How valid is technical analysis?

[jperl]

	Quote:	Soultrader
	Jerry, I want to ask you a question. Can you elaborate on your point of normal distribution? I would like to understand this concept a little better. Thank you

A good general discussion of the Normal or Gaussian distribution can be found here:

Normal distribution - Wikipedia, the free encyclopedia

This is the type of statistical distribution that most TA users assume their TA rules for entry and exit will follow when they do back testing of their rule based trading plan. Unfortunately, most TA users never check this. If they did they would be shocked to discover that their TA rules don't fit a NORMAL or Gaussian distribution of results. The reason for this is quite obvious. The data themselves used to generate the rules, don't follow a Gaussian distribution, so nothing derived from it will either.

The normal distribution is what also occurs on what MARKET PROFILERS call a NORMAL DAY. Market profile is a subset of the volume or price histogram. On a NORMAL DAY, the volume histogram looks like a NORMAL distribution function, hence the name NORMAL DAY.

One of the important characteristics of a NORMAL DAY is the symmetry of the trading volume about the peak in the distribution (the point of control or POC). You can easily tell when a NORMAL DAY is occuring, or when a day is approaching a symmetric distribution (Gaussian or not), by computing the Volume Weighted Average Price (VWAP). When

VWAP=POC we have a symmetric distribution of traded volume.
VWAP>POC we have a skewed distribution of traded volume data toward higher prices.
VWAP<POC we have a skewed distribution of traded volume data toward lower prices.

Knowledge of the VWAP and its dynamic then has some important consequences for price action even when the distribution function cannot be analytically written down.

I dont' have the room here to go into all of the consequences of the VWAP.
At some point perhaps I will contribute and article about it.
JERRY

[drk1t]

TA seems pretty valid to me, I watch the charts, the trends are established, then the stock breaks the trend or it doesn't. That's about it in a nut shell. You can watch the level II and see the bid and ask volume change as it nears the R&S barriers, if it goes through then it's a new ball game until a new trend is established. Think of it like poker, you are delt a hand and then decide what might be best to do with it, same thing when you buy a stock. Some stocks I follow will drop about a certain amount and present a buy or get ahead of themselves and that's a sell, usually, sometimes etc. I look for leading indicators for the sector to help me decide. Then there is Buffet..... if I had his brains and intuition I could be an investor instead of a trader.

[phil895]

I think TA is a very good trading system, for this I must also include concepts that are technical; historical patterns, and price distributions... not just price and volume alone. I used TA almost exclusively sometimes, and the results are much better than FA for short term tradings. For long term setups, it will be another different story. I also find that engineering and math backgrounds are better at TA and economics and accountants are better at FA. When you hear failure rates it's usually from those outside the two academic backgrounds, this doens't mean that engineers and business people are better but they can learn to trade faster, and the success rates are much higher than the rest...

[Soultrader]

	Quote:	jperl
	A good general discussion of the Normal or Gaussian distribution can be found here: Normal distribution - Wikipedia, the free encyclopedia This is the type of statistical distribution that most TA users assume their TA rules for entry and exit will follow when they do back testing of their rule based trading plan. Unfortunately, most TA users never check this. If they did they would be shocked to discover that their TA rules don't fit a NORMAL or Gaussian distribution of results. The reason for this is quite obvious. The data themselves used to generate the rules, don't follow a Gaussian distribution, so nothing derived from it will either. The normal distribution is what also occurs on what MARKET PROFILERS call a NORMAL DAY. Market profile is a subset of the volume or price histogram. On a NORMAL DAY, the volume histogram looks like a NORMAL distribution function, hence the name NORMAL DAY. One of the important characteristics of a NORMAL DAY is the symmetry of the trading volume about the peak in the distribution (the point of control or POC). You can easily tell when a NORMAL DAY is occuring, or when a day is approaching a symmetric distribution (Gaussian or not), by computing the Volume Weighted Average Price (VWAP). When VWAP=POC we have a symmetric distribution of traded volume. VWAP>POC we have a skewed distribution of traded volume data toward higher prices. VWAP<POC we have a skewed distribution of traded volume data toward lower prices. Knowledge of the VWAP and its dynamic then has some important consequences for price action even when the distribution function cannot be analytically written down. I dont' have the room here to go into all of the consequences of the VWAP. At some point perhaps I will contribute and article about it. JERRY

Thank you Jerry. I have never used this reference point but sounds very interesting. I will definitely look deeper into this.