Scales that read consequence, not volume

Growth in turnover says nothing by itself: it can mean a strengthened position or a lost one, and what discriminates is not the volume but what was paid for it. A count of strikes, or the maximum range achieved, is a property of the weapon and of the effort rather than of what is being defended — the same range can mean a strike on a depot or on an assembly node, which are politically and economically different events and indistinguishable in a counter of kilometres.

The conversion is the same in both cases: stop counting what was done, and order what was lost.

The ordering reproduces the incidence principle: the discount is paid by the exporter, the infrastructure by the budget and the pipeline operator, the abandonment of market protection by the domestic producer, access to assets by the owner, and monopsony by all of them at once.

Each rung is filled by a price or a document rather than by a characterisation. Rung 2: gas to China at $258.8 per 1,000 m³ in 2026 against roughly $390 for Europe and Turkey. Rung 3: a pipeline at design capacity, and a second being designed for the same buyer. Rung 4: 98.3% of machine tools imported, China at 60–70% of those imports against roughly 20% before 2022. Rung 5, partly: a special legal regime operating from 1 January 2026, the legal concession complete and observable, while access to assets and subsoil is not yet recorded.

Partial passage is scored by component, not by impression. Rung 5 requires three things — assets, subsoil, legal regime — and one has been conceded in full, so the rung is credited as partial with the missing components named. One detail is logged separately since it works against the instrument itself: the regime restricts disclosure of owners and directors = the concession includes a concession in observability, and the widget says so rather than quietly losing resolution.

In the first half of 2026 China cut purchases of Russian sawn timber by 30% over four months and restarted its own logging in four provinces, while Russia supplies 42% of Chinese sawn timber imports.

The same pattern runs in three sectors, which is what makes it a structure rather than an episode: gas at a third below the European price on infrastructure built at the seller’s expense; coal facing restored import duties from which competitors are exempt under free trade agreements, with buyers additionally requiring the supplier to absorb them; timber cut while domestic logging restarts. In no case was compensation received in another sector for a concession made, and that is what distinguishes dependence from “partnership”.

One measurement makes the price explicit rather than inferred: goods subject to Chinese export controls became 87% more expensive for Russia on average over 2021–2024. That is neither a market price nor logistics. It is the charge for having no alternative supplier.

Four separate ladders — refining and the fuel circuit, export infrastructure and maritime logistics, the productive core, and the symbolic and strategic centre — each ordered by what the regime ceases to be able to protect: first the periphery, then depth, then the productive core, then the centre.

The discriminating statement sits on the widget’s face: a hundred strikes on border-area depots do not change what the regime can promise to defend, while one confirmed strike on an assembly node does. A count treats those as ninety-nine to one; the ladder treats them as two different rungs.

What the rungs record is qualitative in the right way. On the fuel circuit, hits on secondary refining units, where irreversibility often runs 12–18 months, are a different rung from repeated hits on primary capacity, since recovery time is a property of the object rather than of the strike. Distance appears as a rung — beyond 2,000 km from the border — not as a measure of capability but as the point at which a class of interior facility stops being safe by geography.

The upper rungs are defined so that they can be reached with no new capability at all. One requires the rate of damage to exceed the rate of restoration for three consecutive months; the highest requires repair work to cease at half or more of the damaged plants = an admission that what is repaired cannot be protected. Both are statements about the defender’s behaviour, which is the subject of the scale, rather than about the attacker’s reach, which is not.

• The external channel widget does not assess the sufficiency of Chinese support and contains no threshold at which that support becomes decisive. It measures the weakening of a negotiating position without becoming a calculator of how much backing is enough.
• The depth-of-strike widget is one-sided and says so on its face. It measures the result of strikes in one direction only, and a reader who wants a symmetric picture is told, in the widget, that this instrument does not provide one.
• Neither carries a prescription. A scale that answers “how much more is needed” has an objective function, and an objective function invites the question of whose it is.

Refutation follows the same discipline, written as observations rather than as trends. For the external channel: the second pipeline signed on export parity terms; the gas discount narrowing to the transport and insurance premium; China’s share of machine tool and equipment imports falling below 40%; exports to China growing faster than imports for three consecutive quarters.