# Forecasting, Breaking Questions Down, and Probabilities

I'm currently reading *Superforecasters* by Philip E. Tetlock and Dan Gardner (Crown Publishers, 2015). The book addresses how to improve forecasting accuracy. Two ways to do this are: (1) breaking difficult questions or predictions down into parts that can be answered or predicted more easily; and (2) assigning specific probabilities to different possible outcomes rather than simply defaulting to yes, no, or maybe (few people take time to think in terms of specific probabilities). *Id.*

If you're an investor these are helpful concepts to keep in mind. So an investor in Apple might have difficulty answering a question like "will Apple's earnings grow or at least stay flat over the next three to five years?", but this question becomes easier if you break it into parts:

- Will earnings from iPhone sales in the United States grow or at least stay flat over this time?
- Will earnings from iPhone sales in China grow or at least stay flat over this time?
- Will earnings from iPhone sales in India grow or at least stay flat over this time?
- Will earnings from iPhone sales in Brazil grow or at least stay flat over this time?
- Will collective earnings from iPad sales, Macs, and Apple Watches grow or at least stay flat over this time?
- Will Apple move downmarket
*as necessary*to grow earnings in China, India, and Brazil?

And so on . . .

You then assign a probability to each of these question subparts, digging for information that allows you to be as specific and granular as possible (*e.g.*, a 63% estimate is better than a 60% or 70% estimate).

Applying *Superforecasters*, Warren Buffett and Charlie Munger seem to follow a concentrated portfolio strategy so that they can carefully weigh the probability that *each investment* will succeed, thereby improving forecasting accuracy. This strategy would be impractical if Buffett and Munger diversified through a wide array of stocks -- too many holdings would make it difficult for them to carefully weigh each investment's probability of success, thereby hurting forecasting accuracy. This all dovetails with Buffett's famous quote that he'd rather be "certain of a good result [a high probability event] than hopeful of a great one [a 50/50 chance or worse]."

The author owns stock shares of Apple and Berkshire Hathaway.