Skip to contentComing soon: SprintProof for ChatGPT and Codex

Software estimation tool

Software estimation that tells you how sure you are.

Estimate each part of a feature as a range. The tool adds the ranges up over 10,000 simulated runs and gives you a date at the confidence you choose, so you see which parts carry the risk before anyone commits.

Free, no sign-up. Nothing you type is saved.

Three-point estimator
10,000 runs

Parts, in working days

best · likely · worst

Why ranges

A single number hides the risk.

“Two weeks” sounds like an estimate. It’s really three problems the team hasn’t had time to solve yet.

01

It sizes the whole feature at once

Nobody can hold twelve parts in their head. The number ends up being a feeling about the feature, not a sum of its work.

02

It doesn’t say how sure it is

“Two weeks” might mean “probably” or “if nothing goes wrong”. Whoever hears it can’t tell which, and plans as if it were a promise.

03

The missing work isn’t in it

Seat limits, bounced emails, users who already have an account. The parts nobody listed turn up mid-sprint, and the date slips.

How it works

How SprintProof estimates software.

  1. 01Break it into partsThe AI planner drafts the frontend, backend, data and rollout work. You edit, merge and split the parts.
  2. 02Three numbers per partBest case, likely and worst case, in dev-days. Small parts get narrow ranges; risky ones get honest ones.
  3. 03Add what the plan missed“Did you think of…?” questions from the check library become parts with their own ranges, or written decisions.
  4. 04Simulate, then pick a confidence10,000 runs of the plan give a date at 50% for planning, 80% for a commitment, or 90% for a launch. Parts tend to overrun together, so each run shares part of its overrun across the parts, and dates assume engineers work in parallel about 70% of the time.

Methods

Software estimation methods, compared.

Each method answers a different question. Only ranges answer the one your stakeholders ask: when will it ship, and how sure are you?

Estimation methods compared
MethodShows uncertaintyGives a dateFinds missing work
One number for the featureNoYes, with unknown oddsNo
Story pointsPartly, through the gaps in the scaleOnly through average velocityNo
T-shirt sizesPartlyNoNo
SprintProof rangesYes, a range for every partYes, at the confidence you chooseYes, from the check library

FAQ

Software estimation questions

What is software estimation?
Predicting how much work a piece of software will take, usually in days or story points, so a team can plan and commit to a date. A useful estimate says how sure it is, not just how big the work is.
How does three-point estimation work?
You give each part a best case, a likely case and a worst case. SprintProof weights them the PERT way, (best + 4 × likely + worst) ÷ 6, then runs the whole plan 10,000 times to see how often it finishes by each date.
Can we keep using story points?
Yes. Keep points on your board for sizing, and use ranges when someone asks for a date. See how story points and days fit together in the story points guide.
Does it work with Jira and Linear?
In SprintProof, tasks go to Jira or Linear with acceptance criteria, estimates and dependencies.
Is the estimator on this page free?
Yes, with no sign-up. Nothing you type here is saved. SprintProof uses the same simulation for the full plan, with the AI planner and the check library on top.

Team

Your next estimate deserves error bars.

Create your account and name your workspace. Your first feature plan is free. To keep planning, subscribe to Team: the 14-day money-back guarantee applies.

Want to hear when the ChatGPT and Codex plugin ships? Leave your email.

Team size (optional)
One email when it ships. No spam.