Separate the consensus view from the correct one
To outperform, you must be both non-consensus and right — and second-order thinking is how.
Why it works
Marks’ insight is that an edge requires thinking differently and correctly. First-level conclusions are already priced in by everyone who reached them the easy way, so they offer no advantage. Going to a deeper level — what others are missing and why — is the only path to a view that is both different and better. The mechanism is deliberately searching past the obvious shared answer.
How to do it
- Write the first-level, consensus answer everyone reaches.
- Ask what that view assumes and where it could be wrong.
- Form a second-level view only if you can say why the crowd is missing it.
Evidence
A practitioner principle from investing. It is mechanistic reasoning about edge and consensus, not a studied psychological intervention; the value is in the discipline of distinguishing "different" from "correct". (mechanistic)
This is investing wisdom, not a tested technique; its merit is conceptual clarity, not measured outcomes.
Sources
- Howard Marks, "The Most Important Thing" (second-level thinking)
Common mistake
Being contrarian for its own sake — different but not better. Non-consensus and wrong is worse than consensus.
Practice this with IX Coach
7 days free, then $40/month (~$1.30/day).
More practices for Second-Order Thinking: And Then What?
- Ask “and then what?” three times
Don’t stop at the first consequence — follow the chain at least two steps further.
- Anticipate others’ reactions
Second-order effects often come from how people respond to your first move.
- Map the feedback loop
Ask whether the consequence feeds back and amplifies or dampens itself.
- Check the consequence across time horizons
A choice can be good at one day, bad at one year — name the result at each horizon.
- Keep a second-order decision journal
Record the downstream effects you predicted, then check them later.