Expect the Three Deaths
Treat repeated setbacks as tests that refine important high-risk ideas
- Difficulty
- Moderate
- Time to result
- ~ongoing to results
- Steps
- 5
- Confidence
- 93%
High-impact ideas frequently appear dead more than once because early experiments fail, institutions reject them, or the first implementation targets the wrong context. Teams should anticipate these blows instead of treating each one as a unique catastrophe. After a setback, they diagnose whether the failure invalidates the central idea, exposes a correctable implementation problem, or produces a useful discovery for a different path. A revival requires a changed hypothesis and evidence threshold, not mere stubbornness. The framework also includes an explicit stopping rule so romantic attachment to a loonshot does not consume resources indefinitely. Even when the original project ends, teams preserve mechanisms, data, and tools that may enable later work. The goal is resilient learning rather than unconditional persistence.
Origin
Extracted from Masters in Business. Bahcall attributes the phrase to Nobel laureate Sir James Black, who encouraged him after a promising drug experiment failed.
Core principles
- 01Important ideas often suffer several apparent deaths
- 02A setback is information, not automatic proof that the idea is worthless
- 03Persistence must be paired with diagnosis and adaptation
- 04High-risk work requires emotional preparation for repeated disappointment
- 05Projects should be revived through evidence rather than faith alone
How to run it
- 1
Price in repeated setbacks
Assume that a consequential, uncertain idea may appear to die several times. Prepare the team and resource plan for this possibility.
Pro tip Discuss likely failure modes before enthusiasm makes them difficult to acknowledge.
Watch out Anticipating failure does not excuse careless execution.
- 2
Classify each death
Determine whether the setback came from the core premise, the implementation, the test design, timing, or institutional resistance.
Pro tip Ask what evidence changed rather than whether the team still believes.
Watch out Do not label every negative result a temporary setback.
- 3
Extract the learning
Record what the failure revealed and how it changes the next hypothesis. Preserve useful mechanisms, datasets, and techniques.
Pro tip Look for discoveries that remain valuable even if the original objective is abandoned.
Watch out Repeating the same approach is not resilience.
- 4
Set a revival test
Define the smallest experiment capable of showing that a revised path deserves investment. Require stronger evidence after each major setback.
Pro tip Use staged funding or milestones to contain downside.
Watch out Do not restart solely because of sunk costs or executive attachment.
- 5
Revive or stop
Continue when new evidence supports a credible path; stop when the premise is disproven or opportunity cost becomes excessive. Carry forward any reusable discoveries.
Pro tip Make stopping criteria explicit before the next result arrives.
Watch out Persistence without a stopping rule becomes denial.
In the wild
Judah Folkman's proposal to fight tumors by blocking their blood supply was dismissed as foolish and subjected to institutional pressure. Thirty-two years after he proposed it, a major colon-cancer study showed that patients receiving a drug based on his ideas lived longer than previous advanced-disease populations.
→ A repeatedly rejected mechanism became the basis of a clinically valuable treatment strategy.
Elesclomol showed promising early evidence but failed a large metastatic-melanoma trial, ending its development by Bahcall's company. Researchers continued publishing on the molecule and its unusual mitochondrial mechanism, leaving open future applications or related compounds.
→ The original program stopped, but its scientific discovery remained available for later innovation.
Common mistakes
Confusing resilience with repetition
Trying the same failed approach without a revised hypothesis does not generate learning. Every revival should reflect new evidence.
Ignoring terminal evidence
Some failures genuinely disprove a premise or make further work uneconomic. The framework requires stopping rules, not endless persistence.
Discarding all value when a project dies
A failed final product may still reveal mechanisms, tools, or data that enable future work. Preserve those assets deliberately.
Is it for you?
Best for
It is best for high-uncertainty research, innovation, and entrepreneurial projects with potentially large payoffs.
Not ideal for
It is not ideal for routine projects where repeated failure signals negligence, weak controls, or a clearly disproven premise.
From the transcript
“it's not a good drug unless it's been killed three times.”
“you need to expect the fact that you're going to get punches in the stomach many times.”
“most of these things that we take for granted today began life as a loon shot.”
From the episode
Safi Bahcall on Why You Need to Take Crazy Risks
Safi Bahcall