CAISO forecast evidence, without hiding the live gap.
California remains one of the hardest demand forecasting markets because net load depends on weather, rooftop solar, storage behavior, and the evening ramp.
This page now shows current live evidence first. Historical low-MAPE backtests remain visible as research archive, but they are not current production claims.
What is still useful about Gramm in CAISO.
BTM attribution and duck-curve modeling remain the right research direction, but they need to be proven against live h24 ISO references before production claims change.
Customer benchmarks should focus on paired CAISO OASIS forecast and actual rows, with TAC or custom territory scope declared before scoring.
The immediate product value is an auditable benchmark and API integration path, not a claim that current production already beats CAISO.
Live scorecard and archive.
Last updated May 22, 2026. A new CAISO candidate must beat the live row in shadow mode before it can be promoted.
Evaluation path
CAISO evaluation path.
A market-specific page should lead to a market-specific benchmark: confirm the live row, scope the period and horizon, then test payloads only after the evidence is clear.
- 01
Review live scorecard
Start with the current CAISO h24 row and its ISO-reference status before using archive results.
Open scorecard - 02
Request scoped benchmark
Use CAISO as the declared market, then add horizon, audit period, and decision context.
Request CAISO benchmark - 03
Test API payloads
Confirm response shape, region code, horizon, and validation status before wiring production workflows.
View API docs - 04
Move to procurement or paid plan
Use trust, SLA, pricing, and enterprise materials once the scoped benchmark supports review.
Open procurement