Saturday, August 20, 2011

Lesson 3: Repeatability Implies Correctness

If something is worth doing, it's worth doing right. I'd like to add, if you can't do it well, it's best not to do it at all. Paul Tudor Jones, a billionaire super trader, recently gave a lecture at my office and asked us this question. Consider the graph below. I've removed the prices and dates from it, and I'm not specifying the stock. Suppose you had to tell me if it was going to go up or down. Would you be buying or shorting this stock?
Some upward stock graph

Some of might you might want to say that there isn't enough information. And you're entirely
correct. I have no interest in predicting the future, and neither should you. We have no business opening a position without doing proper research, backed by a margin of safety. But let's say you HAVE to go long or short. What should you do? How should you go about it? How can you possibly reason that one decision is better than the other?

It just so happens that there is a beautiful answer, a very mathematically oriented one. And it's the only explanation that I feel gives justice to questions like these which lack so much information. But before I get there, let's consider a different question in hopes that it might shed some light.

Let's play a game. We will flip a coin. You are being forced to pick heads or tails.  If I win, you pay me $100,000. If you win, I'll pay you $105,000. Would you like to play? What should you do? How do you feel about games of chance? Is this gambling? It's quite expensive, to coin flip for six figures. But you have a mathematical advantage here. It is beneficial for you to play. You may however, be ruined by any amount of bad luck. Even if you would like to play, it might be prohibitively expensive for you, or anyone. But what if we agree to do this a million times, and we settle the balance at the end of all the flips? You should take the deal because mathematically speaking, you're guaranteed to come out a winner. Now I'm perfectly aware that flipping a coin once is not the same situation as doing it a million times. What I'm trying to show you is that difficult decisions can suddenly become simple decisions if you think about it in terms of repeatability.

Are you seeing the application to our chart problem now? The shape of the chart has absolutely no bearing on your decision. For any stock, you can find some time frame where some portion of their chart looks like the one above. This could be a chart of any stock since I did not specify the stock or time range. Let's say you were asked to look at some chart and pick a direction, 500 different times. Given no other information, except that the market as a whole averages around 8% to 12% per year, you must go long every single time.

If something is correct, you should be doing it every time. In this case, there are two choices. You can go long every single time or go short every time. You will make a market return, or you will lose a market return. At times where you question the correctness of an action, think about it in terms of its repeatability. How does this apply to you? No method of investing guarantees a profit every single time. You do however, have to make sure that your method is proper. If we changed the numbers around in our coin toss to a net loss, any amount of repeatability will add more losses.

1 comment: