Loading module…
Loading module…
Module 1 of 12
Rung one · the lens — writing that changes a decision
An analysis that nobody acts on has, for every practical purpose, not been done. The analyst may have gathered the right evidence, weighed it honestly and reached a sound judgement — and still fail completely, because the judgement reached the reader too late, or buried on page two, or wrapped in so many qualifiers that it could not be told apart from a shrug, or written about the subject when the reader needed something about their decision. This course is about the last stretch of the analytic road, the stretch where most good analysis is lost: turning a judgement into a product that a decision-maker can read in minutes, trust for the right reasons and act on. It is the discipline of intelligence writing and analytical products, and its premise is a single inversion. Ordinary writing exists to express what the writer knows; an analytic product exists to change what the reader does — or to confirm, on better grounds, what they were already going to do. Every rule in the course follows from moving the reader to the centre. The method rests on six elements, present in every product from a few-line alert to a long strategic assessment. The reader is the person who decides what, by when — the product's reason to exist. The bottom line is the answer, stated first. The judgements are the analyst's claims about what is true or what will happen, each carrying a probability and a separate confidence. The reasoning is the evidence, assumptions and alternatives that support those judgements, kept visible so the reader can test them. The sources are what the reasoning rests on, graded for reliability and credibility and protected so that telling the reader about them never exposes them. And the form is the vehicle — format, structure, visuals and length — chosen for the reader's need rather than the writer's habit. Every product the course builds answers four questions: who decides what, by when? what do we judge? how sure are we, and on what? what must the reader see first? One fictional tasking carries the whole course. Kestrel Motion, the servo-motor and drive maker you may have met in Courses 007 and 010, set up a small market & supply intelligence cell after the flood at its Aldren plant: a lead analyst and one analyst, reporting to the head of procurement. On Day 0, Vashkar publishes a draft export-licensing rule on rare-earth processing, and the COO asks the cell: "Do we need to act on magnets now — and if so, what?" The product is due on Day 5, for the executive committee (ExCo) meeting on Day 7. This module lays out the product anatomy, the transformation the course produces and the thirteen-rung method — and it names the boundary up front. The product serves the reader's decision, and the decision belongs to the reader: the analyst sets out judgements, options and their implications, never what to decide and never what the reader would like to hear. Every claim is labelled fact, assessment or speculation, with its provenance; speculation is never dressed as fact. Sources and people are protected — people are named by role only — and when a judgement proves wrong, it is corrected in the open.
The six elements are the periodic table of this course — the complete alphabet from which every analytic product is built — and the anatomy is those six drawn in the one arrangement that works. The mechanism it shows runs in two directions at once. Analysis runs bottom-up: evidence is gathered and graded, it supports or undermines competing explanations, the surviving explanation becomes a judgement, and the judgements compress into a bottom line. Reading runs top-down: the reader starts at the bottom line, reads the judgements if they need more, the reasoning if they doubt it, and the sources and annex only if they want to check. The single most common failure in analytic writing is to write in the direction the analysis ran — from evidence up to conclusion — so that the reader, reading in the opposite direction, meets the background first and the answer last, if at all. The anatomy fixes that by pinning the elements to positions. The reader sits outside the page because the reader, not the subject, is what the page is organised around: a product about "Vashkar's rare-earth policy" and a product about "whether Kestrel must act on magnets before Day 7" can use the same evidence and be entirely different documents. The bottom line sits at the top because the reader may stop there. The judgements each carry two tags because they answer two different questions — how probable is this? and how much weight will our basis for saying so bear? — and a judgement with only one of them misleads. The reasoning sits in view because a reader who cannot see why the analyst believes something cannot tell a strong judgement from a weak one stated firmly. The sources carry grades, not names, because the reader needs their weight and the source needs protection. And the form is the page itself: its length, its format, its annex. The completeness argument is simple: to act on analysis, a reader must know what the product is for, what the answer is, how probable and how well founded each judgement is, why the analyst believes it, what it rests on, and in what shape it arrives — and every module lights one of these. Module 3 lights the page structure; Module 4 the arrows from evidence to judgement; Module 5 the hypothesis matrix beside the page; Module 6 the probability yardstick and the confidence tags; Module 7 the source chips and their grades; Module 8 the different page shapes; Module 9 the annex visuals; Module 10 the reader and the decision diamond; Module 11 the review marks on the page; Module 12 the page shapes of each discipline. What falls outside the anatomy — telling the reader what to decide, exposing a source or a private individual, presenting speculation as fact — is ruled out by the boundary in Module 2.
Traced instance: the Vashkar tasking. Kestrel's exposure is concrete. Its magnets come through two suppliers, Arvel Components and Dunmore Precision, and both are supplied by Lenhu Magnetics in Vashkar, which buys its oxide from Vashkar Rare Earth Refining (VRR). Kestrel can survive about 5 weeks without new magnets (2 weeks on hand plus about 3 at sea); qualifying the alternative, Pellmark Magnetics, whose recycled feedstock is refined independently of VRR, takes about 22 weeks — a 17-week gap. The cell grades each item as it arrives. The draft rule itself, in Vashkar's official gazette (E1, A1), covers "rare-earth materials and the products of their processing" — wording that could mean oxide only or could reach sintered magnet blocks too. The ministry says the aim is to "keep more value added at home" (E2, A2: official, but motive-laden). A trade-press report cites "ministry sources" saying blocks will be included (E3, C3). Vashkar's sinterers' association, of which Lenhu is a member, is lobbying publicly to keep blocks out (E4, B2: credible, but an interested party). A note from Lenhu's sales team, relayed through Arvel, says Lenhu "expects licences to be a formality" (E5, C4); the product conveys its weight without exposing the route or Arvel's contact. Three years ago Vashkar licensed exports of another mineral, and the licences came, but with delays of several weeks (E6, B2). The consultation closes in about 8 weeks (E7, A1). And an unverified social-media claim says the rule "will be dropped" (E8, F6). Four hypotheses compete — blocks included, oxide only, withdrawn or shelved, adopted with licences a formality — and the one least inconsistent with the evidence is the first. The product the cell delivers on Day 5 is a two-page assessment with an annex, and its bottom line reads: We judge it likely (55–80%) that Vashkar's final rule will require export licences for sintered magnet blocks as well as oxide, which would put licensing between Lenhu and both of Kestrel's magnet suppliers. Moderate confidence: the draft's scope wording is ambiguous and the consultation is open. Because qualifying Pellmark takes ~22 weeks and a licensing interruption could exceed Kestrel's 5-week survival time, the decision window for the magnet options is now. Three key judgements follow, each with both tags; then the options — start Pellmark qualification now, raise magnet stock from 2 to about 10 weeks, or wait for the final rule — laid out with their costs, and the choice left to the ExCo; then the indicators that would change the cell's mind. The course follows this product past delivery: to the ExCo's decision, to Vashkar's final rule, and to the moment one of the three judgements proves wrong and the cell has to say so.
The gap between the two is the whole value of the course, and each of the essayist's habits fails the reader in a different way. The first is the order of discovery. School and academic writing build towards a conclusion, which is sound when the reader will read to the end; a COO reading between meetings may read one paragraph. An essayist's draft on the Vashkar rule opens with the history of the country's rare-earth policy, walks through each report and arrives at "Kestrel may wish to consider its options" on page two — the reader who stops early leaves with background and no answer. Putting the answer first is not a house quirk: military writers call it BLUF, bottom line up front, and Barbara Minto's pyramid principle teaches the same order to consultants, because busy readers everywhere read top-down. The second habit is mixing fact with opinion: "the rule will cover blocks" reads as a fact when it is an assessment built on an ambiguous gazette text and a C3 press report, and a reader who cannot see the difference cannot weigh it. The third is hedging everything or nothing. "Could", "may" and "it is possible that" in every sentence leave the reader with no judgement at all; a flat "will" leaves them with false certainty. Sherman Kent's essay "Words of Estimative Probability" (1964) recounts what happens in between: when he asked colleagues who had agreed an estimate calling an attack on Yugoslavia in 1951 a "serious possibility" what odds they had meant, the answers ranged from about one in five to about four in five — the same words, agreed by the same people, carrying opposite messages (Module 6 dissects why, and what replaces them). The fourth habit is hiding the sources and the reasoning, which forces the reader to trust the writer's tone instead of the evidence. The fifth is writing about the subject rather than the decision — a fine essay on Vashkar's industrial policy that never says whether Kestrel's 5-week survival time is at risk. The sixth is advocacy: the essayist who has formed a view writes "Kestrel must qualify Pellmark immediately", which takes a decision that belongs to the ExCo, or senses what the COO would prefer and softens the judgement to match. The seventh is never correcting: a judgement that proves wrong is left to be forgotten. The intelligence writer replaces each habit with a discipline. The bottom line comes first, graded: likely (55–80%), moderate confidence. Each judgement carries a probability from the course's yardstick and a confidence of high, moderate or low — or not known where the gap is real. The reasoning is on the page: which evidence fits which hypothesis, which assumption carries the most weight, which alternative was weighed and why it lost. The sources are graded — the reader sees that one input is a C4 relayed note and one an A1 official text — while the route by which the note arrived stays protected. The format fits the decision: a two-page assessment with an annex, because the ExCo must choose between options with different lead times and costs. The writer lays out options and implications and leaves the choice to the ExCo. And when a judgement proves wrong, the writer issues a correction and records the miss.
The course is thirteen modules, and the structure follows the four questions. After this orientation, Module 2, The Analyst's Discipline & Boundary, installs the six commitments that govern everything else. Modules 3 to 5 answer what do we judge? The Intelligence Writing Standard teaches the bottom line first, separating what you know from what you judge, precision, structure as a service, objective language and the balance between brevity and completeness. Analytic Reasoning Made Visible puts the argument on the page — claim, evidence and warrant, key judgements and their support, stated assumptions, alternatives, what would change our mind, and the "so what". Structured Analytic Techniques protects the judgement from the mind's shortcuts, drawing on Richards Heuer's Psychology of Intelligence Analysis (1999) and his analysis of competing hypotheses, the key assumptions check, devil's advocacy and red teams, and reframing methods such as Gary Klein's pre-mortem. Modules 6 and 7 answer how sure are we, and on what? Estimative Language & Confidence teaches the seven-band probability yardstick published in the US intelligence community's analytic standards (ICD 203), the difference between probability and confidence, over-hedging and false precision, and calibration and scoring — Brier scores and Philip Tetlock's forecasting research, including the Good Judgment Project. Sourcing, Attribution & Protection teaches the A–F / 1–6 grading system of the NATO and Admiralty tradition, how sourcing can burn a source, and open-source attribution with people in role. Modules 8 to 10 answer who decides what, by when, and what must they see first? Product Formats matches the vehicle to the need — the current-intelligence item, the warning product, the estimate, the deep-dive assessment, the briefing; Visuals & Data in Analytic Products teaches graphics that do not mislead, in the tradition of Edward Tufte's work on honest data display; and Writing for Decision-Makers writes to the decision, tailors without pandering and delivers unwelcome news with integrity. Modules 11 to 13 make the product hold up: Review, Editing & Quality Control covers the self-edit, peer review and dissent, the standards checklist, red-team review and corrections; Analytic Products Across the Disciplines applies the craft to the products of the other Tricellix courses — the country and risk assessment, the due-diligence memo, the competitor brief, the crisis situation report and the OSINT investigation report; and the runs the whole method on the Vashkar assessment. Every module from 2 to 12 is built the same way — six , the sub-techniques of its rung, each dissected through six fixed fields: it is, its (drawn on the product anatomy), the as a numbered procedure, its and failure modes, to use it and when not, and it connects to the rest. The Vashkar tasking is traced through every atom, alongside the other products Kestrel's cell writes: the current-intelligence note that flagged the draft on Day 0, a warning product with the final rule's publication as its tripwire, the flood situation report from Course 010, a counterparty memo on Brenholt Foundry and a competitor brief on , a rival motor maker that also buys from Lenhu. The world is fictional so that the method can be shown whole without real sources, real classified material or anyone's real reputation; the people in it are named by role only — the COO, the CFO, the head of procurement, the lead analyst and the analyst.
This is a dissection course, not a style guide: every technique is broken down to its mechanism and its procedure, so you finish not with opinions about good writing but with a repeatable method for building a product a reader can act on. Three things to hold onto as you go. First, write for the reader's decision — and leave the decision to the reader: before you draft a sentence, write down who decides what, by when, and check every paragraph against it; the product's job is to make the decision better informed, not to make it, and the moment it starts telling the reader what to choose, or what they want to hear, it has stopped being analysis. Second, show your grades: every judgement carries a probability from the yardstick and a confidence of high, moderate or low (or not known), every claim is marked as fact, assessment or speculation with its provenance, and the reasoning stays on the page — a reader who can see why you believe something can use your judgement even when they disagree with it. Third, protect, and correct: convey how much weight each source bears without exposing who or where it is, keep private individuals out of the product, and when a judgement proves wrong, say so first, plainly and in writing — a cell that corrects in the open is trusted the next time; a cell that goes quiet is not. One note on scope: the course teaches writing and analytic method on a fictional universe and on published, well-documented real-world practice; it uses no real classified material and no real sources, and nothing in it is advice on any real country's trade rules. Hold those three and the rest is detailed machinery. Module 2 makes the discipline explicit first, because a product built to move a decision is only safe when it is held to a strict standard of honesty.
Before the first bottom line is drafted, the course installs the discipline that governs the whole method. Module 2 is the analyst's boundary: the product serves a decision, not the writer, so the reader's question sets its content and its order; objectivity, which lays out options and implications and never tells the reader what to decide or what they want to hear; fact, assessment and speculation with provenance on every claim; calibrated language — the seven-band probability yardstick and the separate high / moderate / low confidence scale, with not known for real gaps; protecting sources and people, which conveys a source's weight without its identity or route and names people by role only; and honest dissent and correction, which records disagreement and says plainly when the product was wrong. The product anatomy is a powerful instrument; Module 2 is what keeps it objective and honest.
The signature image of the course: one analytic product drawn as an anatomy. On the left, the reader — a decision diamond for the COO and the ExCo, with its deadline (product Day 5, ExCo Day 7), the three options the ExCo will choose between, and the reason the window is now: a 5-week time to survive against a 22-week time to recover. In the centre, the product as a page, stacked in the order the reader needs it: the bottom line up front, the key judgements with probability and confidence tags, the reasoning with its hypotheses, the indicators that would change the cell's mind, the graded source chips and the annex. On the right, the evidence E1–E8, each graded on the A–F / 1–6 system, feeding the reasoning — the two dashed feeds are the interested or weakly graded items, and E8 is noted but not relied on. Every later module lights one part of this page.
The transformation the course produces. The essayist writes in the order of discovery, so the answer sits below the point where the COO stops reading; mixes fact with opinion; hedges everything or nothing; hides the sources and the reasoning; writes about the subject rather than the decision; slides into telling the reader what to do or what they want to hear; and never corrects. The intelligence writer puts the graded bottom line above that point, grades every judgement for probability and confidence, shows the evidence, assumptions and alternatives, protects sources while conveying their weight, fits the format to the reader's decision, stays objective, and corrects in the open. Same evidence, same deadline; the difference is whose needs set the order of the page.
How the course is built. Phase 1 sets the lens and the boundary (Modules 1–2). Phase 2 builds the judgement — the writing standard, visible reasoning and the techniques that protect it (Modules 3–5). Phase 3 grades it — the language of uncertainty and the sourcing (Modules 6–7). Phase 4 shapes it for the reader — formats, visuals and the decision-maker (Modules 8–10). Phase 5 makes it hold up — review and correction, the craft across the disciplines, and the capstone (Modules 11–13). Serving the decision, analysis not advocacy, graded and protected sources and open correction run under every rung.