Forecasting is as much a philosophy of time and space as it is a series of methodologies. Understand the philosophy and you will understand how to forecast better, which methods to choose in different situations, and you will never give up on forecasting. Here is my philosophy of forecasting:
1) All sentient species were blessed with an ability to see the present and to a greater or lesser degree remember the past. None are known to have been given the ability to see into the future. We can use our knowledge of the past to project into the future. We can anticipate the future. But we cannot see the future, we can only predict it. Yet it is essential to predict the future.
2) Any forecast is based either on an analysis of the past projecting forward, the present projecting forward, or a random choice. The former two are inherently more accurate than the latter. The latter will only be used when the past or present gives no adequate understanding about the future. When people give up on forecasting, because of inaccuracy, they are still forecasting when making decisions. They have only changed their methods from evaluative to random choice. Not knowing you are forecasting is dangerous, because it engenders false confidence.
3) All actions in the present are based on a prior forecast. If you went to work this morning, you were unconsciously forecasting many things: that you would get there without an accident, that you will be paid, that the check will not bounce, and that the currency will have some value to name a few. If you were to get in an accident, would you stop going to work? Probably not. For the latter to occur, you would have to make dramatic changes to your forecast of the value of going to work.
4) Every living thing must forecast. A lion chasing a zebra on the Serengeti plane is forecasting a meal. It won’t chase rocks, even though they are far easier to catch, because it can easily forecast no nutrition. Moreover, a lion will learn from each chase to better forecast the right moves that will best bring down its prey. If in chasing a zebra; it sees that this zebra has consistently turned to the right; it prepares for another; the zebra turns left; and the lion misses - - does it stop chasing zebras convinced that it can’t forecast?
5) The value of a forecast is inversely proportional to its probable accuracy (given that all similar forecasts are equal). Proof: The value of a forecast of the sun rising tomorrow is nil. The value of a forecast of the Sun Microsystems stock rising tomorrow is far higher. This principle can also be demonstrated by the difference in job opportunities and salaries of sun rising analysts, versus stock analysts. Astronomers are not paid more either, even though they can forecast the movement of all the stars with amazing precision.
6) Some things are easier to forecast than others. It is a safe bet that if you flip a light switch up, the lights will go on or that the sun will rise tomorrow. Other things like chip manufacturing markets are more difficult. Moreover, it is far more difficult today than it was twenty years ago. In general, forecasting things where human nature is involved is much more difficult than forecasts of nature or science. That’s because humans have free will. As my forecasting professor once told a group of scientists, “Many of you could precisely place where the stars will be at midnight on Friday. How many of you can precisely place where you teenagers will be at the same time?
7) Forecasting anything becomes easier with more historical data and greater computational power. In olden times, forecasters of the seasons were highly revered. Now anyone can figure this out.
8) Forecasting is the art of comprehending as much of the trends and data as possible with a dash of luck. When you can’t forecast accurately, it is important to know this and then to be positioned to take advantage of luck as it unfolds. Admiral Nelson did this in the days that led up to Trafalgar. Admiral Nimitz did this in the days that led up to Midway. Gordon Moore did this in positioning Intel as a start-up. Bob Graham did it at Applied Materials. Peter Rose did it at Extrion and Nova. Jim Healy did it at Trillium. It is like fishing: you often can’t see the fish, but your odds are greatly improved if you fish in pools where the fish are.
9) Pick your methods based on the probable accuracy of what you are trying to forecast. There are two primary methods: statistics-based and indicator-based. Statistics-based forecasting works best in stable and predictable markets because it relies on long historical series. Indicator-based forecasting works best with unstable and hard-to-predict markets because it relies on instantaneous events.
10) If forecast accuracy in a given market is discernibly not equal, then the more accurate forecasts will be of greater value. One can pick forecasters by their track record. However, rule 1 can often prove this wrong. Since no one can see the future, there is no way of determining exactly which forecaster will actually be the most accurate the next time around. Nobody’s perfect. Don’t trust those that claim perfection. They have only revised their views after the fact.
11) Some are better forecasters than others. Picking a forecast from someone else is made more difficult because everybody has a forecast. Most won’t even be close. Few track what they really believed, which allows their mind to easily distort what they thought. Those that track their forecasts get better. It is learning from errors that improve forecasting ability.
12) Prepare to be wrong. Forecasting is not about being right or wrong. It’s about being less wrong, learning, and improving. When you are exactly right, it is simply luck.
13) The value of a forecast decreases to zero when the target time has passed. The value of an accurate forecast of stock closing prices on the following day is virtually infinite. The value after the close is less than the cost of picking up the actual number from any news source.
14) Forecasting is at the core of planning and planning is at the core of good results. Never forget the five P’s: Proper Planning Prevents Poor Performance.