Quote from: Goomboo on February 08, 2012, 02:49:15 AM

Optimization and the 10 & 21 moving average crossover system

In order to determine the robustness of our trading system, I have decided to see how our parameters hold up when they are varied. To accomplish this, I used NinjaTrader to do an optimization of a variety of periods for our moving averages and I have graphically shown the results below.

This type of analysis shows us very important information - regions of profitability. For nearly every technical indicator, there are parameters which you input (such as 10 and 21 in the moving averages), which determine how much data it uses in calculating values. If you have a trading system which only works with a certain combination of parameters, this is a curve-fit system, or a system which will probably not hold up well in the future.

In order to insure robustness, or its ability to hold up in the future, a system must be rigorously tested across a variety of markets and a variety of settings. In the chart below, you will see 620 moving average combinations with their respective profitability. It is important to note that the data has been normalized, so the number you see is meaningless - it is just a comparison to all of the other data points. Additionally, it is very important not to look for the best value, but rather look for a region of profitability. Markets change over time and the best results today probably will not be the best results tomorrow. This said, it is important to try and place yourself in the region in which you will more than likely be able to survive and profit in the future.

-Red = below average
-Yellow = average
-Green = above average

The image above powerfully demonstrates the use of a simple technical trading concept applied to the markets. When you are creating a trading system, you must question where the inputs came from. Why did I choose 10 for my short moving average? Why is 21 a good moving average length? In order to protect your money, you must begin asking "why" to the decisions you make. I suggest that you critically analyze, quantify, and backtest every aspect of your decision making so that you can behave in a consistent, confident, and disciplined manner as you trade in the markets.

this seems really useful. the region at the top is obviously the largest/greenest (probably the most robust compared to the other regions).

- but what makes it robust?
- what if the largest region was a single square?