Welcome to StockForums.info!   


Technical Analysis questions.......?

Technical Analysis questions.......?

Postby adofo » Mon Dec 31, 2012 11:21 am


1. How do you find what set of indicators work best on a particular investment?
Is there a formalized process for this?

2. What is the difference in usage when it comes to daily, weekly and monthly charts?

Posts: 205
Joined: Mon Jun 25, 2012 5:53 am

Technical Analysis questions.......?

Postby adofo » Mon Dec 31, 2012 11:25 am

One of the best books on the subject of technical market indicators is Colby and Meyer's "Encyclopedia of Technical Market Indicators."

While the book is old, the methodology and conclusions are still valid.

They found that most technical methods don't work. One that seemed to hold up pretty well was Welles Wilder's RSI (Relative Strength Indicator.) I suggest you look at that metric and see if it fits what you need.

The formalized process is called back testing. You take a time series of prices, returns, price changes, relationships, whatever you are charting, and see how your indicator would have done over time, in various markets.

You will find that some indicators work some of the time, depending on market conditions, but that very few work all the time. Some never work.

The problem with most statistical analysis is the people performing it usually ignore the outliers, the extreme conditions that can exist in the markets. Most analysis is performed looking for conditions of central tendency (Things vary around a mean value) and usually analyzed under a Gaussian or normal distribution -- the bell curve. However, there has been, for at least the past 50 years, conclusive evidence that this assumption is wrong. See B. Mandelbrot's work on fractal markets and stable distributions.

In fact, Mandelbrot's work calls into question the whole edifice of modern capital market theory. You may not be able to follow the math, but the conclusions are easy to grasp: Large movements are not once in a century events, but are much more frequent than we like to believe. See also Schrodinger waves in the ocean.

Here's where it gets really interesting. Economic time series are what's called affine. If you look at charts of tick by tick, minute by minute, hour by hour, week by week, month by month and year by year price changes and throw away the units, you won't be able to tell the difference. They are all the same, of self-similar.

If you are familiar with the Dow Theory and the work of Bob Prechter in the Elliot Wave, then the concept of self-similarity will be familiar.

Choose the time frame that fits your personal temperament and have at it. The only problem with choosing a short time frame is that transaction costs may eat you alive. A long time frame may expose you to ruinous runs of down markets or losses. Most non-prop traders stick with an intermediate time frame, one week to two months.

I suggest you look at the recent works that have come out about randomness in the markets. Start with Fooled by Randomness and go from there.

Good luck
Posts: 205
Joined: Mon Jun 25, 2012 5:53 am

Technical Analysis questions.......?

Postby rickie » Mon Dec 31, 2012 11:52 am

You won't like this answer.

There is no "best" set of indicators. There is no "formalized process". TA is more "ART" than "SCIENCE". The key is finding what works best for you.

There is no "difference" between different time frames... except which chart works best for your trading style.

Posts: 45
Joined: Wed Oct 03, 2012 1:27 pm

Return to The Dow Theory


  • Related topics
    Last post