I have been working through the workshop on gaining an understanding of the coding of functions to use Zorro. I have a basic understanding of coding from working with DB programming, SQL, Python, etc., but am willing to learn.

I wanted to put out some thoughts about how I would like to go about building an algorithm and wanted to see if I could get some feedback if it is a viable approach for use with trading Forex. I have been trading FX for almost a year now, both live and demo, and I feel it could use some improvement and greater discipline gained from a more mechanical process.

My understanding is that a basic algorithm has various components, such as a confirmation indicator, a continuation indicator, a baseline, some measure of volume / volatility, an exit strategy/indicator and money and risk management. Is Zorro a good choice for me to go through the process of testing each indicator to determine the best indicator for a particular function of the algorithm, and where each of the additional components are added to the algorithm, improve the overall results? This algorithm would be based on using a daily time-frame and will be primarily a trend trading strategy. Each of the components and the final algorithm comprised of all the selected components (indicators & rules) would be back-tested, forward tested, etc. to arrive at a final algorithm that would produce a viable and reliable ROI. Ultimately if this were to work with Zorro, I wouldn't want this to be an automatic trading system, but rather something that would give me a report each day about which trades I should make, with Zorro providing me all pertinent details about entries, stop losses, trailing stops and exits (for trades already underway).

Is this something feasible with Zorro (although I gather you can do so much more) or do you think Zorro would be overkill? I am hoping to build my algorithm in this fashion to avoid over-fitting any single indicator and / or part of the process. I would like to try creating a robust algorithm that I can use Zorro to help provide feedback about when I should maybe go back and revisit / re-test different parts of the strategy to see if it requires updating or improvement.

Hopefully this isn't too much of a noob sounding question and I do understand I have a lot to learn; but I really want to do things the right way.

Thanks

Of course, you can use Zorro as a trading advisor. You could print out the information you want to it's output window. But I wonder how you would want to back test such a strategy, where you are part of the indicators....? You could find some format for your daily bias (like: Date, time: strongly bullish on EURUSD) and honestly note it down every day, but not in hindsight...! Then in like 2-5 years from now, you could backtest your strategy (including yourself as an indicator) and see if it is any worth.

My personal opinion it this: Don't interfere with the machine. Either trade by hand, or let the machine trade for you.

the back-testing part would be in the assembling of the different components/indicators/rules that would comprise the complete algorithm. For example, I might use a crossover of some sort to act as my initial confirmation indicator for a trade and I would test it (back test) to see how well it performs in picking profitable points in which to enter a trade. I would be looking for such an indicator with the highest win rate. There are many different indicators like this, some good (i.e. more accurate) and some not so good. But still this one indicator wouldn't be sufficient in which to base an entire trading strategy on. Then I would add additional indicators or rules (like using ATR for setting stops, etc.) to help fill in the deficiencies that are lacking in the confirmation indicator to help improve the wins or limit losses, etc. The back-testing would be performed using these indicators across multiple pairs to see if the indicator isn't just useful for a couple of pairs but maintain a win rate across many (28) pairs. Then once I have assembled, the hopefully right combination or indicators & rules into an algorithm, it too will be back tested and forward tested to determine at the end of the day if its something I am willing to risk an investment.

I don't know if this is the best use of Zorro, because the only reference point I have is using MT4 and testing different indicators in this manner. Maybe there is a better workflow in building a strategy with Zorro and I have to maybe think of it differently. My idea is fairly linear, but I feel its the only way (at least using MT4) to assemble the different parts for a final algorithm, by testing each component as a build and then test the interaction of each additional component so as to not break the algorithm strategy at the end. This will also help to determine in the future if one of the components has broken down and needs to be replaced. I am probably thinking too simplistically for most in this forum; but just trying to put my thoughts out there to see if my thinking is off and maybe because I don't know Zorro and there may be a better way.

I am not opposed to letting the machine do the trading ... especially to keep my psyche out of the picture which is always a good thing. I will need to develop a level of trust to finally take that step.

Thanks