Try it.
Check it. Challenge it.
GERT invites independent review of its historical load analysis. The most useful contribution is a reproduced result, a mistaken assumption you identify, or a concrete improvement to the way evidence is explained.
Start with a five-minute counterexample.
Not ready for a full technical review? Make two invented worlds produce identical meter readings and inspect why their unserved energy differs. Includes a short facilitator guide; no account or installation needed.
Open Same meter, two worlds →An introductory teaching activity, not a completed review of the historical dataset.
One complete review, about 10 minutes.
- Open the summer case at 80,000 MW and 0% reduction. Note the hours above assumed capacity and the gap energy.
- Switch to 5% reduction. Explain why the hours and energy change. Can you distinguish the actual observations from the two assumptions?
- Download the JSON report and hourly CSV. Recompute one row as max(0, adjusted load − capacity). Sum the hourly gaps to check the total in MWh.
- Inspect the winter case. Why would measured load during outages be insufficient to reconstruct unmet demand?
- Share one result you checked and one change you would make. A negative finding is useful.
Reproduce without the website
The archive extractor and independent report checker use Python's standard library. The source files, data version, checksums, formulas and limits are documented in the repository.
Reproduction instructionsUse it in a small workshop
A ready-to-use 20-minute exercise covers load, power versus energy, counterfactual assumptions, and the difference between arithmetic and a reliable operational decision.
Open the facilitator guideLeave an inspectable review.
Use the structured GitHub form for a reproducible review, or add one concise field note to the public discussion. You choose whether to submit. GitHub posts are public: do not include personal contact information, private datasets, credentials, or another person's words without permission.
See the community evidence ledger →A review records what you actually completed. Feedback is not automatically an endorsement; hosting a session, contributing a critique, and adopting software are different outcomes.
What this release establishes
Three public-data cases, reproducible deterministic calculations, downloadable reports, and an open review process. User-defined capacity is not historical capacity; counterfactual gap energy is not expected unserved energy. The probabilistic research candidate remains gated until its validation requirements pass.