Apprenticed Investor: The Folly of Forecasting

That doesn't mean A will happen or C cannot -- only that there's a specific probability of these events occurring out of the millions of ways the future might unfold. Whether any particular scenario plays out is determined by how the countless variables interact over time.

Looking at the future in terms of various probabilities is a productive way to position assets and manage risk. Why? If your expectations for the future recognize that this is but one possible outcome, then you are more likely to consider and plan for other contingencies. It builds in an expectation that other scenarios can and will occur.

For example, one signal I use is to determine when to sell (the subject of a future column) is after a long uptrend is broken. It's not that stocks cannot go higher after breaking their trend line -- they sometimes do. However, most of the time this happens it signals a significant change in institutional behavior towards the stock. Typically, it reflects a shift from fund accumulation to distribution.

For those people who have been enjoying the ride in Google (GOOG - Cramer's Take - Stockpickr) -- especially the near vertical move since April -- this is a high probability strategy. Once that trend line gets broken, say adios muchachos, take your profits and move along.

Again, a trend break is not a guarantee that the upside is finished, but it's a fairly good probability assessment.

The second type of good prediction is the risk-based discussion. These forecasts care less about price targets -- instead, they are an assessment of danger. In other words, to buyers of stocks under the present conditions, when this, that and the other are happening, you are taking on more (or less risk) than is typical. Saying the markets contain more or less risk at given times is a very different statement than: "I think the Dow is going to go to X."