A self-assessment scale can only tell you what you already believe about yourself. A ratio whose numerator is somebody else's behaviour cannot.1
⭐ The Hanumanic Strategy framework's measurement document contains four or five of the second kind, scattered across two of its three layers, never named as a class — and they are the answer to a criticism this build has twice levelled at the corpus's instruments.23
The measurement document is built in three tiers.1
Quantitative — revenue trajectory, product-validation rate, inbound-to-outbound enquiry ratio, referral and retention percentages, learning-to-implementation time. Qualitative — a Dharmic Alignment Scale self-rated one to ten daily, energy-sustainability assessment, warrior-sage balance evaluation.
⚠ This build has already read those two layers and read them correctly: everything countable is counted honestly, everything uncountable gets a self-rated scale invented for it, and a scale scored by the person being scored cannot come back low.23
⚑⚑ But there is a third tier, headed Compound Development Indicators, and it does something neither of the other two does.1 It stops asking the practitioner how he is doing and starts asking what other people did.
Read the compound layer's entries as arithmetic and the pattern is immediate.1
"Recognition received vs. recognition sought." ⚑ Both terms are events and both are countable. Asks made, pitches sent, applications submitted — that is the denominator, and it is entirely the practitioner's. Invitations, citations, approaches, referrals — that is the numerator, and none of it is.
"Influence emergence through competence vs. self-promotion." "Natural positioning vs. forced market creation." "Sustainable authority building vs. temporary visibility." ⚑ Same shape each time: something the practitioner did on the bottom, something other people did on top.
⭐⭐ A ratio of that construction has a property no scale in the document has: it can go down while the practitioner is trying harder. ⚠ Indeed working harder at the wrong thing pushes the denominator up and the ratio down, which is precisely what a measurement is supposed to do and precisely what a self-rating cannot.
Gathered from both layers, the framework holds five instruments of this construction and does not know it.1
The inbound-to-outbound enquiry ratio — sitting in the quantitative tier, and already the strongest thing in the document. Referral rate and client retention — other people's repeat decisions. Speaking and consulting opportunity frequency — other people's invitations. The recognition-received-to-recognition-sought ratio. And "problems solved for others without expectation", which is a count of a thing that can only be verified by whether the other party thought a problem was solved.
⚑ Five, and they are the only entries anywhere in the document that a determined self-deceiver could not survive.
⚠ They are also presented in three different tiers under three different headings, in a list also containing self-rated ten-point scales, with nothing marking them out. ⭐ The corpus built the good instrument five times and never noticed it was building the same thing.
⚑ The usual defence against a gameable metric is to add more metrics, and it does not work — a practitioner who can score his own alignment at eight can score five dimensions at eight.
⚑⚑ What the compound ratio does instead is relocate the evidence. It does not ask the practitioner to be honest; it asks him to count — and the thing being counted happened in somebody else's calendar. ⭐ Honesty is no longer load-bearing, which is the only durable way to build an instrument for a person assessing himself.
⚠ Two conditions have to hold, and the document states neither. The denominator must be recorded as it happens — asks are forgettable, and a practitioner reconstructing them a month later will remember the ones that worked. And the ratio has to be kept over a window long enough that a single lucky invitation does not swamp it.
⭐ Get both and the instrument is unusually strong for something a solo operator can run alone. Miss either and it degrades into a story about how things are going, which is what the tier above it already is.
One entry in the compound tier does not have this shape and it is worth separating.1
The Integration Coefficient measures cross-track reinforcement — "how cognitive development enhances AI collaboration", "how spiritual practice improves strategic thinking", "overall synergy vs. isolated track development."
⚠ There is no external party anywhere in it. ⚑ Nobody but the practitioner is positioned to observe whether his contemplative practice improved his strategy, and the claim is not the sort of thing that produces a countable event.
⭐ Which makes the coefficient a good idea filed in the wrong tier. It belongs with the qualitative scales, and its presence among the compound indicators is probably why the compound tier reads as more of the same — one unfalsifiable entry in a list of five falsifiable ones is enough to make a reader stop reading the list.
The document closes with success thresholds at three stages, and each stage has two of them.1
Foundation, months one to six. Development, months seven to eighteen. Mastery, year two onward. ⚑ Each carries a Minimum Viable Progress line and an Optimal Development line — a floor and a target, stated separately.
⭐ The two-tier threshold is a better design than it looks and it is rare in practice literature, which almost always states only the target. ⚑ A single stated standard has to serve as both aspiration and pass mark, and it cannot: set at the aspiration it fails everybody, set at the pass mark it inspires nobody.
⚠ The floors are also the more informative half and the document treats them as the afterthought. Foundation's floor is daily practice established, basic competency in each area, one validated offering, active participation. ⚑ Every item is binary and none is a matter of degree — which is what a floor has to be if it is going to be checkable in a bad month.
The document ends by naming the failure it was built against: the apparatus exists "preventing optimization of metrics at the expense of core principles while maintaining accountability for real-world effectiveness."1
⚑⚑ That is Goodhart's law stated without the name, and stated as a design constraint rather than as a lament — which is unusual, because the observation is normally deployed to explain why measurement failed after it failed.
⚠ But naming the hazard is not the same as escaping it, and the document escapes it only in the tier it does not draw attention to. ⭐ The self-rated scales are exactly the entries that can be optimised at the expense of the thing they stand for, because optimising them costs nothing but a shifted judgement. The five ratios cannot be, and the document does not say so.
Log the denominator live. ⛔ Every ask, pitch, application and approach, recorded at the moment it is made. This is the single point of failure: a reconstructed denominator is a flattering one.
Keep the five ratios and drop the scales. ⚑ Inbound-to-outbound, referral and retention, invitations received, recognition received over recognition sought, problems solved that the other party would confirm. ⭐ Five is enough, and they are the five that can disappoint you.
Move the Integration Coefficient down a tier. ⚠ It is a genuine question and it is not a measurement. Keeping it among the ratios makes the whole tier look self-reported.
State a floor and a target for each stage, separately. ⚑ The floor must be binary and checkable in a bad week; the target may be as demanding as you like.
And set the window before you set the metric. ⚠ A ratio over a quarter is an instrument; a ratio over a fortnight is noise wearing an instrument's clothes.
Source. "Hanumanic Strategy: Measurement Framework", one of the framework-specification documents of the extension corpus, read in full at the phase-10b sweep 2026-08-19.1 personal-notes. [EXTENSION] — the three-tier structure, the compound indicators, the cadence and the two-tier thresholds are the author's construction.
⚑ This page corrects a reading this build has made twice. Both the-anonymity-test and the-rama-test-and-the-decision-filters cite this document for the split between countable and uncountable metrics, and both conclude that where the corpus cannot count it invents a scale.23 ⭐ That is true of the qualitative tier and false of the compound tier, which is in the same document, below both, and does something the earlier reading did not register. ⚠ The earlier judgement is not withdrawn — it is bounded to the tier it was drawn from.
⚠ The five-instrument count is this page's, not the document's. The document names no class, draws no line and presents the ratios interleaved with self-rated scales across three headings. The grouping is an inference from their construction and should be read as one.
⚠ Bounded on the corpus's other measurement documents. At least two further assessment systems exist in the folder and are treated on their own pages; whether either contains ratios of the same construction has not been checked.
Open question. ⚑ Has the author ever run any of the five? The document specifies a daily five-minute and ten-minute cadence, a weekly synthesis, a monthly review and a quarterly assessment, and nothing found in the corpus reports a figure produced by any of them.
Second open question. ⚠ What is the correct window? The framework's cadence implies the ratios are consulted monthly, but a solo operator's invitation rate is lumpy enough that a month may be indistinguishable from noise. The document sets no window and the question decides whether the instrument works at all.
⚑ The author built his best instrument and buried it under his weakest.
⚠ The document's visible architecture puts the self-rated scales in the middle tier, headed Qualitative Assessment Dimensions, with the confident presentation of a system's soft-skills half. ⭐ The compound tier arrives afterwards, under a heading that sounds like a summary, and reads at first pass as a synthesis of the two above it rather than as a third and different thing.
⚑⚑ So the ordering is doing the damage. A reader who has just been offered a daily one-to-ten scale for their own dharmic alignment has already formed a view of what kind of document this is, and the five ratios arrive too late to change it. ⚠ This build formed exactly that view, twice.
⭐ The convergence is that the author knows what a good instrument looks like — he wrote the Anonymity Test, which is the sharpest self-assessment question in the whole corpus, and he wrote the inbound-to-outbound ratio, which is the sharpest measurement.2 ⚠ What he does not do is separate them from the material they sit next to. ⚑ The corpus's problem here is not judgement. It is layout.
The Recognition Signs Are Other People's Reports. ⚑ The same corpus, the same move, opposite result — that page holds a set of spiritual progression markers that also rely on other people's behaviour, and it concludes they are worthless because of it.
⚑⚑ The pairing produces the rule neither page states: relying on other people is not what makes an instrument strong or weak. Having a denominator is. ⭐ The recognition signs ask whether anyone has noticed a change in you — and the population who might have noticed is unbounded, so no amount of silence counts as a negative result. The compound ratios ask how many invitations arrived against how many were sought, and the sought are finite, dated and yours.
⚠ Same evidentiary source, same corpus, one page apart in the same author's thinking, and the difference between an unfalsifiable claim and a working measurement turns entirely on whether the bottom of the fraction was written down. ⭐ Which is a general instrument: given any indicator that rests on other people's behaviour, ask what the denominator is. If there isn't one, it is a mood, not a metric.
The Four-Minute Fix That Wasted Twenty-Five. ⚑ Sutherland's rail case is Goodhart in its classic form: journey time was legible to a finance department and passenger experience was not, so the organisation spent enormous sums shaving minutes off a ride nobody wanted shortened.
⭐ The compound ratio is the structural answer to that case rather than a moral one. ⚑⚑ The Euro-Star metric was gameable because both its terms were inside the organisation's control — the engineers could move the number without any passenger changing their mind. A ratio whose numerator is supplied by the customer cannot be moved that way at all: the only route to a better number runs through somebody outside deciding differently.
⚠ So gaming-resistance is not a property of what a metric measures. It is a property of who supplies its terms. ⭐ Which is a testable design rule and a cheap one: for any dashboard, mark each number by who produced it, and treat the ones produced entirely in-house as management information rather than evidence.
⚑ Five instruments that require other people to act, in a corpus written by and for a person working alone.
🧭 The first open boundary: the compound ratios only work at volume, and the corpus's own reported volume is very small. ⚠ Elsewhere it concedes four paid subscribers and a few hundred pounds a year. A recognition-received-over-recognition-sought ratio computed on single-digit numerators is not a measurement of authority; it is a record of two or three events. ⭐ Which means the framework's best instrument is the one its author was least able to use when he wrote it, and nothing in the document acknowledges a floor below which it does not function.
🧭 The second, and it is the more interesting: an instrument that depends on other people's behaviour is measuring the market as much as the practitioner. ⚑⚑ A falling inbound ratio may mean the work got worse, or that the field got crowded, or that the people who used to invite you moved on. The framework treats the number as a report on the self. ⚠ It is at least as much a report on the world, and separating the two is a problem the document does not know it has.
🧭 The third is the chapter, if there is one here. ⭐ Almost every self-improvement measurement system in circulation is built from self-report, and almost every one is unfalsifiable for the reason this document's middle tier is. ⚑ The repair is not more rigour or more honesty. It is arithmetic: put something you did on the bottom and something someone else did on top. 🧭 Nobody appears to have written that down as a general design rule, and this corpus came within one heading of doing it by accident.