In technical analysis, many times, people regard stock price data as normal distribution data, for example, calculate its standard deviation STD. However, in fact, the distribution of stock price data does not conform to the normal distribution. Fisher transform is a method that can transform stock price data into normal distribution.

According to the definition of Fisher transform:

y = 0.5 * ln(( 1+x)/( 1-x))

Using this change to deal with the price, we can get the index in foreign exchange trading: Fisher index.

According to an article published by John Ehlers in the Journal of Technical Analysis of Stock Futures, the Fisher Index is calculated as follows: (here, the calculation period is 10 day).

Price = (highest price+lowest price) /2

Price change = coefficient *2* (minimum price-10 daily minimum)/(maximum price 10 daily minimum-10 daily minimum) -0.5)+( 1- coefficient) * yesterday's price change.

If the price change is > 0.99, the price change is = 0.999.

If the valence changes

Fisher =0.5*Log(( 1+ valence change) /( 1- valence change)) +0.5* Fisher yesterday.

Where the coefficient can be 0.33 or 0.5.

The advantage of Fisher index is that it reduces the lag of commonly used technical indexes. Many trading systems generate trading signals according to technical indicators, but usually the trading signals are seriously lagging behind. How to solve the lag of trading signal? One method is to use Fisher transform. The transaction signal processed in this way can be more sensitive and the performance of the trading system can be improved.

Fisher indicator is mainly used for:

1, which is used to judge the direction and intensity of the trend;

2. Used to generate trading signals: for example, buy when the overall trend is strong and the short-term price is adjusted back.