Overview
Three-Point Estimation, also called PERT (Program Evaluation and Review Technique) estimation, is a method for estimating task duration using three scenarios to account for uncertainty and provide more realistic project timelines.
The Three Estimates
Optimistic (O)
Definition: Best-case scenario time
Assumptions:
- Everything goes perfectly
- No obstacles or delays
- Ideal conditions
- Maximum efficiency
Question: "What's the fastest this could possibly be done?"
Most Likely (M)
Definition: Realistic scenario time
Assumptions:
- Normal working conditions
- Expected challenges accounted for
- Based on experience
- Most probable outcome
Question: "Given normal conditions, how long will this take?"
Pessimistic (P)
Definition: Worst-case scenario time
Assumptions:
- Significant obstacles encountered
- Multiple setbacks
- Unfavorable conditions
- Still completable (not catastrophic)
Question: "If things go badly, what's the maximum reasonable time?"
Calculation Methods
Triangular Distribution
Formula: (O + M + P) / 3
When to Use:
- Equal weight to all three estimates
- Simpler calculation
- Less historical data available
Example:
- Optimistic: 4 days
- Most Likely: 6 days
- Pessimistic: 10 days
- Estimate: (4 + 6 + 10) / 3 = 6.67 days
Formula: (O + 4M + P) / 6
When to Use:
- Weight most likely estimate more heavily
- More accurate with experienced estimators
- Industry standard for project management
Example:
- Optimistic: 4 days
- Most Likely: 6 days
- Pessimistic: 10 days
- Estimate: (4 + 4×6 + 10) / 6 = 6.33 days
Standard Deviation
Formula: (P - O) / 6
Purpose: Measures uncertainty
- Higher SD = more uncertainty
- Lower SD = more confidence
- Helps assess risk
Confidence Intervals:
- 68% confidence: Estimate ± 1 SD
- 95% confidence: Estimate ± 2 SD
- 99.7% confidence: Estimate ± 3 SD
Benefits
More Realistic Estimates
- Accounts for uncertainty
- Balances optimism and pessimism
- Based on range rather than single point
- Reflects real-world variability
Risk Assessment
- Standard deviation shows uncertainty level
- Identifies high-risk tasks
- Supports contingency planning
- Improves decision-making
Better Than Single-Point
- Single estimates tend toward optimism
- Doesn't account for unknowns
- Three-point captures range of outcomes
- More defendable to stakeholders
How to Apply
Step 1: Break Down Work
- Divide project into tasks
- Make tasks estimable (not too large)
- Identify dependencies
- Involve people who'll do the work
- Discuss assumptions
- Consider historical data
- Account for known risks
Step 3: Create Three Estimates
For each task:
- Optimistic: What if everything goes right?
- Most Likely: What's our realistic estimate?
- Pessimistic: What if we hit obstacles?
Step 4: Calculate
- Apply formula (usually PERT: (O+4M+P)/6)
- Calculate standard deviation
- Sum estimates for total project time
- Add buffer based on uncertainty
Step 5: Review and Refine
- Sanity check results
- Compare to historical projects
- Adjust for unique factors
- Get team buy-in
Common Mistakes
Too Optimistic Across Board
- All three estimates are optimistic
- Doesn't truly represent worst case
- Solution: Challenge assumptions
Pessimistic = Catastrophic
- Worst case includes disasters
- Not realistic for planning
- Solution: Pessimistic should be bad but plausible
Anchor Bias
- First estimate influences others
- All three cluster too closely
- Solution: Estimate independently or use ranges
Forgetting Dependencies
- Estimates assume parallel work
- Ignore sequential constraints
- Solution: Consider critical path
Use in Agile
Story Point Estimation
Can use three-point for complexity:
- Optimistic: 3 points
- Most Likely: 5 points
- Pessimistic: 8 points
- Weighted estimate: ~5 points
Sprint Planning
- Estimate capacity with three scenarios
- Plan for most likely
- Understand risk with pessimistic
Release Planning
- Project completion dates
- Account for velocity variance
- Communicate uncertainty to stakeholders
Planning Poker Variant
- Team estimates optimistic
- Team estimates most likely
- Team estimates pessimistic
- Calculate using PERT formula
Monte Carlo Simulation
- Run thousands of simulations
- Use three-point estimates as inputs
- Generate probability distributions
- More sophisticated analysis
When to Use
Good For:
- Complex, uncertain projects
- High-stakes deliverables
- Projects with significant unknowns
- First-time initiatives
- Communicating risk to stakeholders
Less Useful For:
- Routine, well-understood tasks
- Very short tasks (overhead not worth it)
- When single estimate is accurate enough
Ideal For
- Project managers
- Agile teams doing estimation
- Anyone planning complex work
- Organizations needing realistic timelines
- Teams learning to estimate better
