Skip to content

Velocity panel

This is for Scrum Masters, Product Owners, and PMs who want to know how much work a team can realistically get done. Velocity is how many story points a team completes per sprint on average — it’s the basis for forecasting how many sprints are left to finish the backlog. This panel shows a bar chart of the team’s last 8 closed sprints, color-coded by how much of what they committed to they actually finished, along with a rolling average (with its statistical spread) and a forecast range for the next sprint.

Step 7 (Forecast — Monte Carlo across both worlds) of the hybrid PM flow — the velocity history that turns sprint cadence into a defensible probability for Janet’s exec view.

  • Primary stat: {avg} ± {stdev} pts (last N) using .tppm-mono
  • Forecast chip: Forecast {low}–{high} pts right-aligned (range is avg ± 1 stdev, rounded to int)
  • Bar chart: 8 bars (or as many closed sprints as exist), colored by completion ratio:
    • ≥ 0.85 → semantic-on-track
    • 0.6–0.85 → semantic-at-risk
    • < 0.6 → semantic-critical
    • no committed points (none recorded, or zero) → neutral bar, labeled “no committed points”; never red, because there is no commitment to fall short of
  • Footer: the plain text Velocity feeds CPM duration estimates (a link to ADR-0036 before 0.5)

The ? button beside the panel heading explains what the average counts: closed sprints only, excluded sprints held out, and a forecast range that is the average plus or minus one standard deviation.

The active-sprint panel at the top of the Board carries a compact version of the same data:

  • History chart with a range band: the 8-sprint bars sit behind a shaded min–max band with a dashed P50 (median) line, so the typical throughput reads at a glance.
  • Delivery forecast line: a one-line answer to “when does it ship?” —
    • if the active sprint is bound to a milestone, the milestone’s reforecast delivery estimate (from the sprint-close reforecast bridge, ADR-0106) — labeled a velocity-based estimate, not a simulation, because the bridge reforecast is a deterministic velocity-band heuristic; true probabilistic P50/P80 lands once agile-aware Monte Carlo backs it;
    • otherwise the remaining committed backlog re-paced into “~N–M more sprints to clear X pts (by ~date)”;
    • falling back to “Need at least 3 closed sprints to forecast delivery” until there is enough history.
  • Team-private by default: velocity is gated by the team-signal privacy ladder (ADR-0104). When the reader’s tier is below the velocity audience, the card shows a “Velocity is team-private” state and the forecast is not fetched — never a misleading empty chart.
  • Route: /projects/:projectId/sprints (right column of the metrics row, bottom half)
  • Board: top sprint panel → Velocity card (history band + forecast line)

A backlog delivery forecast was added in 0.3. It answers “when is the whole backlog done?” without spreadsheet math, from a velocity Monte Carlo: it bootstrap-samples the team’s historical per-sprint throughput a thousand times to clear the remaining committed backlog, and reports P50 / P80 sprint counts and calendar dates.

Because this is a real simulation — not the deterministic velocity band the milestone reforecast uses — the P50/P80 percentile vocabulary is honest here and is shown as such. It surfaces in two places, both reading the same /sprint-forecast/ endpoint (no duplicate computation):

  • Project overview — a Backlog forecast card: “At current velocity, the remaining backlog is forecast to clear by [P50 date]”, with the P80 date and sprint counts below.
  • Board sprint panel — two compact projection chips beneath the burndown: a sprint-finish chip (“on track to finish ahead” / “N pts behind at this pace”, from the current sprint’s burn pace) and a release-horizon chip (“at this pace, the backlog clears in ~N sprints”). Both link to the overview forecast.

Like every velocity surface it is team-private by default (ADR-0104): a reader below the velocity audience sees a “team-private” state, never the horizon, since the backlog horizon is reversible into the team’s throughput. A forecast appears only once two sprints have closed — one observation cannot express the velocity variance the forecast is built on.

MethodEndpointPurpose
GET/api/v1/projects/{id}/velocity/Last-8 closed sprint points/tasks (each with exclude_from_velocity) + rolling stats + forecast range + team_velocity_per_day + excluded_count
GET/api/v1/projects/{id}/sprint-forecast/Backlog delivery forecast — P50/P80 sprint counts + calendar dates from a velocity Monte Carlo (computed on read, cached; velocity_suppressed for below-audience readers)

The team_velocity_per_day field is the same rolling per-day average used by velocity calibration suggestions on sprint close. The response also carries an excluded_count and a per-sprint exclude_from_velocity flag (see Setup work & Sprint 0).

A setup or ramp-up iteration — what agile teams often call a “Sprint 0” — is the period before real delivery begins: standing up the environment, forming the team, building the initial backlog, spiking the architecture. Its throughput is low or zero, so if it is run as a sprint and closed, it drags the rolling average down and widens the forecast band until it ages out of the window — and that contamination flows into the milestone delivery forecast too.

TruePPM is scheduling-first, so it has an opinion about where this work belongs:

  • Model mobilization as schedule tasks, not a sprint. Setup work is real work with durations and dependencies — put it on the schedule (the waterfall side of the hybrid bridge), where it shows up in the critical path and the Gantt rather than as a misleading zero-velocity sprint. This keeps your velocity baseline clean from the start.
  • If you do run a ramp-up sprint, exclude it from velocity. A Scheduler (or above) can flag any sprint Exclude from velocity from the sprint workspace’s Sprint options (⋯) menu in the page header; an excluded sprint also shows a standing Excluded from velocity chip beside it. An excluded sprint is held out of the rolling average, the forecast band, and the milestone reforecast — but it stays fully visible in your history.
  • In the velocity chart the bar is muted and hatched (and hollow in the Board sparkline), labeled excl, so it is marked rather than silently dropped — you can always see what was excluded and why. The Board sparkline also carries a small N excluded pill so the reduced-set average is signaled at a glance.
  • The panel shows an ⌀ N excluded chip and the rolling-average label counts only the sprints that still feed velocity.
  • The flag is team-owned: it is set by the team’s Scheduler-and-above, with no PMO override, and every change is recorded in the sprint’s history audit trail. It can be set after a sprint closes — teams usually realize a sprint skewed their numbers only in hindsight.

The exclusion is honored everywhere velocity is consumed: the panel, the Board sparkline, the delivery forecast, and the CPM velocity-calibration source — a single rule, applied once.

Stakeholders trust a range more than a single number. A forecast that says “5–8 sprints remaining” defends itself against the team’s natural variance; a point estimate is brittle.

The terminology is forecast range, not “velocity confidence band” — confidence band has a specific statistical meaning we are not claiming here. Velocity is XP-origin (not part of core Scrum) but real-world standard practice; we surface it as a practice-layer tool, not a process mandate.

  • ADR-0036 — Hybrid PM philosophy: how velocity feeds CPM duration estimates
  • ADR-0037 §Q3 — Velocity storage decision (snapshot on close, not computed from history)
  • Janet — when the team’s signal audience includes your tier, read the forecast range chip; otherwise velocity stays team-private (ADR-0104) and you work from the aggregate milestone-health signals. The footer ADR link explains why the range is defensible.
  • Sarah — the rolling avg drives the work package re-forecast on the Schedule view at Step 6.
  • Alex — own the trajectory. If the bar color is shifting amber over multiple sprints, the team is signaling something.