Automated trading provides several unique benefits over traditional, discretionary trading. Some of the most notable are their ability to operate continuously, execute trades quickly and adhere to its trading logic without the emotion that humans have. Human emotion often becomes a hinderance to the ability to succeed at trading due to the fear of loss and greed clouding one’s ability to truly think logically. Of course, there is a lot of depth to human intuition and insights which are not easily emulated in a computer program however I feel that the benefits outweigh those specific limitations.
Developing an automated system has never been easier with tools such as BloodHound and NinjaTrader’s strategy builder. With these you don’t have to strictly rely on a programmer to begin building your system and testing to see how the system performs through historical backtests and live simulations on a paper trading account. The only limitation you have is time and your ability to come up with new ideas to test and build.
While building a system is easy, the biggest hurdle that most new system builders have is understanding the reality and expectations between backtesting (historical, simulated trading) and live trading. There are many best practices in system testing and optimization that are overlooked which will ultimately lead to loss and system failure with unanticipated performance.
Backtesting systems provide a tremendous amount of insight into how a system can perform but is not infallible due to how the backtest processes data and generates execution logs. This can often lead to misrepresentations of performance and expectations in live trading. Even the smallest discrepancies per trade can cause a ripple effect that can turn a system which once looked promising into one that causes significant risk and losses.
It’s imperative that thorough and practical testing is done to give confidence that there are no surprises when you let the program run autonomously. Let Lucrum Trading Systems help with that confidence by employing their years of system testing experience on your next project and schedule a consultation today.
There are many questions to ask yourself; Is your data accurate, do you have enough data, is your slippage factored in appropriately, if using limit orders do you account for possible unfilled orders, is your order management configured to not close or reverse orders which haven’t been filled, are the bar types used showing a hindsight bias, do you have sufficient cushion for intra-trade drawdowns, are the system parameters fit to your testing data, are there ample failsafe in place in the occurrence of specific market conditions, is the system running in a stable environment with redundancy in the event of power/data loss? The list goes on about possible oversights which can dramatically affect a system’s performance moving forward and add additional risk you may not have expected.
The transition between simulated testing and live trading is not one to be treated lightly and a significant amount of time should be invested into building confidence that a system will operate exactly how you intend it to without any surprises. Be sure to have your system analyzed to prevent significant losses before you turn it on.