Best way to BACK TEST a system refinement??

eegozi

Active member
Messages
126
Likes
11
So, I have been trading for about 3.5 years now and rapidly getting better and better. Not quite break even or profitable,...YET... but certainly not losing MEGA cash. Basically my system is a MACD type system looking for crossovers of the fastline with the signal line and the 0 line (up = LONG and down = SHORT) and always trading based on 3 different time frames all aligning in the same direction. I am in US and trade mostly stocks and ETFs offered here.

The problem I have had is that my stops are either too close or too far away to allow me to make sufficient $$ on my winners and I end up at other times getting stopped out just before a reversal to skyrocket in the direction I anticipated.

So the dillema I have been working on is reviewing trades and trying to refine it to minimize these issues. I have come up with what SEEMS to be a GREAT refinement. I have "BACK TESTED..." this by starting in 2007 to 2009 almost DAY BY DAY and documenting my "trade" then advancing the chart by 1 day to document what my new stop would have been and where my trade positions would have opened or closed. I figured this was better than using a computer to "back test" as my system is partly a discretionary system based on market sentiment and general fundamental news. Doing this I would have been WILDLY successful with my new refinement. I realize that 2007-2009 was a strong DOWNTREND so I chose 2004-2005 where the market was quite choppy and flat and repeated the process. Here I was less succesful (as expected) but I still would have beat the market.

In the past I have just traded each new refinement. I won or lost money over about 30 -50 trades before I would make any decisions on if my refinement was worth a damn. However, I think this method described above would have worked better.

BOTTOM LINE: HOW DO YOU GUYS "BACKTEST A SYSTEM REFINEMENT"?? ANY BETTER IDEAS

Thanks for your input...
 
WOW,...62 reads and not one comment.:confused:

Doesn't anyone backtest their system?
 
i use matlab, backtesting and optimising is fun but i would never trade any "system" the backtesting and optimizing results in

what i've read countless times on here and from personal experience is that "systems" based on indicators and statistics are WAY harder to trade

the markets are subjective, mathematical precision would be amazing but it's theoretical. one exception to this is if you have millions of dollars and an amazing team of mathematicians to devise some sort of model based on applications from physics, biology, statistics, economics and game theory you might just produce something that could work but why not just sit on the millions...

i use support and resistance and an EMA just to cut out some noise when i first look at the chart and ichi-moku on another chart cos it looks pretty (lol)
 
Top