Background
This is level 1 of 5. The records for this level are in workspace/t1/.
Task
Subject: PJM day-ahead: what is the bar?
Hi,
We are going ahead with scoping PJM for the day-ahead product. Before anyone quotes a number to a prospect I want our own view of the bar we have to clear: how well PJM's own day-ahead forecast, the one it files with EIA every morning, has tracked the demand series our products are scored on over the last four full years, 2022 to 2025. Whatever we end up saying has to be something a prospect can reproduce from public data, so work from the EIA files we hold.
One prospect forwarded the open-source surge benchmark (the checkout is in github/surge). It puts PJM's operator error for 2025 at 3,297 MW with surge beating it by half, and Dev would like that line on the PJM one-pager; his mail is in mail/. Tell me whether it stands up.
Please put in workspace/t1/results/:
- a workbook with PJM's day-ahead accuracy year by year and through the hours of the day, with a sheet saying how you scored it;
- a short note, plain text is fine, that I can forward to Jonah: the figure we should use as PJM's bar for 2025, and what it does to the surge comparison.
Our evaluation standards and the draft PJM product spec are under forecasting/.
Maren
Input
workspace/t1/: the records supplied with the work (company/,eia861/,eia930/,forecasting/,github/,mail/,meetings/,projects/,reading/,weather/).
Output
workspace/t1/results/pjm_da_benchmark.xlsxworkspace/t1/results/note_for_jonah.txt