More bad news.. another failed mechanical system

I tried another mechanical system to see if I can beat the results for my discretionary practice. This system is straightforward:

• Using 2.5*ATR(14), if price is currently below this ATR, buy when it closes above it. Conversely, if price is currently above this ATR, sell when it closes below it.
• An additional criteria is that the closing price has to be above EMA50 for buy and below it for sell.
• Initial stop loss is based on 2.5*ATR(14).
• For buys, exit if either of the following criteria are met:
  • initial stop loss is breached
  • price closes below the 2.5*ATR(14) trailing stop
  • price moves more than 3% below the 2.5*ATR(14) trailing stop
• For sells, exit if either of the following criteria are met:
  • initial stop loss is breached
  • price closes above the 2.5*ATR(14) trailing stop
  • price moves more than 3% above the 2.5*ATR(14) trailing stop

I tested this system on MU and SWN.

For MU, using strategy 4 for position sizing (ignoring days to exit), I ended up with a final balance of $150,141.80 and a maximum drawdown of -20.14%.

For the discretionary practice sessions with MU, my first practice session resulted in a final balance of $261,086.86 and a maximum drawdown of -16.53%. My second practice session resulted in a final balance of $104,445.67 and a maximum drawdown of -19.99%.

Therefore, the mechanical system beat the second practice session but performed worst than the first practice session. Given the wide disparity between the results for the two discretionary practice sessions, I am reasonably satisfied with this mechanical system.

That is.. until I tested it out on SWN as well. This time, despite catching a few good trends, the results are abysmal. I ended up with a final balance of $63,548.79 and a maximum drawdown of -37.00%.

For the discretionary practice sessions with SWN, I ended up with a final balance of $98,514.13 and $114,660.12, with a maximum drawdown of -17.76% and -27.73% for the first and second sessions, respectively.

Hence, the result I got with the mechanical system is significantly worse.

Moving forward

I am still keen to design a mechanical system because this gets me more excited than a discretionary system. Designing a mechanical system is like solving a math puzzle, and that is what I am most passionate about.

However, I am worried that I am wasting my time. I do not want to look back in my 80s and conclude that I have wasted my life trying to chase the ‘holy grail’.

I’ve tried designing a mechanical system before without any luck. I only started to make money when I used fundamental analysis. But I want to make more a decent income from trading (and not just buy and hold and wait for the stocks to appreciate in value) and I figure I have no edge in fundamental analysis. Therefore, I am back to technical analysis.

Given how inconsistent (though still consistently profitable) I am in using a discretionary system, I figure a mechanical system works better for me. Using a discretionary system is stressful and gives me gastric issues (or is the oily tingkat food the culprit?).

So will I be able to develop a profitable mechanical system with an edge? I have to try, right?

My next step is to code my Darvas trailing stop indicator properly. Then I can code a system based on that. Hopefully I can get this done by Fri.

Well, no more practice sessions for this week, I’ll focus on coding the trailing stop indicator and get ready for CNY, plus think more about what really makes me happy. Been kind of depressed and moody lately.

Update

It’s amazing how some concentration goes a long way. I read through the Darvas indicator carefully and managed to correct the previous issues.

Next up, I’ll check to ensure the indicator does not suffer from any ‘Look-forward bias’. I will also code a mechanical system that

• Looks to enter a trade when a n-day high or low is breached
• Requires price to be at most x% from a moving average
• Exit using the darvas indicator

With this system, I am hoping to take advantage of the two edges I identified previously:

• Getting out asap when I am wrong, and riding a trend for as long as possible
• Trading a pullback to get in at a better price