GARCH Improved Nelder Mead MT4 外汇EA

Introduction to GARCH Indicator

GARCH is the short initial for Generalized Autoregressive Conditional Heteroskedasticity and it is the volatility prediction model commonly used in financial industry. GARCH model was first appeared in the work by Danish Economist, Tim Peter Bollerslev in 1986. The 2003 Nobel Prize winner, Robert F Engle also added much contribution for the refinement of GARCH model with Tim’s work. Our GARCH INM predictor took the original method of Nelder Mead for GARCH model building. However, the original Nelder Mead’s method occasionally misses convergence and therefore, we improved the original Nelder Mead method by incorporating intermediate ARMA (Autoregressive Moving Average) step before GARCH step.

How to use GARCH Indicator

The whole purpose of GARCH INM predictor is to improve trading performance. Here are few tips.

1. Our GARCH INM predictor uses point as basic unit. Therefore, the predicted volatility can provide trader scalable idea on future volatility. For example, if GARCH predicted 1120 point (112 pips) for current candle with 95% confidence interval, then you might expect that there is 95% of chance for current market to stay within the 112 pips range.

2. Detecting anomaly market is also another important function with GARCH. If price range for current market moved outside the predicted volatility (95% confidence interval), then this is the anomaly event with 5% Probability to occur.

3. Volatility has very close connection to trend. Increasing volatility often tell you start of trend. Filtering out range market using our GARCH INM predictor, you can aim to enter trend market.

Mathematical Description of GARCH

GARCH INM predictor uses standard GARCH (1, 1) model. Its mathematical description is simply relating the variance of time series to previous variance and previous squared return. In mathematical notation, this would be:

Variance at t = omega + alpha * squared return at t-1 + beta * Variance at t-1

Where time series return y at t = mean + error and t = time

Description for input parameters

• NumberOfObs: Number of observations (data point) to be used for calculation. We recommend minimum 1000 data point.
• UseHighLow: With use of High-Low mode (true), the range is calculated using high – low. If UseHighLow = false, then close – previous close will be used to calculate range. We recommend using default for your trading.
• ConfidenceInterval: Confidence Interval for GARCH prediction. We provide 90%, 95% and 99% confidence interval. This confidence interval provides you probabilistic sense for future market range.
• SafetyFactor: Safety Factor is added for extra safety of trading operation. Safety Factor should be greater than 1. If Safety Factor = 1, then the predicted volatility is equal to original GARCH volatility. If Safety Factor = 1.5, then the predicted volatility will have 50% safe margin over the original GARCH volatility.

Some Trading Ideas with GARCH Nelder Mead

User can adjust Safety Factor from 1.0 to 3.0 to work best for your trading. If GARCH prediction envelope does not touch the price at all, then you can decrease the Safety factor. If GARCH prediction envelope touch the price too often, then you can increase the Safety factor. The main idea is to tune the GARCH prediction to pick extreme bullish or bearish shock in the market. In the shock, use the mean reversion style trading to profit.

Trading Strategy Guide

You can find the detaild trading strategy guide about how to use volatility indicator from the link below. We recommend to read this article before using this volatiltiy indicator for your trading.

Guide for Volatility Indicator

In addition, you can find the trading stratgy for Forex from links below. 

Guide for Volume Spread Analysis

Guide for Harmonic Pattern Detection Indicator

Guide for Momentum Indicator

Guide for Supply Demand Indicator

Guide for Triangle Pattern, Falling Wedge Pattern, Rising Wedge Pattern and Channel

Guide for Elliott Wave Indicator

Important Note

This is the off the shelf product. Therefore, we do not accept any modification or customization request on this product. In addition, we do not provide any code library or any support for your coding for this product.