The only delivery forecast willing to be graded.
Every other tool sells you confidence. KalMatrix shows you its track record — it back-tests every forecast against your own closed sprints and publishes the hit rate, right next to the prediction. A forecast you can’t check is just a guess in a nicer font.
Were we right? Here’s the scorecard.
Every closed sprint is re-forecast from only the history available before it, then checked against what actually happened. These are the live demo workspace’s numbers, read from the same endpoint the product reads — your own appear on your data as sprints close.
Last published figures — the live scorecard is unreachable right now · reconstructed by re-running the forecaster over closed sprints — calibration, not early warning. Open the demo to read it from the product itself.
Every high-confidence call is scored against what actually shipped — including the ones we got wrong.
A green is graded the same way a red is — so it means something.
It back-tests on your own history.
- Replay, don’t peek. For each closed sprint, it fits the S-curve from only the sprints before it — no hindsight.
- Forecast mid-flight. It reconstructs where the sprint stood at mid-sprint and makes the call it would have made live.
- Check reality. It compares that call to what actually shipped — code-verified, not just what the board claimed.
- Publish the score. Hit rate, Brier score, and every prediction-vs-outcome row, shown next to today’s forecasts.
A measurement tool’s identity is the dashboard. The moment it publishes a falsifiable prediction and grades it, it can be wrong out loud — and most won’t take that risk. KalMatrix is built the opposite way: the willingness to be graded is the product. It’s also why a skeptical CTO will act on the number two days early instead of confirming the slip two days late.
One track record, three jobs it earns trust for.
Put a number in front of the board you can defend — from a tool that shows how often it’s been right, and that stores zero per-person data.
Know a sprint is slipping while it’s still saveable — and get the exact scope to cut, backed by a forecast with receipts.
Bring correction factors to planning — calibrated on your own history, not a folklore buffer — and a scorecard the steering committee can check.
Don’t take our word for it. Read the scorecard.
Open the live demo and see the accuracy page — real back-tested predictions, graded — in the next 30 seconds.