Financial Market Chart Analysis Help

[dasilvja]

Hi,

This is my first time posting here and I'm not 100% sure this is the correct category. If not, please direct me to where I should go. I'd love to get deeper into all this. Anyways, to the problem.

When analyzing a price chart, traders know that levels where prices reverse are significant. The more times prices reverse at a level, the more significant it is. These levels are known as support and resistance levels. Obviously, when price reacts at these levels it's seldom the exact same level. However, it will be within a range of the level. Rather than think of it as a specific level, we can consider it an area.

I am able to record all the turning points myself. This in itself is very subjective so the support and resistance levels will be relative to my own identification of the turning points. So, once I've got a list of these turning points, is there a formula or function I can use to identify the most "popular" areas?

I know there are a lot of variables here. I just thought I'd start with this and see what I learn then go from there. To help illustrate my goal imagine that I have listed the following levels:

2, 9, 21, 18, 17, 16, 35, 68, 72, 66, 70, 99

Just looking this over quickly you can see there are a couple areas of significance. The area around 18 and the area around 70. I'd like to input all the levels, and input a formula into excel to identify the groups that are close together then create an average of those groups. That, I believe will be how I can identify these levels.

Any help is appreciated. Thank you in advance.

 

[JeffM]

As you understand, what you are doing is subjective. (And in the view of many, but not all, economists theoretically wrong. I do not fully agree with that majority, but I thought I ought to warn you of the professional academic view.)

In any case, you are looking at time series, but in your example time is ignored. Moreover your example is much more dispersed than is typical of prices so it is not a good example.

What I am going to suggest requires a lot of data points.

Organize your prices in a column in date order. To the left of that column build a column of positive integers, 1, 2, 3, etc.

Now use the graphing feature of excel to graph the data, price on the vertical axis and time on the horizontal axis.

OK. You should get what looks like a standard price chart. Now comes the subjective part. There may be periods when prices seem to be rising fairly consistently and periods when they seem to be falling fairly consistently. The periods between those periods with obvious trends are periods that show, according to technical analysis, resistance and support levels. Identify such periods by eye (you could use mathematical techniques to refine your eye, but that is why Wall Street pays big money to "quants.") Now simply take the arithmetic mean and the population standard deviation of the prices in those periods using the excel functions for computing those statistics. Finally, divide the standard deviation by the mean to get the relative standard deviation. Assuming you have enough points in such a period, the mean will represent a typical price in the support or resistance range, and the relative standard deviation will indicate how stable that price is (the lower the relative standard deviation, the more stable is that price.)

One thing to think about are double tops or bottoms or even triple tops or bottoms. You can combine the data from periods that seem to show repeated tops or bottoms. This is in accordance with what technical analyts do.

You could try getting much more sophisticated by using the ratio of individual stock prices to an appropriate broad market index rather than the prices themselves, but that is about where I start charging money for doing consulting work.

 

[dasilvja]

I appreciate the reply but it's not quite what I'm going for. I want to remove all subjectivity and calculate the levels that are historically most significant in a given period of time. The only subjectivity is how I decide where the turning points are. That's all the numbers are going to represent. So they wouldn't move in an ascending or descending order.

I have a different formula and method to calculate these turning points and to collect those figures for me. Once I have them though, I need a formula to automatically calculate the price levels (or price areas) that have the most activity. So on the list of numbers, groups of numbers that are close together would represent an area of significance. Time is irrelevant. Chronological order is irrelevant. I just want to group the areas of price where there were the most "reactions" (each number in the list represents a reaction).

I've heard of quants before no idea what that is. I assume not a resource at my disposal. lol

 

[JeffM]

The problem is that you cannot remove subjectivity entirely. What is historically most important is not a just question of computation but rather a question of human judgment informed by but not determined by both historical knowledge and quantitative data. The moment you decide to exclude some data, you have made an at least partially subjective determination that the data excluded is completely irrelevant.

What can be done is to use mathematics to suggest or test test specific theories. The very well trained mathematicians on Wall Street who do that work are disparagingly called "quants," short for "quantitive analysts."

I hope you find what you are looking for.