There are two general categories of limitations on using the past data approach:
1. The past doesn't predict the future.
No matter how much math you apply to how much data, you're still looking backward.
Trying to say what's going to happen based on what has happened is a dangerous game, particularly with anything involving as many non-quantifiable variables as investing.
The best thing you can do with the models is to gain a better understanding of what happened in the past, but you can't be sure it will happen again in much the same way; in fact, you can be pretty sure it won't happen the same way again.
2. There are too many moving parts.
External, internal and personal factors all come into play, and no model can take everything into account.
A stock may have played predictably in the past (and a company's earnings may have played predictably, too, thus the stock price predictability), but what happens when something changes?
What happens when customers suddenly decide they don't like a product anymore or, for that matter, when investors decide they don't like a stock (or gold or corn or a bond or real estate) anymore, or as much as they did?
You can't predict all the factors that influence the future. Nobody can. Again, if you could, who would take the other side of the trade?
No comments:
Post a Comment