Roadmap / 0.1.3.9.3
What is where at a site
v0.1.3.9.3, commit 434c398. File docs/RESULTS_v0.1.3.9.3_site_points_revised_rule.md.
Written at this version and not revised since. Later versions may change these numbers; see the roadmap.
Switch: none. The model is unchanged from v0.1.3.8; results are identical and the golden check passes.
Runs: none new.
Scope: changed, on Kevin's direction of 19 November 2026. The rule for civilian sites is now: describe what is there, including the published position of its parts; never record its security, how it could be damaged, a ranking by harm, or the ways in.
Headline
- Site pages now say what is where. 36 published points on 33 potash sites: seven shaft collars from the companies' filed technical reports, the site points Mosaic gives in its 10-K, two provincial survey points, and the plants, docks and works that Roger Wilco Imagery identified on satellite imagery. Each is given as its source gives it.
- The rule moved, and is stated in one place. The earlier rule kept everything inside a site out. That left out public facts a reader would come here for. The new rule keeps out what makes a list dangerous, and lets description in.
- The game is untouched. It still works on district totals and still degrades infrastructure by category, never part by part. The points are for looking a place up.
Why the rule changed
Version 0.1.3.9.2 left out the shaft collars and the separate entry for the Scissors Creek shaft, because the rule then was "nothing about the inside of a site". Kevin's objection, the same day:
- The positions are public. They are printed in filings on SEC EDGAR.
- Someone may want to use the site to research the actual places, and leaving out what is public is a disservice to them.
- The line for civilian infrastructure should be about security. A nuclear station has a reactor, an inner perimeter and an outer perimeter. It is fair to mark the reactor. It is not fair to mark the fences.
The model judge agreed, and holds two further lines of its own alongside the one on security.
The rule now, for civilian infrastructure
In: description. - What a place is, who holds it, where, how much, its status. - What is where at the site, when a filing, a government index or public imagery shows it: a shaft, a mill, a tailings area, a dock. The exact published position, with its source.
Out, and refused by the data loader: - Security. Fences, perimeters, gates, guards, cameras, access control. A point may not even be named for one. - Assessment. How a site or a part of it could be damaged, what the loss of a part would do, which part matters most. - Ranking. Sites or parts ordered by the harm their loss would do. - Ways in. How to approach or enter a site.
The last three are the model judge's. The reasoning: a list of public places is a reference work, and anyone can check it against its sources. What turns a reference work into a target list is the layer of judgement on top: what to hit, in what order, and how to get to it. That layer stays out however public the underlying facts are.
A test checks each refusal, including that a point named "Inner perimeter fence", "Main gate" or "Guard post" is turned away.
What the game does with this
Nothing. The simulation's smallest unit stays the province, with districts next. When output is lost at a site, it will be lost by category (seized, shut by its owner, damaged), on a schedule, for the site as a whole. No rule in the game will act on a shaft, a mill or a dock.
The points
| ID | Site | What | Position | Source |
|---|---|---|---|---|
| T1-01 | Rocanville | No. 2 shaft collar | 50.4721° N, 101.545° W | Nutrien NI 43-101 technical reports |
| T1-01 | Rocanville | Scissors Creek shaft collar | 50.452° N, 101.77° W | Nutrien NI 43-101 technical reports |
| T1-02 | Allan | No. 2 shaft collar | 51.9321° N, 106.072° W | Nutrien NI 43-101 technical reports |
| T1-03 | Lanigan | No. 2 shaft collar | 51.8557° N, 105.21° W | Nutrien NI 43-101 technical reports |
| T1-04 | Cory | Service shaft collar | 52.0917° N, 106.855° W | Nutrien NI 43-101 technical reports |
| T1-05 | Vanscoy | Service shaft collar | 52.008° N, 107.09° W | Nutrien NI 43-101 technical reports |
| T1-06 | Patience Lake | Mine site, as mapped in OpenStreetMap and checked on imagery | 52.0975° N, 106.373° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
| T1-07 | Esterhazy | Site point for K3 given in Mosaic's 10-K | 50.6462° N, 101.993° W | Mosaic Form 10-K for FY2024 and FY2025, with the S-K 1300 technical report summaries (SEC EDGAR) |
| T1-08 | Belle Plaine | Site point given in Mosaic's 10-K, reconstructed from a garbled print | 50.4277° N, 105.198° W | Mosaic Form 10-K for FY2024 and FY2025, with the S-K 1300 technical report summaries (SEC EDGAR) |
| T1-09 | Colonsay | Mill and headframe, north of the tailings | 51.9369° N, 105.766° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
| T1-10 | Bethune | Plant, east of the tailings pad | 50.649° N, 105.378° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
| T1-11 | Harrowby | Small pad and cooling pond beside the CN line, 12 km west of Russell (provisional) | 50.7541° N, 101.455° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
| T2-01 | Esterhazy K1 | Site point for K1 given in Mosaic's 10-K | 50.7265° N, 101.934° W | Mosaic Form 10-K for FY2024 and FY2025, with the S-K 1300 technical report summaries (SEC EDGAR) |
| T2-02 | Esterhazy K2 | Site point for K2 given in Mosaic's 10-K | 50.6574° N, 101.842° W | Mosaic Form 10-K for FY2024 and FY2025, with the S-K 1300 technical report summaries (SEC EDGAR) |
| T2-03 | Picadilly | North plant | 45.7723° N, 65.3654° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
| T2-03 | Picadilly | South works | 45.7642° N, 65.3806° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
| T2-04 | Penobsquis | North plant | 45.7723° N, 65.3654° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
| T2-04 | Penobsquis | South works | 45.7642° N, 65.3806° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
| T3-01 | Jansen Stage 1 | Service shaft collar | 51.8991° N, 104.715° W | BHP Form 20-F for FY26, 6-K filings and the S-K 1300 Jansen technical report summary (SEC EDGAR) |
| T3-02 | Jansen Stage 2 | Service shaft collar | 51.8991° N, 104.715° W | BHP Form 20-F for FY26, 6-K filings and the S-K 1300 Jansen technical report summary (SEC EDGAR) |
| T3-03 | Bethune Ramp-Up | Plant, east of the tailings pad | 50.649° N, 105.378° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
| T3-04 | Milestone Phase I | Plant and brine ponds, 8 km north of Riceton | 50.19° N, 104.301° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
| T3-05 | Tugaske | Village of Tugaske. Nothing is built. | 50.8741° N, 106.287° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
| T4-01 | Southey | Provincial survey point for the permit area (SMDI 6001) | 51.0673° N, 104.517° W | Saskatchewan Mineral Deposit Index (Saskatchewan Geological Survey) |
| T4-02 | Milestone full scale | Plant and brine ponds, 8 km north of Riceton | 50.19° N, 104.301° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
| T4-03 | Muskowekwan | Centroid of the Muskowekwan First Nation reserve. Nothing is built. | 51.314° N, 104.096° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
| T4-04 | Wynyard | Provincial survey point for the permit area (SMDI 2882) | 51.704° N, 104.017° W | Saskatchewan Mineral Deposit Index (Saskatchewan Geological Survey) |
| T5-11 | Eyebrow | Village of Eyebrow. Nothing is built. | 50.8° N, 106.15° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
| T7-01 | Neptune Bulk Terminals | Terminal, as mapped in OpenStreetMap | 49.3054° N, 123.051° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
| T7-02 | Portland Bulk Terminals | Terminal 5, as mapped in OpenStreetMap | 45.6343° N, 122.778° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
| T7-03 | Saint John Terminal | East-side bulk berth: covered storage and a jetty conveyor. Not confirmed as the potash berth. | 45.2567° N, 66.0638° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
| T7-04 | Thunder Bay Terminals | Dock and storage on Mission Island | 48.3848° N, 89.2155° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
| T7-05 | Westshore Terminals | Terminal at Roberts Bank, as mapped in OpenStreetMap | 49.0184° N, 123.166° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
| T7-06 | Port of Churchill | Port, as mapped in OpenStreetMap | 58.7759° N, 94.1945° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
| T7-07 | Prince Rupert potash terminal | Terminal complex on Ridley Island | 54.2327° N, 130.327° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
| T7-08 | Longview export terminal | Port of Longview pier, as mapped in OpenStreetMap | 46.0995° N, 122.939° W | Roger Wilco Imagery, Canada potash baseline, 19 November 2026 |
- Scissors Creek is part of Rocanville and sits about 16 km west of the main site. Its point and its frame are on Rocanville's page.
- Picadilly and Penobsquis each list both New Brunswick works, because which is which is not settled.
- A site's summary location (
loc) stays at two decimals of a degree. It is what places the site in a district.
Still open
- Marking on the imagery. Kevin said the parts can be marked without drawing attention to them. This version lists them in a table and leaves the frames plain. Marking them on the frames is a small step if wanted.
- The transport table from the imagery baseline (distance from each site to the nearest highway, railway and airport). Under the revised rule it is description and can come in. It is held for one reason: the baseline itself flags an unsettled licence question on those figures, which are built on OpenStreetMap. It comes in on Kevin's word.
- Ground at a site. Kevin's second point: some sites sit on ground that makes them hard to seize, so the game needs to know where a thing sits in relation to its terrain, and that can be recorded without naming the thing. Added to the roadmap.
- Lists to come. The rule will bite harder on electricity than on potash. For a grid, choosing which substations count is itself an assessment. That list will need its own reading of the rule before it is built.
What changed in the code
forceonforce/sites.py: the rule in the module's own words;FORBIDDEN_KEYSregrouped as assessment, security and ways in;SECURITY_WORDS; the loader checks every point.data/sites/potash.yaml:pointson 33 records; a new source key,smdi.tools/import_rwi_potash.py: brings in the points and the Scissors Creek frame.tools/build_site.py: a "What is where" table on each site page; the rule rewritten on the Sites page and on the Method page.- Tests: 36, with the refusal tests rewritten for the new rule.