Overcome the planning fallacy by getting calibrated time estimates with built-in multipliers for complexity, interruptions, and the unexpected
You are a time estimation calibration coach who helps people overcome the planning fallacy, the universal tendency to underestimate how long tasks actually take. Your approach combines reference class forecasting, segmented estimation, and pre-mortem thinking to produce realistic time predictions. The task I need to estimate is [TASK_DESCRIPTION]. My initial gut feeling is that this will take approximately [INITIAL_ESTIMATE]. This task falls into the category of [TASK_TYPE:select:creative work like writing or design,technical work like coding or analysis,administrative work like emails and scheduling,learning something new,physical project like organizing or building,communication intensive like meetings and coordination,research and information gathering,mixed or undefined]. My familiarity with this type of work is [FAMILIARITY:select:I do this regularly and have done similar tasks many times,I have done something like this a few times before,I have done this once or twice,this is mostly new territory for me]. When I reflect on my past estimation accuracy, I typically [ESTIMATION_PATTERN:select:significantly underestimate by 50 percent or more,somewhat underestimate by 25 to 50 percent,slightly underestimate by 10 to 25 percent,estimate fairly accurately,tend to overestimate to be safe]. A similar task I have done before is [REFERENCE_TASK?]. If you remember, share what you estimated versus how long it actually took, as this reference class data dramatically improves future predictions. Additional context that might affect timing includes [CONTEXT?]. This could include dependencies on other people, potential blockers, competing priorities, or environmental factors like interruptions. Walk me through a calibrated estimation process. Start by breaking down my task into its component subtasks, because aggregated estimates are more accurate than single estimates for complex work. For each subtask, provide a time estimate and note any hidden steps I might be forgetting. Then apply appropriate adjustment factors. Consider my historical estimation pattern and apply a correction multiplier. Account for task novelty since unfamiliar work typically takes 2 to 3 times longer than expected. Factor in the coordination tax if this involves waiting on others or scheduling. Add buffer for interruptions, context switching, and the inevitable unexpected issues. Also apply the outside view by asking: if someone else with my skill level were doing this task, how long would you expect it to take them? This third-person perspective reduces optimism bias because we are more realistic about others than ourselves. Present your analysis with three scenarios. The optimistic estimate assumes everything goes smoothly with no interruptions or surprises. The realistic estimate accounts for normal friction, minor setbacks, and typical workday interruptions. The pessimistic estimate considers significant blockers, scope creep, or learning curves that reveal hidden complexity. For each scenario, show your math so I can understand the reasoning. The realistic estimate should be the one I actually plan around, not the optimistic one. Finally, give me a brief pre-mortem. Imagine we are looking back at this task after it took twice as long as my initial estimate. What are the three most likely reasons that happened? This helps me anticipate and prevent common pitfalls. Be direct about where my initial estimate seems off. I want calibration, not validation.
Use this prompt anywhere
10,000+ expert prompts for ChatGPT, Claude, Gemini, and wherever you use AI.
Get Early AccessAsk someone how long a task will take and they'll picture the best version of doing it, no interruptions, no re-work, no waiting on a reviewer. That's the planning fallacy, and it's why almost every estimate comes in low. The fix isn't trying harder to guess right, it's building in the overhead you already know is coming.
Describe the work in [TASK_DESCRIPTION] and give your gut number as [INITIAL_ESTIMATE]. The coach runs it through bottom-up breakdown and three-point analysis (best case, likely case, worst case) to produce a calibrated range instead of one number you'll blow past. Add a [REFERENCE_TASK] if something similar shipped before. Real history beats a guess every time.
Pair this with the deadline tracker to see how your calibrated estimate holds up against the real delivery date, or the sprint planning assistant to turn individual estimates into a realistic sprint commitment.
Calibrate your next estimate in the Dock Editor before you promise a delivery date you can't hit.
Open ChatGPT, Claude, Gemini, or the Dock Editor and paste the full Time Estimator Coach prompt to get a structured analysis of your time estimate for any task or project.
Enter what you need to estimate and your gut-feel estimate. The AI uses your initial number as a starting point and stress-tests it rather than replacing it entirely.
Specify the type of work (development, design, research, writing, planning) and how familiar you are with similar tasks. These factors directly affect how much padding the AI recommends.
Describe whether you tend to underestimate or overestimate, and provide a similar past task with its actual completion time. This calibrates the coaching to your personal estimation bias.
The AI produces an adjusted estimate with confidence ranges (best case, likely, worst case), a breakdown of hidden time costs you may have missed, and specific questions to ask before committing to a deadline.
Get calibrated time estimates for development tasks by accounting for code review, testing, deployment, and the inevitable scope creep that turns a 2-hour task into a full day.
Generate three-point estimates (optimistic, likely, pessimistic) for project milestones that give stakeholders honest expectations instead of overly optimistic deadlines.
Produce realistic project estimates that include hidden time costs like client communication, revision rounds, and administrative overhead so you stop underpricing your work.
Build awareness of your estimation bias by comparing your initial estimates against calibrated projections. Over time, this feedback loop improves your natural estimation accuracy.
The planning fallacy is the main culprit. People focus on the ideal scenario rather than accounting for realistic obstacles. Other factors include forgetting overhead tasks (meetings, communication, context switching), anchoring to the first number that comes to mind, and social pressure to give optimistic estimates.
Five widely used methods cover most scenarios. Bottom-up estimation (break work into small pieces), top-down (estimate the whole then allocate), three-point analysis (optimistic, most likely, pessimistic), analogous estimation (base on similar past projects), and expert judgment. This prompt combines elements of all five using your [TASK_DESCRIPTION] and [INITIAL_ESTIMATE].
Break the unknown task into smaller pieces where at least some parts are familiar. Research how long similar tasks took other people or teams. Add a larger buffer (50 to 100 percent for truly novel work versus 20 to 30 percent for familiar work). Set intermediate checkpoints to validate your [INITIAL_ESTIMATE] early.
PERT uses three estimates: Optimistic (O), Most Likely (M), and Pessimistic (P). The weighted formula is (O + 4M + P) / 6, giving more weight to the most likely scenario. Standard deviation is (P - O) / 6. PERT is useful because it produces both an expected duration and a measure of uncertainty around your [INITIAL_ESTIMATE].
Present a range rather than a single number. Use your [INITIAL_ESTIMATE] as the baseline, then share best-case, expected, and worst-case timelines. Explain the assumptions behind each scenario. Stakeholders respect honest ranges more than precise numbers that consistently miss. Update estimates as you learn more about the [TASK_DESCRIPTION].
20 to 30 percent for familiar work, 50 to 100 percent for novel tasks. The buffer covers unknowns that your [INITIAL_ESTIMATE] cannot predict: unexpected dependencies, scope clarifications, and integration issues. Track your actual-versus-estimated ratios over time to calibrate your personal buffer based on real data rather than rules of thumb.
Discover more prompts that could help with your workflow.
Analyze overcrowded calendars and receive a strategic reorganization plan that protects focus time, batches meetings efficiently, and reclaims hours for meaningful work
Analyze how you spend your time, identify hidden time wasters, and get actionable strategies to reclaim lost hours
Create a personalized Getting Things Done productivity system tailored to your work, tools, and lifestyle.
10,000+ expert-curated prompts for ChatGPT, Claude, Gemini, and wherever you use AI. Our extension helps any prompt deliver better results.