Diagnose problems to get at their roots

A person’s mistake may be a trigger, but the deeper question is why the system allowed the mistake to pass unchecked. If you only blame individuals, you miss the machine.

Diagnosis requires evidence. What exactly happened? Under what conditions? What information was missing? What standard was unclear? When the diagnosis is specific, the solution becomes repeatable.

Diagnosis also demands emotional control. People fear it because it can expose weakness. A healthy culture treats it as problem-solving, not humiliation.

Diagnosis is the step Dalio says people skip most and pay for most. Before designing any fix, you must understand what's really going wrong and why — and specifically get past the proximate cause (the immediate what-happened) to the root cause (the deeper why, almost always a person or a design). His example pattern: 'the report was late' is proximate; 'this person consistently fails to prioritize under pressure' is root.

He argues root causes are usually about people and their wiring, and that facing them requires overriding the instinct to be diplomatic. A good diagnosis names, honestly and specifically, who or what is responsible and in what way — not to punish, but because you cannot redesign a machine whose failure you refuse to describe accurately. Vague or blame-avoiding diagnoses produce vague, ineffective fixes.

Dalio prescribes a repeatable drill: identify the symptoms, trace the chain of cause and effect backward, distinguish proximate from root, and check whether the root is a recurring pattern (a machine issue) or a one-off. Doing this in the open, with the people involved and believability-weighting, keeps the diagnosis honest and turns each failure into durable knowledge about how the organization actually works.

The chapter's warning is against the two failure modes of diagnosis: superficiality (stopping at the proximate cause and treating the symptom) and personalization (making it about blame instead of cause). Get to the root, keep it about the machine, and the design step that follows has something real to fix. Skip it, and you're redecorating around a structural crack.

Dalio offers a practical tell for whether you've reached the root: a good diagnosis lets you say specifically what would have to change — in a person or a design — for the problem never to recur. If your conclusion is only 'we'll try harder next time,' you have stopped at the symptom. The root cause, uncomfortable as it often is, is the one that points to a concrete redesign.