THE REVISION LEDGER
Separate a changed estimate from a changed market
Choose one observation year and one official row. The path moves across publication editions, not across economic time.
Five first-party editions paired
MCS 2022–2026 · Gold chapters plus official data records · local normalized rows · no live provider.
Every point has a direct value label. Move the pointer, tap or drag, or focus the chart and use Left, Right, Home and End.
SAME YEAR, EVERY ROW
2024 revision matrix
The delta compares the first edition containing 2024 with MCS 2026. Zero means the displayed value did not change.
| Official row | MCS 2025 | MCS 2026 | First → latest |
|---|---|---|---|
| 160 te | 163 t | +3 t | |
| 170 te | 180 t | +10 t | |
| 90 te | 89 t | −1 t | |
| 150 te | 190 t | +40 t | |
| 300 te | 289 t | −11 t | |
| 200 te | 210 t | +10 t | |
| 8,130 te | 8,130 t | 0 t |
Open five source identities, hashes and evidence limits
2017–2021
Same-edition Gold PDF · official data record
- Chapter
- 138,375 bytes
- DOIs
- 10.3133/mcs2022 · 10.5066/P9KKMCP4
- SHA-256
8c7b3cf8d69871080e6ce4140cf6b7d6cd9c5d31cf208bc258e0c83902945cc5- Version
- First posted January 31, 2022
2018–2022
Same-edition Gold PDF · official data record
- Chapter
- 180,109 bytes
- DOIs
- 10.3133/mcs2023 · 10.5066/P9WCYUI6
- SHA-256
f9274555eb8a985a04f578ddf6a76669bf1b6d813923443206f27c3f1c145b2b- Version
- First posted January 31, 2023
2019–2023
Same-edition Gold PDF · official data record
- Chapter
- 743,608 bytes
- DOIs
- 10.3133/mcs2024 · 10.5066/P144BA54
- SHA-256
3b551dcc4dbc7e21b3b1789c4552dab12d370e7a839288744196105e4767946c- Version
- Data release version 2.0, March 2024
2020–2024
Same-edition Gold PDF · official data record
- Chapter
- 743,291 bytes
- DOIs
- 10.3133/mcs2025 · 10.5066/P13XCP3R
- SHA-256
d4ec1750249005b4ad91dd59aaa7978932ecb7b7a410d8f6a135e6f8f8cc136f- Version
- Report version 1.2, March 2025
2021–2025
Same-edition Gold PDF · official data record
- Chapter
- 138,864 bytes
- DOIs
- 10.3133/mcs2026 · 10.5066/P1WKQ63T
- SHA-256
c2fef62f665d3334302b8b9bd32e2da6f40b3f8136d1ae00ffa102e196943627- Version
- Report version 1.3, May 2026
Interpretation boundary: an unmarked prior year is not labelled final or audited. Reported consumption is not apparent consumption, and Treasury stocks are not industry inventory. USGS does not provide a cause for every changed cell; this tool is not a supply forecast or price forecast.
THE DIRECT ANSWER
Publication versions can materially change an apparently precise statistic
For 2021 primary refinery production, MCS 2022 reported 180 metric tonnes. MCS 2023 printed e170 for that cell, and MCS 2024 onward reported 181 tonnes. The first-to-latest displayed change is +1 metric tonnes.
For 2024 mine production, MCS 2025 marked the observation as estimated at 160 tonnes. MCS 2026 reports 163 tonnes without an estimate mark. The value changed by +3 metric tonnes; that does not establish why it changed or imply a market shock.
WHAT THE LEDGER CAN PROVE
Edition identity is evidence, not a forecasting signal
Fact
The selected official row differs across named USGS editions, with direct chapter links, report and data DOIs, and chapter hashes shown.
Inference boundary
Later reporting, classification or rounding may matter, but USGS does not publish a complete cell-level revision-cause ledger.
Market boundary
Reported consumption and Treasury stocks have their own definitions. They do not close a supply balance, measure retail availability or predict gold prices.
CONTINUE THE EVIDENCE CHAIN
Use each page for a different question
QUESTIONS
USGS gold revision FAQ
What is a USGS gold data revision?
It is a change in the value published for the same observation year across later Mineral Commodity Summaries editions. It is not a change measured between two economic years.
Why can an estimate change in the next edition?
Later reporting, classification work or rounding can change a published cell, but USGS does not assign one cell-level cause to every revision. GoldObserve shows the difference without inventing a reason.
Why does the comparison stop at 2024?
The 2025 observation appears only in MCS 2026, so it has no later edition to compare yet. The tool requires at least two official editions for every displayed observation year.
Does an observation without an e mark mean final or audited?
No. It only means the same-edition Gold PDF did not mark that displayed value as estimated. Later editions can still revise it.
Is reported consumption a supply balance or price signal?
No. It is shown under the USGS published label and definition. GoldObserve does not turn it into apparent consumption, a closed supply balance or a price forecast.