Point in Time what the site said on any publish, never restated
The vintage store proves what inputs we had on a day. This ledger proves what we published. Every daily run appends its headline readings to an append-only file in the repository; nothing is ever edited. In a claim or a change order the counterparty cannot argue the history was revised — the row is right here, with the publish timestamp, and the git commit behind it.
Export data
Cite
MacroGauge Macrogauge (CPI-comparable) YoY as published, 2026-10-02, 2018-01=100, 3.2% vs official 3.4% — https://macrogauge.vercel.app/as-of| Reading as published 2026-10-02 | Value | As of |
|---|---|---|
| Macrogauge (CPI-comparable) | +3.2% | 2026-10-02 |
| CPI-Tracker | +3.7% | 2026-10-02 |
| Cost of Living | +5.8% | 2026-10-02 |
| Supercore | +2.7% | 2026-08-01 |
| PCE-weighted | +3.2% | 2026-10-02 |
| Official CPI | +3.4% | 2026-08-01 |
| DC Build | +8.4% | 2026-10-02 |
| DC Ops | +5.6% | 2026-09-01 |
| DC Hardware | +32.1% | 2026-10-02 |
The solid line is the headline exactly as each day's publish stated it — this ledger is append-only and never rewritten. The dashed line is what today's history says for the same dates. Where they differ is the honest revision footprint of a live-data gauge: a late print, a re-published source value, a splice moving. Nothing on this page is recomputed; a row is read back verbatim.
96 publishes on record since 2026-07-08. Rows before 2026-09-03 were backfilled from the git history of pulse.json (scripts/backfill_ledger.py) using the same row builder the live run uses; fields absent on early rows (DC index, Cost of Living) read as — because those artifacts did not exist yet. Comparison months come from compare.json; source: 2026-10-02 publish.