This is a great talk from Howard Marks. What a joy to see a brilliant mind in action. My notes and thoughts follow.
You shouldn’t act as if the things that should happen are the things that will happen. Humility is needed.
Beware of outcome bias. You can’t tell from an investment outcome whether a decision was good or bad because of randomness and luck. The longer the successful track record, the more likely skill is involved.
Balancing Probabilities and Consequences Is Key
Thinking in terms of probabilities and the expected value from a course of action is the first step. But then you must think about the consequences of being wrong. A very small probability event that you can’t survive may make you choose a different course of action.
As Marks puts it, you don’t want to be the skydiver who is right 98% of the time.
“It’s not sufficient to survive on average. We have to survive on the bad days.”
— Howard Marks
Focus on Avoiding Losers
Charles Ellis says that if the game isn’t controllable, it’s better to work to avoid losers than to try for winners.
Risk control is Oaktree’s primary focus. They don’t swing for the fences. Oaktree’s motto: “If we avoid the losers, the winners take care of themselves.” Just lop off the left tail of the probability distribution.
Weed out the problems, like tending a garden. Focus on consistency.
The Relationship Between Price and Value
If you buy a high quality asset but you overpay for it, you are in big trouble. Remember the Nifty Fifty in the late 1960s and early 1970s.
“The secret for success in investing is buying things for less than they are worth.”
— Howard Marks
Planning Assumptions Versus Macro Forecasts
Macro forecasts are worthless, but planning assumptions are necessary with individual companies. Just don’t make big one-way bets based on these assumptions. Assume a range of outcomes and seek survivability.
Don’t Chase the Crowd
You make no money doing the things that everybody wants to do, you make money by doing the things nobody wants to do that then turn out to have value.