We rebuild the cash-flow model behind each gold and copper study and show where our results match the published numbers and where they differ. Subscribers can compare projects on the same prices and discount rate, and test their own cases in the models.
Open the free screener Browse the library| Project | Country | Primary metal | Study | After-tax NPV, as published | Rebuilt model |
|---|---|---|---|---|---|
| Troilus | Canada | Gold | FS | US$3,206M | |
| Black Pine | USA | Gold | FS | US$2,397M | |
| Green Bay | Canada | Copper | PEA | C$2,135M |
A complete sample model is free to view in your browser: open the sample.
When a mining company publishes an economic study, its headline NPV is an output you are asked to trust. We rebuild the entire cash-flow model from the report's own published inputs — every assumption footnoted to its page and table, every tax line built from statute, verified by real recalculation — and show you, in plain language, what drives the value and how the rebuilt figures compare with the report.
The same discipline on every report, whatever the commodity or code.
A formula-driven Excel model reconstructs the study's economics from its lowest-level published inputs — unit costs × physical drivers, never the report's own totals pasted in. The report's figures become reconciliation targets the model must independently hit.
Every model must reproduce the report's headline NPV within a disclosed tolerance under real spreadsheet recalculation. Where a report cannot be reconciled to its own tables, we publish the divergence and its decomposition rather than force a tie.
Fiscal lines are built from the governing law — rates, depreciation regimes, loss carry-forwards, royalties — and checked against the report's published tax. Where the report and the statute disagree, both treatments are built and selectable.
Every deliverable is reviewed line-by-line by the founder — a CFA with 15+ years in valuation and a mining specialty — before it ships. The review is the product.
Recent additions to the library.
| Report after-tax NPV at 7% | C$2,135M |
| Re-created NPV | C$2,150M (+0.7%) |
The study prints a single total-taxes row, with no pools, allowances or credit timing. Its taxes are C$140M below statute over the life of mine, with refunds in Years 1 and 2: about C$340M of deductions it does not name. The model builds tax from federal and Newfoundland and Labrador statute; on that basis the value is below the study's, while pre-tax the two are close. With the study's published taxes selected, the model lands 0.7% above the report.
| Report after-tax NPV at 5% | US$1,214M |
| Re-created NPV | US$1,213M (−0.1%) |
The study publishes no annual cash-flow table, only a chart, so each annual line is built from its unit rates and yearly mine and mill schedules. It states unit-of-production depreciation, but its charted Year 1 tax is closer to the Income Tax Act's three-year write-off of mining capital, which the model uses; unit of production gives a lower value. Built from Namibian statute at the 37.5% rate, the value is below the study's. With the study's 37% rate and stated settings selected, the model lands 0.1% below the report.
| Report after-tax NPV at 5% | US$1,009M |
| Re-created NPV | US$1,007M (−0.2%) |
The study publishes no tax line: its tax is the difference between two rows of its cash-flow table. That table ends in 2037, so the study leaves out the tax on 2037 income, due in 2038. The model builds tax from Senegalese statute and the 2015 Mining Convention terms, that final payment included; the value is below the study's, while pre-tax the two are close. Without working capital, as the study carries none, the model lands 0.2% below the report. The 2016 Mining Code royalty and the removal of the investment tax credit are both selectable, and both lower the value.
Why the variance column matters. Anyone can quote a report's NPV. Re-creating it in our model from the report's own inputs and landing within a percent — or proving precisely why it cannot be reconciled to its own tables, as our Santa Cruz coverage did — is the difference between repeating a number and verifying one. That check runs on every publication, and the result is printed on each model's dashboard.
Members receive email updates as we add new projects. Paid subscriptions open at launch, with every rebuilt model viewable in the browser and new analyses as they release.