For any task, write three numbers: the best case, the most likely case, and the worst case. Use them to get one expected number and to see how risky the task is. It takes one extra minute and it stops the "you said two days" conversation.
One number is a trap
"How long?" "Two days." Now two days is a promise. If it takes three, you were wrong. If it takes one, you were padding. One number cannot be honest, because you do not actually know. You know a range.
Three numbers say what you really know.
The three numbers
- Best case. If everything goes right. No interruptions, no surprises.
- Likely case. A normal day. Some bumps.
- Worst case. The thing you are quietly worried about happens.
For a report: best case 1 day, likely 2, worst 4 if the data needs cleaning.
The simple math
To get one expected number from three, use this: (best + 4 × likely + worst) ÷ 6.
For the report: (1 + 4 × 2 + 4) ÷ 6 = 13 ÷ 6 = about 2.2 days.
Why count "likely" four times? Because it is the most probable. The best and worst cases pull the number a little, but not much. This formula has been used in project planning for decades. The three-point estimate calculator does the math and shows the spread.
The spread tells you the risk
Look at the gap between best and worst.
- Best 1, worst 1.5: you know this task well. Plan on it.
- Best 1, worst 4: there is a real unknown. Find out what it is before you promise anything.
A wide spread is not a bad estimate. It is an honest one. It is also a signal: do a small piece first, or get the unknown answered, then re-estimate.
How to say it
To your manager: "Likely two days. Could be one if the data is clean. Could be four if it is not. I will know by tomorrow lunch."
To a client: "We are planning on two days. If the data needs cleaning, it could be up to four, and we would tell you by tomorrow."
Both are more trustworthy than "two days," and both set up a cheap check-in instead of an expensive surprise.
Example
Priya is asked how long the FAQ page will take.
| Days | Why | |
|---|---|---|
| Best | 2 | Screenshots arrive, content is approved first time. |
| Likely | 3 | One round of edits. |
| Worst | 6 | Screenshots are late and content needs two rounds. |
Expected: (2 + 12 + 6) ÷ 6 = 3.3 days. Spread: 4 days. The risk is the screenshots. Priya asks for them today instead of on day three. The worst case shrinks to 4.
Where M-Tracker fits
In M-Tracker, tasks carry an estimate and the timer records the actual time. You and your manager can see estimate against actual across tasks, people, and projects. The pattern is visible, so estimates improve, and so do the plans built on them.
Make it a habit
Add three boxes to your task template: best, likely, worst. It takes a minute. After a month, compare your likely numbers with what actually happened. Most people find they are low by a steady amount. Once you know yours, your estimates get better without any other change.
Try M-Tracker free14-day trial. No credit card. No employee surveillance.
Work you can trust, without redoing it.
Your team gets feedback before work reaches you, or your client.
Copyright © 2020 – 2026 M-Tracker. All Rights Reserved. Privacy Policy
Trackbacks/Pingbacks