so I diversify as much as I can.

I trade bundles of portfolios that contain many strategies within them......and the more the merrier.

As I risk weight the portfolio, you will find that as a portfolio builds a larger drawdown, the position size is decreased....hence I never actually turn strategies or portfolios off at all. I continually add new portfolios to the collection. It is a never ending process. The only think that changes is that you need to progressively build up your VPS servers and broker live accounts as your portfolios build. At the moment I am operating off about 10 separate VPS accounts each operating off a separate live broker account ....and this is likely to continually grow as time progresses and the workflow processes continue. I use FX Blue to do most of the heavy lifting with live trade data collection which I import into my data analysis tools.

Make sure that if possible you use an account that offers rebates off volume transactions as this can be a nice bonus for high trade activity.

You seem to have a very similar ambition in your trading goals that I do which appear very realistic.

Nice catching up Mike.

Cheers

Rich

Hi Geektrader thanks for sharing your reports. May I know how did you use Quant Analyzer to generate such a report for strategies that were created with EA Studio?