The GERT Historical Lab is a deterministic, browser-based teaching exercise built on 216 hourly observations of ERCOT public native-load data, in three deliberately selected 72-hour windows: August 9-11, 2023; April 10-12, 2023; and February 14-16, 2021. The windows illustrate different load profiles and are not a representative validation sample or a forecast backtest. Learners open https://gert-d.vercel.app/history, set a capacity (MW) and a uniform reduction (%) as explicit assumptions, and watch the lab compute gap power and gap energy hour by hour. They then download the JSON report and run an independent checker — Python standard library only, no packages — that recomputes every hourly value against the bundled source data and fails on any checksum mismatch, changed observation, invalid assumption, or wrong result: python3 scripts/verify_historical_report.py --report <downloaded-json>. The checker's Python code is independent of the browser's TypeScript. A verified reproduction, a failed check, or a sharp criticism all count as useful results; results are filed with the repo's "Historical Lab independent review" issue template. Honest framing, stated up front in the materials: this is a deterministic teaching exercise, not a validated forecast. Capacity is a user assumption, not historical capacity; the reduction is a counterfactual, not a measured intervention. Separately, the project's probabilistic research candidate v1.4 was publicly rejected for production inference after missing its predeclared calibration gate. Nothing here should be described as a validated prediction system. Who can use it: a high school or university STEM/energy/data club, educator, or study group. Browser-only, zero install. Status: pilot teaching exercise, not a completed or evaluated educational program.
