ankit,
Thanks again for sending the data. There is some setup information I need to place in the information screen on Amibroker for the Nifty_F1 data you sent me. I need to know:
Round Lot Size
Margin Deposit
Tick Size
Point Value
Once I get those values I can run some backtests using the data.
I got my code reworked and it slices through the backtest and optimization process very fast now.
This enabled me to finally run an exhaustive analysis on the S&P 500 E-mini futures contract (ES), only to find the ORB's do not work consistently. In fact I found they fail to work for far greater amount of times then when they are profitable. It will likely blowout your account if you were to trade it mechanically for a period of years.
This was a pretty hard blow given all the work I've put into this. At least on ES, ORB's have been on a winning streak for most of this year. In fact, if you had traded them every single day with no filter whatsoever you would have made a massive profit for the first half of 2009.
In my longer term analysis I used 2 years worth of ES data spanning the years 2006 and 2007. I optimized using just about everything imaginable:
Varied the Average Range Min from 8 to 15 days
Varied the profit target using a multiplier from 0.5 to 1.35
Varied the Stop Amount using a multiplier from 0.25 to 1.2
Varied the BreakOut levels, by applying a multiplier from 0.7 to 1.3
I used filters such as ID, NR4, NR7, 2BarNR, 3BarNR, BullHook, BearHook.... (Including various combinations such as ID/NR4, NR4/NR7).
I even tried taking trades on days where the previous 2-3 days had no patterns whatsoever. (that actually worked from 1/1/06 through 6/30/06, and then it blew up during the rest of the in sample period)
I also tested taking trades on a specific day of the week. Found a great combo on the 1/1/06 through 6/30/06 time period. Trade the NR4 and NR7 combo only on Mondays. Worked great until I expanded it to the rest of my in-sample period.
So, while the ORB trades on ES are presently hitting very well, there is no telling when they stop working again. And how is one to know? How do you tell the difference between a short string of losing trades and the end of the run???
Over the previous 12 months I have coded and tested dozens of different trading methodologies and none have proved to work in the long term. It seems to me the only reliable tool to use is the human mind. One that is highly trained and experienced in the market, that knows instinctively when to switch tactics. Anyone know where I can get one of those? My search for a purely mechanical trading system, one which performs consistently over the long term has so far met with no result.
Once I receive the setup data required to test the ORBs on the Nifty futures data I will run it through the same vigorous analysis and let you know my results. But in my opinion, one years worth of data is not enough to prove that the ORBs are here to stay.
Good trading everyone
Dry
