Estimate how long a project will take
When to use
You need to estimate the time, effort, or cost of a project before committing to a deadline.
What you'll get
A realistic time estimate with an optimistic / most likely / pessimistic range drawn from comparable projects, plus a recommendation for what number to commit to externally.
The prompt
I need to estimate the time [and/or cost] for [PROJECT]. Here's what's involved: [DESCRIPTION OF SCOPE].
My instinctive estimate: [YOUR ESTIMATE].
Before accepting that, let's check it properly:
- My instinctive estimate is almost certainly too optimistic. This is one of the most reliably replicated findings in psychology — people systematically underestimate how long things take, especially for projects they're excited about or personally invested in. Please treat my estimate with that in mind.
- Ask me about the last 2-3 comparable projects and how long they actually took vs. what I estimated upfront. That track record is more predictive than my current plan.
- Help me think about this from the outside: for projects of this type and complexity, what does the typical range of outcomes look like, independent of what I hope this particular one will look like?
- Walk me through the categories of delay most common in work like this — dependencies, scope creep, review cycles, unexpected technical issues — and ask which ones apply here and how I've accounted for them.
Then give me a realistic estimate range: optimistic, most likely, and pessimistic. Tell me which number I should actually commit to with the people counting on me, and why.Why this prompt works
The reference class approach — asking "how long do projects like this actually take?" rather than "how long will this one take?" — is the most effective corrective for the planning fallacy. The prompt forces both a track record review and the outside view before accepting any inside-view estimate.
The psychology behind this
The Planning Fallacy
Read the full experiment →35_planning_fallacy.md