California benchmark memo: day-ahead error drops from 4.8% to 2.9%.

This memo summarizes the clearest public example of how Gramm compares with an incumbent baseline. It is meant to help a desk or operating team decide whether the same readout is worth requesting for its own market.

Benchmark summary

Market
California (CAISO)

Public day-ahead demand benchmark.

Metric
Day-ahead MAPE

Lower values mean lower forecast error.

Observed change
4.8% → 2.9%

39.6% lower error versus the public baseline.

Why it matters
Tomorrow's load view

Useful for bids, staffing, dispatch, and storage posture.

Why California is a useful starting market

The California day-ahead market is easy to understand and operationally important. If the forecast improves here, buyers can quickly judge whether it is worth testing against their own workflow.

How the comparison is measured

This comparison uses day-ahead MAPE, with the public baseline shown next to Gramm. The point is not to compress results into a blended claim. It is to show the gap in a way a trader, operator, or engineer can verify.

Why the gap matters in practice

Lower day-ahead load error affects bids, hedges, staffing, generation schedules, and storage posture. The number only matters if the same forecast can then move into live use on the same delivery surface.

What this memo should establish.

The baseline is visible.
The forecast run is clear.
The metric is familiar.
The same delivery surface is available in production.

If this is close to your workflow, request the same memo for your market.

The public benchmark is the starting point. The next step is a market-specific package with the dates, issuance timing, and horizon detail relevant to your desk or operating team.