Developing an Automated Trading System – 2/1/15

It has been my dream to have a money tree. It would take some nurturing, adding nutrients and water but US dollar notes would sprout from the tree. Alright, I know that is not in the likely realm of occurrences! My proxy dream is to be able to develop an automated trading system. I would be trading e-mini Nasdaq 100 futures (or oil futures) on a very short time frame and preferably either just the morning or the afternoon. The computer would have my algorithm saved in its brain. It would include a time solver so that at the start time it would wake up and look for signals and at the end of the session it would close any open trades... On receiving a buy signal the computer would automatically place the trade. A short time later after it reached its profit objectives it would take profit. All along the process, it would also have exit plan so that if the futures behaved differently from what was expected, it would exit the trade for a small loss and wait for the next signal.

Till recently, I did not think I could do that as I am not a C++ code developer. I think I am pretty knowledgeable on technical analysis but converting that knowledge to an algorithm and then back test it seemed impossible with my limited capabilities (COBOL, Basic and Fortran) in programming. But that was till 2 months back when I ran across Sharkindicators Bloodhound software. 

See link for its home.        http://www.sharkindicators.com/

 Quick Start tutorial          https://www.youtube.com/watch?v=0viC-gSQIw8

With this software one can take technical indicators, even with multi-time frames and using visual drop and drag boxes that have my indicators and logic, combined with and/or and many other node features create quite sophisticated trading systems. Additionally Bloodhound is written for ninjatrader software which can go to any of several brokers who specialize in low cost futures trading, placing buy and sell trades. And did I say that short term futures trades are treated as long term capital gains for taxation? Go figure.

So here is a chart of a quick system I was back testing. It uses MACD’s sweet spot when it breaks out close to 0 and a medium speed EMA (exponential moving average) like 20 to decide which direction it will trade and when. The chart shown below is a 2 minute chart of NQ. I started practicing back testing with it using 1 minute past data from last year. I am testing it during the first 3 months of actual data from 2014 and where the results look profitable, I tested it again for the next 3 months. I changed the Take-profit level from 6 ticks to 48 ticks in 6 tick increments. Each tick is worth $20 in profit on one contract of /NQ. The results show that keeping a take profit in the range of 24 – 48 ticks is much better than keeping it tight at 6-12 ticks. Small stop loss values are not very effective with this system as it is not super well timed for entry. I may consider adding stochastics to it and improve timing. I can enter only when stochastics is below 80 or 90, a typically classic signal to enter on weakness but in the direction of the long trend. I left the stop loss as unspecified and allowed the system to exit when the opposite signal was generated.

The beauty of the software is that I can take these systems with all its variables through its paces and better understand the hard and soft regions of profit and losses. This trading system is certainly not always profitable but it has promise. I have many ideas for trading and this software has given me the capability to start with taking a few simpler systems through its paces before I get into ones that are more complex. I purchased Sharkindicator’s Bloodhound Ultimate version which sells for $495 and also the BacktestRenko option which is additional.

I look forward to spending many hours listening to Youtube training videos by Jeremy Tang the creator of the software and Zac, one of their experts.  I can fine tune and testing many ideas using technical analysis coupled with logic algorithms to come up with my money tree. I don’t expect any if this to happen quickly and expect many hours on weekends to make this work. But it will be worth it….

---

 [RM1]