An Introduction to Trading Systems
Trading Systems are built on algorithms which are developed by computer scientists and mathematicians. Algorithms are a step-by-step procedure of how to do a procedure. For example, there are algorithms for booking trades, pricing trades, portfolio information. When you make tea or coffee, this is an algorithm, you follow a set of rules. These algorithms are developed to execute buy and sell orders, on a particular date. These algorithms can be also be used to execute limit buy and limit sell orders. Front office systems are used to price transactions, positions, risks and profit and loss. These systems are mainly used by the trading/sales division in an investment bank and by hedge fund managers. The front office systems are quite complicated; they can be used for individual transactions or portfolios of complex deals. Middle office systems are used by the risk control division to monitor risk. Back office systems are used for settlement of transactions and market operations such as option exercises, expiry, fixings, and early terminations (early termination of a position).
Recently algorithmic trading has become more desirable, mainly due to the economic climate. Computers are cost-efficient and faster than humans. A subset of algorithmic trading is HFT (High Frequency trading). This system trades on a smaller time-frame, usually on a ticker or a 1 minute chart. Many strategies can be implemented into these algorithms including moving averages cross-over. If you are building a system for the FX market you will need to incorporate economic events, as these events will make the market more volatile. In addition, these algorithms have a short-shelf life. Every week, algorithms are created to keep up with the changing market conditions and it is often that these algorithms can become outdated, even before they are put on the shelf. Most of algorithmic trading is developed in-house (keeping secrets of models in-house!) and some are very complicated and have automatic sales/purchases in accordance with model with direct connections to the markets with almost zero latency.
There are many trading systems, including Murex, Calypso, Wall Street and Summit. We can create algorithms in MetaTrader using MetaEditor and here we can write some code to tell the computer what to do at a particular price. The programming language used in MetaEditor is MQL# (# is the version of your MetaTrader). The programming language is fairly easy to learn. Once we have built an algorithm, we can back-test it. In addition, we can base a trading system on the Twitter API (application programming interface) which is a very clever system. Twitter is a faster news source than the news channel on the TV and the algorithm trawls through the news feed and uses this data to base trades.
Back testing, is similar to data-mining. We can test the algorithm on previous data and we can then see how profitable the algorithm is. However, this is very risky as the algorithm is based on previous data and it does not necessarily mean that the algorithm will make a profit in the future. In addition, before starting to build an algorithm it is good practice to be familiar with indicators you are using so that you can set parameters for these indicators. We can also back-test these algorithms in MetaTrader.
CFDs, spread betting and FX can result in losses exceeding your initial deposit. They are not suitable for everyone, so please ensure you understand the risks. Seek independent financial advice if necessary.
Nothing in this article should be considered a personal recommendation. It does not account for your personal circumstances or appetite for risk.
