Hi,  I have been struggling to learn to code but love QuantConnect with the good documentation, great commnunity, and education.  I have an idea for a strategy that I think would be interesting to explore.  If you open a 4 hour chart on any highly traded security and zoom it out dramatically and extend/condense the bars to many months/years back you will see larger trends emerge.  What you will notice is there is a lot of "noise" in the market with reversals(mean reversion) happening more than trends.

What I think to myself when I see a chart in this way is that at some point you could have placed a trade in that noise and it would have been profitable at some point.  Time is the key issue because if I were manually watching the trade emotions and draw down would affect me and cause me to exit the trade and that's where the noise comes in.  Thinking about this premise I would like to code a simple algo to test it.  

We would use one indicator, a 30 SMA on a 4 hour chart.  I will use Forex for the example, say EURUSD.  Turn the algo on, if the current price is above the 30 SMA, just Buy, if below it Initiates a Sell order.  Then we give it time.......  The rule would be that the trade has 10 days to become profitable, if at 10 days it hasn't hit the profit target the trade is closed and a new order is opened in the direction of the SMA rule above, then a new 10 days is given.  The profit target could be 20 pips, and say it hits the profit target it would just initiate a new order in the direction based on the SMA above.  The variables that could be tinkered with would be the profit target and the number of days to let the trade run. Drawdown could get awful but could Time win? I would just be interested to see how many profitable trades you get just by placing them and letting Time run its course?  Maybe this has been tried before... let me know your thoughts.  Thanks!

Author