I've been doing this for more than 25 years and have learned never to take mistakes lightly. What's most important for us, though, is to stay focused on the discipline of only investing in companies with the characteristics of leaders, laggards, and innovators that we've seen work as investments over a long period of time. [...]

We've many times sold way too soon. To try to avoid this, we've forced ourselves to look over long periods at where margin and sentiment peaks have been in individual stocks, to really vet how high something might reasonably go.

We've developed a quantitative model that's designed to reflect the attractiveness of every company in our 1,000-company database based on three factors: expected earnings growth relative to the P/E multiple (the higher the better), valuation relative the company's past history (the lower the better), and the trend in consensus Wall Street estimates (upward movement gives [...]

Our experience shows there's a positive correlation between improvements in a company's return on invested capital and its stock performance. That makes sense, given that a company's earnings today are the result of project spending it made in the past. We obsess over ascribing value to today's capital projects, as well as on deconstructing businesses [...]

Small-cap investing can be more labor intensive due to the sheer number of companies, but at the same time you can more quickly know just about everything you need to know about a company to make an investment judgment. I can't say that in looking at a company like AIG, for example.