A Forex Trading System

Filed under: Forex Investing

“Price Smoothing Algorithms”

The Forex trading system algorithm is a set of procedures that indicate the way the system manages entries, and exists at losses and profits. The steps require to be coded into an information system to allow the automation of trading, and to implement the actual algorithm. The trading system algorithm utilizes price smoothing algorithms.

Price Smoothing

Traders have to change the data series of a price into signals of trading. However, price data is very noisy, and it is almost the same as trying to adjust a radio station over a lot of crackling noise. It is very difficult to differentiate between the essential and the random noise.

Noise is an element of price data that can never be traded; trading noise might result in minimizing your profits. It is important to set the noise and the signals apart. This will enable smoothing the price series, which will result in highlighting the fundamental direction.

This problem is demonstrated in signal processing, and there are some existing advanced techniques, however, traders usually use very crude approaches. The following section will demonstrate a set of the traditional approaches.

Crude Approaches

(The Breakout and the Moving Average)

There are two filters of crude noise, the moving average and the breakout including its variants. The breakout is a signal of entry or exit, it takes place when the existing price exceeds for instance a 20 day high, or decreases below a 20 day low. The attributes that can be pulled out are the number of periods and the balance that the price must be greater or lower than the high or low.

Filtering the noise through this approach is done through a volatility filter. The system tries to reduce the price volatility parameter to noise, and presumes that the price that is greater than a specific level presents an actual signal (not noise). The breakout can be demonstrated in an algorithm through the following equation:

If price + trigger amount > high of n periods then buy

If price – trigger amount < low of n periods then sell

The second approach is the moving average; it is the average of the last, for instance, 20 periods. The outcome will be a line that is smoother than the actual price series, but still falling behind about half the chosen period. The longer the number of periods, the smoother the line is. The existence of more lags to price action, and a short number of periods results in making the line presenting the noise less smooth, but still more reactive to conversions.

Noise is removed by the moving average approach through minimizing the effect of a specific noisy value by averaging out this specific value. This average might have been affected by the extreme values; this will probably result in having an unrealistic average value. In other words, this approach doesn’t work very well with noisy data, except if you selected a long moving average period, which induces lags. The algorithm for a moving average (n period, where n is an integer, e.g. 20) is demonstrated through the following equation:

Sum last n periods, then divide by n

Move forward 1 period, then recalculate

Comments are closed.