[With] closet indexing....you're paying a manager a fortune and he has 85% of his assets invested parallel to the indexes. If you have such a system, you're being played for a sucker.

The game of investing is one of making better predictions about the future than other people. How are you going to do that? One way is to limit your tries to areas of competence. If you try to predict the future of everything, you attempt too much.

If the same family were always on the bottom, then you'd have big resentments. But if DuPonts go down and Pampered Chef up, [that's good]. That much churn makes people think the system is fairer. Buffett: We don't like churn now, but we liked it more 30-40 years ago.

Hard work, honesty, if you keep at it, will get you almost anything.

In investment management today, everybody wants not only to win, but to have a yearly outcome path that never diverges very much from a standard path except on the upside. Well, that is a very artificial, crazy construct. That's the equivalent in investment management to the custom of binding the feet of Chinese women

The normal expectancy of the average investor - for example, the pension funds of AT&T or IBM - is 6% for a long time.

[GEICO] got to thinking that, because they were making a lot of money, they knew everything. And they suffered huge losses. All they had to do was to cut out all the folly and go back to the perfectly wonderful business that was lying there.