Having built strategies in both Ninja and Tradestation, and being a programmer and computer consultant, I have a natural bent and desire for automated trading strategies.
1. Forces one to trade with a deterministic set of entry and exit rules.
2. One typically can enter a trade much faster with an automated strategy.
3. Help to eliminate the fear factor.
4. Can bring great reward and profits.
5. Consistency - applies to profits ALONG with losses.
6. Generally, the computer should be able to consider more trading factors than the human mind. HOWEVER, the human brain is the BEST analytical computer I have ever seen. Trying to implement the logic and visual scanning ability of the human brain is a DAUNTING task.
1. Trying to replicate in computer code what the brain does is a very difficult task. Not impossible but difficult.
2. You will spend days and weeks building trading code. The hedge funds spend MILLIONS of dollars on programming and analytical resources we can only dream of.
3. Simple concepts will typically only allow one to lose more money faster and CAN EMPTY ones account in a flash. Happens to the big boys too. Back in August, Knight Capital lost $440M in the blink of an eye when they installed a new automated trading program that had a quote "technical glitch"
4. My current automated platform contains 2,000+ lines of Ninja C code. (Before Zondor (BTW, he offers great programming efficiency comments) blasts me here, the main line code is very tight with very compact logic in the OnBarUpdate routine. There are 5 major trade entry techniques with 25 entry filters, ATM like entries for stops, targets, breakeven, etc., and 15 automated exit filters. Obviously, not all the filters are used at once, but I hate to maintain tens of different trading strategies. If your computer is too slow, buy more memory and cpu cycles.
5. WHEN WILL Ninja bring multitasking and multithreading to platform. WHAT A PAIN to only be able to have ONE active window open at time. Hopefully, NT8???????
6. I still believe in Automated trading even with its liabilities.
7. My next project is to rip out all my "ATM" like code (which is there so I can backtest and optimize my settings) and implement the SuperDom and Ninja ATM. That along with chart trader will allow one the best of both worlds. Automated entries/exits along with realtime ability to modify stops, targets and exits.
8. NEVER. NEVER. NEVER walk away from your automated trading platform. Ninja at times does strange things (COULD NEVER BE MY CODE ) and trades do go off into never never land. Just another fully automated fumble fingers effect.
9. BACKTESTS and OPTIMIZATIONS are by basic NATURE, LIARS and DECEIVERS. Cannot tell you how many times my backtest has reported me to be the next millionaire but I'm still waiting. Never use trailing stops and force your strategy to make entries and exits on end of bar close. Not necessarily representative of a live trading environment but it will make you results more accuracte. Read up on how trading programs simulate live price action when only having 4 data points, OHLC, per bar. EVERY trading platform, Ninja, TS, MC, etc exhibit similar problems. You've been warned!
Automated Trading Rule of the Day.
Remember that most entry techniques LAG big time. Many times your entry is 3 to 5 bars after the pivot low/high and the same applies to your exits. Trading CL using a 10 Range chart, I have to assume a 20 to 40 tick cost of trade, not including commissions BTW, whenever I take a trade. What does that mean, you have to have movement which means volatility so I need a market that can move that much. Thats why I trade CL and GC which have multiple moves per day of more than 50 ticks.
And then there is chop, sideways action, etc. Bain of ones existence. I love automated trading in spite of all the shortcomings. Just don't think it is a panacea.
Hope this helps.
The following user says Thank You to djarvis for this post: