Monday, November 23, 2009

Pros and Cons of Momentum Investing

Stock chart showing levels of support (4,5,6, ...Image via Wikipedia
There are a LOT of myths passed around by various investment advisors, and some involve timing the market.

I've actually heard one speaker at a seminar who actually claims that "Santa Claus Rally", "Sell in May, Stay Away" and so on, are valid advice. You probably heard some others as well. And a lot of months have a myth attached to them. There's the "January Effect", "October Surprise", "Take Profit Just before St. Patrick's Day", and so on.

What I can say is... What utter... nonsense! If trend exists, then you can bet money on someone anticipating it, thus destroying the myth.

Let us take Santa Claus Rally as an example. Santa Claus Rally basically says, "Stocks always rally in December." Just think about it... If you KNOW the stocks will rise in December, you will SELL in December to take the profit! And the selling will destroy the myth, because selling means prices will DROP, thus countering the effect of the alleged December rise!

In fact, the same reasoning applies to all attempts to tie a certain month to ups and downs in the stock market, no matter what the month is.

Frankly, it sounds good (until you do some fact checking and deeper thinking, like I did here), and it's simple and easy to understand, just what seminar speakers need. Whether its actually true or not is a different matter altogether. It may have been true in the past, but that is no guarantee that it will continue to be true in the future.

Another problem is trend prediction without studying the underlying reason for the trend. The word "trend" here can go by many forms... but the most popular term is probably "cycle". One such famous "cycle" is called the Presidential cycle. The claim is the stock market will rise upon the inaugural of the new president, rise to a peak in the second year, then drop in the third to the bottom, and recover in the final year of the term.

If you don' thave a reason, just a trend, you may as well as forget the whole thing, because something doesn't happen just because it had happend yesterday, and the day before that, and the day before that. Or to give another example, you observe that the sun rises from the east, day after day. So you predict the sun will rise from the east tomorrow. Surely you'd be right, but the REASON of your prediction is NOT because it has always risen from the east before. "It's always been that way" is not a "reason". The actual reason is due to the earth's rotation. There is a big difference: one is simply extrapolating a trend, while the other is actually UNDERSTANDING the trend's cause/reason.

While it's true that you don't need to understand the pattern to utilize it, if you don't KNOW WHY the pattern occurs, can you really predict this really IS the beginning of the pattern?

Trend analyzers, esp. the "technical analysis" folks, loves to spot "patterns" in the trading chaos for buy or sell signals, or at least what they THINK are signals. And they pour over all the historical data they can find for patterns they claim will prove their hypotheses. However, that's what I called "lazy research". They are searching for data that fits their hypotheses, not analyzing the data.

In real research, the cycle goes like this

Start -> form hypothesis -> devise experiment / obtain data -> analyze data, proves / disproves hypothesis -> update hypothesis -> restart

And you repeat the cycle until you have a THEORY that will explain ALL of data, not just some of it.

However, if you do "lazy research", you end up with

Start -> form hypothesis -> devise experiment / obtain data -> pick out data that supports your hypothesis -> hypothesis "proven"

If you can pick what data to use, you can prove almost anything. And frankly, the llazy researchers WANT the stuff proven, because it would validate themselves. Then they can write books that cost $15-20 and you will buy them, hoping that they actually KNOW something.

And we've already explained that there's a difference between trend extrapolation and trend cause and effect. Trend extrapolation is just that: a prediction. It has no LOGIC behind it. It just has the HOW, but not the WHY. And if you don't know the WHY, you cannot be sure that it WILL happen again, esp. with something as volatile as the stock market.

I've read a dozen or more different papers and books on technical analysis, and so far I am not impressed. I've read "gap analysis", "1-2-3 pattern", "Japanese Candlestick", and many many more. Some even went as far as explaining each pattern as manifestations of investor psychological states. However, most have one or two examples, and some don't even name a real example, just a "theoretical illustration". If it's a real pattern, it would apply to all stocks, yes? But no, it only applies to the specific examples they found. That is lazy research to me. They leave the task of actually spotting the patterns up to you, the investors, and if the pattern didn't work out, well, they've already sold you the book.

The most complicated, of course, is Japanese Candlestick charts, and how to read them. They have those fancy Japanese names for the "patterns", and there are dozens and dozens of them, and most of them require yet ANOTHER pattern as "confirmation"! To me, that just smacks of spotting something in nothing. In other words, fitting a pattern onto chaos. I mean, if people can spot image of Virgin Mary on a piece of grilled cheese sandwich, or Jesus on the wood grain of a cheap door, what ELSE are you going to see, if you EXPECT to see something? And they made a SCIENCE out of that?

And of course, they have no DATA to back up the accuracy of their predictions. Most Japanese Candlestick books just explains it's ancient Japanese, long history, blah blah blah. Sorry, but I want to see some REAL data. Like some REAL predictions, performed by UNRELATED PEOPLE, and how they turned out. I've yet to see any, just more "lazy research" into more "pseudo-science".

If you enjoy technical analysis, i.e. spotting patterns in the chaos, then candlestick charts can be a powerful tool in your arsenal. However, it has been in my experience that candlestick charts, while visually more pleasing than lines with tickmarks, do not really convey any advantages. All these so-called patterns are spotting something out of nothing. Heck, some books even claim there are over SEVENTY patterns! And they can't even agree that are the basic patterns... Most pin it at about 18, but some say 15, 16, or 17. Just a search on Google will tell you that there sure are a LOT of books on the subject, and a lot of articles as well, but all claims "amazing accuracy" with little data to back up the claim.

Remember what Reagan said... "Trust, but verify". Take all information you receive, but verify all of them through other sources and make up your own conclusions. yes, even this article.

Reblog this post [with Zemanta]