I ran an optimization using a simple MA cross-over strategy on HAR between May 1, 2011 to Jan 20, 2012; about 20.5 months.
The rules were:
1.) Buy 100 shares when the optimized fast MA crosses above the optimized slow MA within the last 4 bars provided the close of day is above the daily pivot point.
2.) Sell Short 100 shares when the optimized fast MA crosses below the optimized slow MA within the last 4 bars provided the close of day is below the daily pivot.
3.) Close long positions if the price closes below the optimizsd slow MA.
4.) Close short position if the price closes above the optimized slow MA.
5.) Exit positions on an optimized stoploss $ amount.
6.) Take profits on an optimized $ profit target.
The results were:
The top ten winners used the 10 day SMA and the 40 day SMA for fast and slow, respectively.
The optimum stop-loss's varied between $175 and $275.
The optimum take-profit targets varied between $700 and $1000.
The Net Profits before commissions varied between $4244 and $4624 (trading 100 shares).
The top performer used the 10 and 40 SMA's; a stop-loss of $175; and a take-profit target of $800.
The gross profits were $5203 and gross losses were $579 giving a net profit of $4624 before commissions.
There was a total of 18 trades and 36 executions (36 commissions paid). 9 were long and 9 short.
13 or 72.22% of the trades were profitable.
5 or 55.56% of the long trades were profitable.
8 or 88.89% of the short trades were profitable.
The maximum drawdown for all trades was $345; $392 for long trades; and $12 for short trades.
The largest winning long and short trades were $800; the optimized take profit target.
The largest losing long trade was the optimized stop-loss amount of $175.
The largest losing short trade was $12.
A profit of $4600 over 20 months is not bad considering that the maximum investment at any one time over this time period would have been about $2500 to $5000. HAR has a history of being a a good trender. I will probably run it for a longer duration and see what I get; but I'm liking what I'm seeing so far.
The results highlight the importance of cutting losses and letting profits run. The optimized stoploss at $175 represents 7% of $2500 and 3.5% of $5000. I know for a fact that some don't give this much thought but consider that if a long goes from $10 to $9 it is a 10% loss. To get back to even it needs to rise 11.1%. That is acceptable. But it gets worse fast. If it goes down to $8, a 20% drop it needs to rise 25%. If it drops to $6; a 40% drop it needs to rise 67% to get back to even and if it drops 50% it needs to rise 100%, double in price, to get back to even....forget about it, it could take years. SKUL has dropped over 50% in just 5 months and may take years if ever to get back to the $16's.. My intent is not to rub in the salt but to point out the utter futility and the missed opportunities of hanging on waiting for your ship to come in. Come back to this when it shows a sign of strength.
Okay, the same model (same rules as for HAR above) run on trading 100 shares of SKUL beginning July 20, 2011 (1st day of public trading). Except the model didn't trade until Oct. 24, 2011.
Optimum parameters are the 13 SMA cross 40 SMA with $200 stop-loss and $300 profit target #$%$ risk/reward compared to the others). For 100 shares the net profit would have been $921. There were 6 trades, 12 executions/commissions. The average investment was $1443. That's a 64% return over one year! The first entry was 10/24/11 and the last exit was optimized $300 profit taking on 11/2/12 @ $10.01. So that is a 64% annual return! A great return but high risk.
3 long trades; 2 profitable. 3 short trades; all profitable. The short trades netted $619 and the long trades netted $302.
So based on that, I guess you wait for the 13 to cross above the 40 to get long. Unfortunately they've been diverging since the last drop.
NOTE: One difference from the HAR rules 3 and 4: Instead of exiting positions when the PRICE crosses the slow (40) MA, exit when the 2-day MA crosses the 40 MA. This thing is volatile, well, has been anyway.
Also, a caveat: the model is only based on about 18 months of data; a small sample population.
I just ran the same optimization on WHR, without any tweaking, from Jan 1, 2000; 13 years. Trading 100 shares with crosses of the 12 SMA and the 45 SMA within the previous 4 days with a close on the right side of the daily pivot, a stop-loss of $400 and take-profit target of $3500 resulted in a net profit over 13 years of $21,472. Assuming an average investment at any one time of about $7000, that's a a 400% return or an annual return of 11.25%.
Again note the huge difference between the optimized stop loss and the take-profit target; $400 vs. $3500.
FYI, the model took the $3500 optimized profit on Oct 24th, 2012 at $96.23 and has since been flat. The current price is $102.31 after reaching a high of 107.94. The model is not greedy.