This thread is being started to document live trading of a swing oriented bot on the NQ. The primary focus will be trade analysis and differentials from back testing, kind of a journal. Right now, I am not looking to changing code, or do further development at the moment. This is an primarily an operational thread; however, discussion on methodologies or about the market in general is certainly welcome.
To give a little background on how the bot operates:
Entries
1) Trend following with a volume component that helps identify the beginning or continuation of a trend. Do not expect the bot to get the best entries.
Exits
1) Volatility band (custom) limit exit
2) Target at which a trailing stop is implemented.
Stops
1) Time exit at market if unprofitable at a certain point. No traditional stops.
Money Management
The bot has a setting that allows me to adjust the number of real-time point gain required to add another contract. The point gain is calculated as if you are trading only one contract.
For example if it is set to 100 points:
The bot will start trading 2 contracts when over 100 points is gained ($20 per point * 100 = $2000). However, to get to 3 contracts it has to trade another 100 actual points ($20 per point * 100 * 2 contracts = $4000). I picked-up this method from somewhere on this board (Thank-You!). It will save you from the exponential crash when a bot fails because you are basing your size on the bots performance over time. If it takes 6 months to go from one contract to two, then it will take 6 months to go from two to three if the bot maintains its performance.
Eye Candy
The below chart is the entire period the bot has been operational plus a few days. It went into operation 1/23/12. This is a back test chart so that I can show the yellow bands which represent my exit limit order prices or trailing stop prices. A bit large scale but should give a better feel for the time in and out of the market.
Back test equity curve
Real equity curve
As you can see up until this point there are discrepancies between back test and real life; however, they are tracking basically the same.
Reasons for discrepancies
1) Figuring out what the synchronize button is good for (oops)
2) Accidentally allowing the bot take a trade in the wrong direction (as an exit) after it did not see the real position get closed the day before at the close. Was not sure how NT7 would handle this. I learned.
3) Missing a signal due to a time stamp issue (IB data) and having to manually enter the trade.
4) Having to roll over contract months while in a trade in early March.
As you can see, having a bot is not totally hands off... :sarcastic:
Performance updates will come slowly...the bot is a slow mover...but I hope this will generate some discussion.