Start here: the v1 report uses the CMT24 relaxed-proxy calendar-tenor spread as the primary full-history series, then keeps strict longer-tenor spreads as diagnostics.
V1 primary calendar-tenor proxy:CMT12/CMT24_RELAXED_3M, built as CMT12_STRICT versus CMT24_RELAXED_3M. Primary status is ok with business-day coverage of 96.3%.
Label discipline: do not relabel this as clean strict CMT24. The full-history strict CMT24, CMT36, and CMT60 diagnostics remain sparse. The input CMT proxy quality gate is blocking with 13 blocking rows. Publishing remains disabled unless explicitly approved.
Anchor on CMT12, then read the heatmap by spread segment.
Data source: /Users/agentr/Claude/data-local/bbg/bql_exports/copper_curve_spreads/copper_curve_spreads_cmt_proxy_20260708_1044/local_output/processed/copper_cmt_proxy_regime_map_panel.csv. Chart starts from 2011-01-01.
Panels 1-3: CMT12 price, walk-forward regime classification, and COMEX listed-contract prices beside the CMT24 proxy and CMT calendar-tenor anchors.
Use these panels to audit the percent, absolute, and percentile inputs behind the heatmap.
Data source: /Users/agentr/Claude/data-local/bbg/bql_exports/copper_curve_spreads/copper_curve_spreads_cmt_proxy_20260708_1044/local_output/processed/copper_cmt_proxy_regime_map_panel.csv. Chart starts from 2011-01-01.
This v1 report keeps the old Copper Timespreads report structure but changes the primary full-history calendar-tenor series to a clearly labelled CMT24 relaxed proxy.
The report remains curve-only and clean-split: BQNT writes raw/check/reference CSVs, while the local builder owns quality gates, charting, report rendering, and local artifacts.
The primary regime map ranks log(CMT12_STRICT) - log(CMT24_RELAXED_3M) against its own expanding history. Low spread percentiles are treated as bullish because CMT12 is unusually cheap versus the two-year proxy anchor; high spread percentiles are treated as bearish or mature-cycle.
Primary v1 calendar-tenor spread:CMT12/CMT24_RELAXED_3M. It is the best full-history compromise because CMT12 is clean on the business-day grid and CMT24 becomes usable with a small, explicitly labelled +/-3 month relaxation.
Diagnostics: keep CMT12/CMT24_STRICT to audit exact 24m availability, CMT12/CMT36_STRICT and CMT12/CMT36_RELAXED_6M to show the 36m sparsity problem, and CMT12/CMT60_STRICT as a long-end diagnostic rather than the primary full-history signal.
Hard rule: if a relaxed/proxy series is shown, the label must include RELAXED or PROXY. Do not silently widen a tenor and call it strict CMT.
Primary:CMT12/CMT24_RELAXED_3M. The near leg is strict CMT12. The far leg is the CMT24 relaxed method with a +/-3 month tolerance. It is a full-history proxy and must stay labelled as relaxed/proxy.
Diagnostics:CMT12/CMT24_STRICT, CMT12/CMT36_STRICT, CMT12/CMT36_RELAXED_6M, and CMT12/CMT60_STRICT are retained to show what fails or improves under stricter definitions.
Grid semantics: quality tables distinguish calendar-row coverage, business-day coverage, and usable-grid coverage. Business-day coverage is the main gating metric for the primary spread.
Relaxed CMT methods are compared with strict CMT where strict rows exist. Nearest-contract fallback is shown separately and is not silently blended into strict CMT.
