Predictive ETA
P50, P85, and P95 completion date forecasts powered by Monte Carlo simulation — with a what-if velocity slider.
Given how your team is actually delivering work, when will this project finish? Predictive ETA answers that by running a Monte Carlo simulation over your historical velocity and giving you three dates: likely (P50), confident (P85), and safe (P95).
Screenshot needed — add an annotated image showing this UI
What P50 / P85 / P95 Mean
| Percentile | Interpretation |
|---|---|
| P50 | There's a 50/50 chance the project finishes by this date. The "expected" date. |
| P85 | 85% of simulations finish by this date. A solid commitment date. |
| P95 | 95% of simulations finish by this date. Your safety-margin date. |
If P50 is June 10 and P95 is July 2, that's saying "we'll probably wrap early-to-mid June, but give us until the first week of July to be safe."
Where to See It
Predictive ETA appears on:
- Project overview — forecast card with all three dates and a confidence band
- Epic detail — for epic-level forecasts
- Roadmap — confidence bands on each epic bar
How the Forecast Is Built
Workestra looks at:
- How many tasks remain open
- The total remaining estimate
- Your team's completed estimate per week over the last 12 weeks
- Historical cycle predictability (see Reports)
It then runs thousands of simulated project timelines, each sampling from your actual delivery data. The distribution of end dates is what you see on the chart.
Forecasts need at least a few weeks of completion history to be meaningful. Brand-new projects will show "Not enough data yet" until velocity stabilizes.
What-If Analysis
Every forecast card includes a velocity multiplier slider:
- 0.5× — "what if we only had half our usual capacity?"
- 1.0× — baseline (your current velocity)
- 1.5× — "what if we added headcount and sped up 50%?"
- 2.0× — "what if we doubled capacity?"
Sliding the multiplier re-runs the simulation live. Use it for budget conversations, staffing decisions, or to show stakeholders why adding one more engineer buys you X weeks.
Forecast Caveats
Predictions are only as good as your data. The forecast assumes:
- Historical velocity is a reasonable predictor of future velocity
- Remaining scope is captured as tasks with estimates
- The team isn't about to change size or composition dramatically
When those assumptions don't hold, treat the forecast as directional — not a contract.
Next Steps
- Reports — velocity and predictability data behind the forecast
- Cycles — cycle completion feeds the model
- Capacity Planning — forward-looking workload