# Compact operations HUD projection

Measured the main dashboard refresh: approximately9.75ms, including3.64ms spent obtaining a complete OperationsCoordinator snapshot for a short status label. Added get_dashboard_summary on the coordinator and get_operations_dashboard_report on Runtime. They return only live flags, action points, total arrears, terminal/loss/legacy status and the latest close day/profit. UI uses this compact API with old-report fallback for fixtures. Full reports and save snapshots are unchanged. Returned dictionaries/arrays are newly allocated and contain no mutable references to stored summaries.

Dashboard136 checks PASS. Same production fixture: operations text average3635us before /11.75us after; full x1 average34.410ms /30.374ms. x3 P95124.341ms /122.392ms, still above12ms. No claim of budget compliance or GPU/Windows performance.

Opening/save/replay test completed with58 checks,0 failures,exit0 (compact-operations-opening.log), including10 new projection-equivalence and defensive-copy checks. Evidence: compact-operations-ui.log, production-ui-costs.json, production-compact-operations.json.

## Pause/speed projection follow-up

The dashboard still fetched the full operations snapshot inside `_update_pause_and_speed`. Added `deferred_runtime_minute` to the compact API and switched this consumer to it, retaining fallback for older fixtures. No simulation or save contract changed.

Targeted dashboard139 checks PASS,exit0: includes injected deferred-state warning/retry labels and clear-to-resume checks. These validate projection/display, not a complete fault-recovery scenario. Production diagnostic (`production-compact-pause.json`) reports x1 average22.057ms, x3 P9575.145ms, dashboard refresh2.009ms. Compared with the preceding compact-text run30.374ms/122.392ms, the measured update is faster; both budgets remain failed. Runs are headless, small fixture, no GPU/Windows acceptance. Guide subprofile (`production-guide-costs.json`) measured refresh1.115ms, so it did not explain the whole previous dashboard cost.

Latest initial-day/save/replay test completed with59 checks,0 failures,exit0 (`compact-pause-opening.log`), including the added deferred-field equivalence check. Previous58-check pass applies to the preceding increment.