Territory taken and losses incurred can be related in either direction, and the two forms behave differently exactly where the reading matters. One goes to infinity at the moment the phenomenon reaches its extreme; the other stops at zero and keeps working.
| Form | Behaviour as advance falls | As an instrument |
|---|---|---|
| km² per 1,000 losses — the form on the scale | Falls toward zero and stops there | Bounded below, degrades smoothly, stays readable at any rate of advance |
| Losses per km² — the inverse | Rises without limit; undefined at zero advance | Unbounded above, dominated by the last few points, breaks when the reading matters most |
An instrument that goes to infinity when the phenomenon reaches its extreme is not measuring the extreme, it is failing at it. The bounded form carries the zones.
The inverse is still published, in one place and for one reason. The cost block gives about 845 losses per square kilometre, since historical cases were counted that way and comparability is the point there. The rule is not that one form is correct: the form carrying the zone must be the bounded one, and the other form may appear where a reader needs to compare with an outside series.
The current value is 1.18 km² per 1,000 losses. Absolutely, that number means nothing to a reader without a yardstick, and any absolute threshold an analyst picks invites the question of where it came from.
Base: the 2025 average 4,831 km² against 416,600 losses = 11.6 km² per 1,000 Zones as shares of that base green ≥ 100% yellow 70–100% orange 30–70% red under 30%, second component also met Current: 1.18 = 10% of base
Two things follow. The base is the measured party’s own performance a year earlier, under the same command, against the same adversary, with the same definitions on both series = a critic disputing it is disputing a fact rather than a judgement. And the scale carries an explicit refusal: an offensive costing a thousand men per square kilometre is not a “norm” merely because it has lasted long enough to look like one.
The refutation threshold is set the same way and deliberately loosely: a return to 8.1 km² per 1,000, which is 70% of the 2025 level, with a positive net territorial balance for three consecutive months. That sits below the 2024 average of about 9.8, thus the falsifier counts even at a rate worse than the year before last. A falsifier calibrated where it can plausibly trigger is worth more than one placed where it never will.
Numerator and denominator are both estimates, and for each there is a choice of series that would produce a more dramatic figure. Advance is taken as geolocated confirmed control, not the larger figure that includes infiltration, since infiltration is presence without holding and the publishing institute acknowledged in July 2026 that its infiltration map overstates presence. Losses are taken at the median of an independent estimate rather than at the higher figures published by a party to the conflict.
Both choices work against the panel’s reading: on the wider advance measure and the higher loss series the figure would be 2.84 km² per 1,000 rather than 1.18. The smaller variant is taken on both.
The cost is a lag. Infiltration converts into confirmed control within weeks, and higher losses surface in family payments and funerals within one to three months, thus the indicator confirms what has already happened rather than what is supposed — which is what makes it usable by someone who does not share the author’s priors.
Alongside that sits a disclosure cutting the other way, stated in the widget: three independent sources of understatement — the acknowledged overstatement of presence, units not reporting lost positions, and the fog of war maintained by the other side — all push in the same direction, toward understating losses and overstating advance. The real price of ground is most probably higher than the published one, and the published one is used anyway.
It has been argued that the red zone ought to carry a second, political condition: that the loss of initiative be registered in Russian public discourse. The argument is reasonable and the condition was rejected.
A regime that acknowledges nothing would hold the widget at orange indefinitely, whatever happens on the ground = the instrument would be measuring the willingness to admit a fact rather than the fact.
The zone reflects a physical fact and is not conditional on that fact being acknowledged by the party it describes.
Acknowledgement keeps its significance and is carried separately. It determines whether the war can be sold at home as a success, not whether it can be continued. As an observation it is a useful leading marker, the cheapest version being complaints by Russian military bloggers about unreported losses of positions. It sits outside the zone definition, where it cannot be used to freeze the scale.
A related objection, and why it creates no category. A low rate of advance may be a choice rather than a limit. Not choosing reduces to one of three things: 1) incapacity; 2) restraint under an external constraint, which operationally is the same incapacity; 3) a preference for inflicting damage another way, which is a choice for greater rather than lesser effectiveness. The widget measures the price of ground, not intent, and in that formulation the distinction is not required.
None of this is specific to a war.
Three failure modes.
A scale the measured party can freeze by saying nothing is not an instrument, it is a courtesy.