## Saturday, April 28, 2007

### On Technical Analysis or TA

One of the books recommended on this blog has this interesting story about technical analysis or TA. A professor ask a class of students to play a game. They are all given each a pencil, paper and a coin. They are to draw a stock chart with stock price at \$1 starting from Day 1. Every toss of a coin represents how the stock will do for that day, and if it lands on head, the stock price goes up 3%, if it lands on tail, the stock price goes down 3%. And from there, they can plot the stock charts for 100 days (i.e. 100 tosses).

Guess what the resulting charts look like?

They look exactly like actual stock charts with famous patterns like head and shoulders (btw this is not a shampoo brand hor, this is a technical signal in stock charts), double tops, double bottoms, flag formation, cup formation etc.

Why is this so?

According to the professor, in the short run (short run means anything less than 10 yrs hor) stock prices move on positive and negative news, and news flow as such are random, like tossing a coin. Hence by looking at how the stock has moved in the past cannot help you predict what it will do in the future. On every new day, the stock has 50% chance of going up and 50% chance of going down, (like tossing a coin), depending on whether good or bad news will come out on that day. So how can you try to determine which way it will go by seeing what coin toss you have done in the past 10 or 20 days?

Then why is there all those studies about technical analysis, head and shoulders, double tops etc? To give them the benefit of the doubt. I think these things work a bit. They probably work 52% of the time and fail you 48% of the time. Btw these are quite good statistics bcos if you go casino it becomes more like 80:20, meaning 80% chance you will lose.

The main reasons why TA work are probably:

1) self fulfilling prophecy: people think that they work and then strive to make it happen, eg when you see a double bottom, you and 10,000 other TAcians buy the stock, of course it goes up.

2) human/investment crowd psychology does not change: this is the basis for TA as explained by TA textbooks, support and resistance levels are formed bcos investor crowd psychology dictates these levels until the next driver pushes the stocks to another paradigm.

But having said that, we must understand that stock markets are complex systems and hence TA can only help you win 52% of the time. There are times that TA can drive the stock prices, and there are times other information like macro outlook, earnings announcement, sentiments etc drive the stocks.

TA is only marginally useful in predicting short run stock peformance and not useful at all in predicting long term stock performance. On the other hand, value investing has zero use in predicting short run stock performance but gives you a little bit of an edge in predicting long term stock performance (probably 54% or so). The good thing about value investing is if you have done work homework, even if you are wrong, you will not lose your shirt.