How Tom Clancy built a secret world from things found in plain sight

In 1976, information about the Riga incident reached the Western press gradually. Clancy says he read the story in the Washington Post, noticed that it was unusual and filed it away. In the talk he almost performs the gesture: a fact enters a mental shelf and gathers dust. There is no novel, character or conclusion yet. There is an anomaly with potential.

Good working archives contain such suspended objects. They are neither ordinary bookmarks nor claims ready for citation. They are observations that do not yet fit a model. Retaining them without conclusion leaves room for a later connection: a new doctrine, a comparable actor or a capability that changes the stakes.

When Clancy returned to the story, he did not copy the event. He extracted a mechanism: an insider seizes a military platform and forces two states to infer intent before time expires. He changed the frigate into a strategic submarine, changed the motive and destination. A historical episode became a machine for generating questions.

For intelligence, the discipline has two faces. Retain the anomaly; when you reuse it, retain the source version that existed at the time. Storozhevoy shows why. The Western defection narrative was coherent enough to generate an extraordinary novel, while Sablin’s actual purpose was probably different.

In 1980, Clancy joins the U.S. Naval Institute and attends its annual meeting. The ceremonious consensus irritates him, and he sends a letter saying so. A week later he receives $35. He recalls the moment as revelation: someone had paid him for something he wrote.

An article follows on a possible basing scheme for the MX intercontinental ballistic missile. Editors request a rewrite and, in the process, connect him with a naval officer in the Office of the Secretary of Defense. Clancy invites him to lunch on the Monday after Argentina occupies the Falklands. The officer spends roughly ten minutes explaining how Britain will retake the islands, then an hour telling sea stories. Clancy remembers the prediction as remarkably close to what followed.

What he takes from the table is more than information about a campaign. He learns “what kind of person drives submarines”. He had assumed the technical data would be findable. He did not know how someone thinks when going to work in a vessel designed to sink. He discovers a combination of fighter pilot, engineer and believer in his own competence.

The professional institution supplied two things that a library could not: feedback from people who recognised the domain and access to the questions the domain considered alive. A community of practice validates the finished product and corrects the model while it is being built.

He begins writing seriously in August 1982. By 28 October he has roughly 300 pages and asks editor Marty Callahan for a professional judgement. Callahan reads two chapters and tells him to finish the book. Clancy returns to page one. On 27 February 1983 he completes the manuscript; the next day he walks into the Naval Institute with 723 pages.

The manuscript goes to two submarine officers. Clancy summarises the verdict in a useful distinction: he had made many stupid little mistakes, while the big things held together. Small errors matter because practitioners see them immediately. The large system matters because the story survives their correction. Technical review requires both scales.

Then comes Connie Buchanan, the editor Clancy describes with his characteristic mixture of admiration and dread. Marginal comments, deletions and months of negotiation feel like tattoos placed on his child. The comparisons are comic; the result is serious. Detail does not remain in the book simply because the author worked hard to acquire it.

For an intelligence product, editing controls cognitive load. A paragraph can be true, relevant to the field and fascinating to its author, then conceal the judgement that must be retained. Specialist review checks the world; editorial review checks the reader’s route through it.

After editing, the book finds its readers through a chain no marketing plan would have drawn. Jerry O’Leary, a retired Marine colonel writing for the Washington Times, gives a copy to someone flying to Buenos Aires so that it can reach US ambassador Frank Ortiz. She reads it on the flight, later buys a case, and one copy finds its way under Ronald Reagan’s Christmas tree. The president’s comment propels the novel onto the bestseller lists. Clancy tells the route as evidence for Irish superstition; for an intelligence product it is also a reminder that a judgement does not circulate by itself. It needs a carrier, a moment and a network that grants it attention.

“The principal sources for my information were three books and a war game,” Clancy says at minute 21. The books were Ships and Aircraft of the U.S. Fleet, Guide to the Soviet Navy and Combat Fleets of the World. They supplied the inventory: classes, dimensions, propulsion, crews, weapons and range.

Harpoon supplied something else. It was a naval wargameA structured simulation of conflict in which participants make decisions under defined conditions and rules. It is useful for relationships and dependencies, not as prophecy. played with maps, counters, tables and dice, designed by Larry Bond for people wishing to simulate modern conflict at sea. Its first manual explained mechanics: how a sensor searches, how sound propagates, how a contact is classified, how platforms move and what happens when a weapon hits. The second was a database of ships, aircraft, radars, sonars and ordnance.

Larry Bond later noticed an almost comic trace of the manual in the novel: the Soviet “Type F” torpedo carried a designation he had invented for the Harpoon rules when little public nomenclature was available. The detail shows how deeply the instrument entered the fictional world.

The point becomes convincing when the object starts working in front of the room. Clancy asks for the name of a ship; someone calls out Coral Sea. He finds the carrier and reads out speed, displacement, propulsion, crew, weapons and range. Then he turns to the HARM anti-radiation missile: reach, warhead, weight and probability of killThe estimated probability that a weapon will produce the intended effect on a target under specified conditions.. For the price of a cheap meal, he says, he had received the equivalent of thousands of dollars in reference works: compressed into a form he could query while writing.

The real value was not numerical density. The manual was a bridge document: it had been forced to explain relationships. A treatise on acoustics may describe refraction perfectly. The rules of Harpoon had to say what a convergence zoneA distant region in which refraction returns ocean sound towards the surface, sometimes enabling sonar detections at unintuitive ranges. does to submarine search and when it changes a decision in the game. Clancy learned the system in a form he could manipulate.

He tore his copy apart and placed it in a ring binder. A book became a working instrument: pages could be compared and recovered in the middle of a scene. The digital equivalent is not a folder containing hundreds of PDFs. It is a structure in which claim, source page, definition, contradiction and judgement remain seconds apart.

Clancy says without false modesty that technical research was easy and fun. The blank paper was hard. Between the numbers and the page, however, lay a problem the reference works could not solve: tacit knowledgeKnowledge acquired through practice and difficult to exhaust in a manual: what operators notice, which exceptions matter and how decisions are actually made.. What does a credible order sound like? What does an operator notice before checking an instrument? Which rule is followed rigidly and which is bypassed by everyone competent?

His insurance agency stood near the Calvert Cliffs nuclear plant. Clients included former members of the nuclear Navy. They carried the profession in their voices, jokes and memories. One explained “Crazy Ivan”, the sudden turn by which Soviet submarines checked the blind area astern. Clancy later could not identify which client; the knowledge had entered memory through ordinary conversations.

He then sought direct experience: an Air Force fighter squadron scrambling, a week aboard the frigate USS Gallery, driving and firing an M1 tank at Aberdeen, travelling in England for Patriot Games and watching the FBI Hostage Rescue Team enter a room with live ammunition and volunteers on a sofa.

Such visits did not make him a submariner, pilot or agent. They corrected the bodily imagination: how long an action takes, what noise it makes, how much space it occupies and what “minimum force” means to a police unit. For the analyst, contact with practice identifies where abstract data has lost friction and where it needs verification.

Larry Bond recalled a conversation with Arkady Shevchenko, the former UN under-secretary-general and adviser to the Soviet foreign minister. Bond asked about processes. Clancy wanted to know where the Politburo met, what the room looked like and what people drank. The question appears decorative only while the institution is imagined as a diagram.

A room tells you who enters, who waits and who can approach the centre. Speaking order reveals informal authority. Drinks, notes, telephones and the way a meeting ends describe ceremony, control and intimacy. Domestic detail becomes political infrastructure.

The same logic appears in his account of the FBI Hostage Rescue Team and Delta Force. From the outside, both can look like elite units entering with force. Mission and legal authority separate them. Police seek rescue, preservation of evidence and arrest. A military unit is organised to defeat resistance and complete an objective. Equipment may resemble; success criteria change behaviour.

For institutional analysis, “what can the unit do?” is not enough. Ask who authorises action, what must survive it, which failure is tolerable and which result produces prestige or professional shame. That is where the decision model resides.

In the talk, the distinction widens when he turns to Northern Ireland. Clancy argues that London sent the army into a public-order crisis after local institutions had lost legitimacy, without a British equivalent of US federal intervention in civil-rights cases. His account is polemical and compresses a much more complicated history. The analytical question remains excellent: what happens when an institution is sent to solve a problem for which its authority, training and definition of success are badly matched? Sometimes the more powerful capability worsens the problem precisely because it optimises a different outcome.

Clancy was drawn to problems where direct information was absent but constraints remained visible. In the NSA talk he used, among other examples, a low-observable American aircraft and began with the dangerous mission of suppressing air defences. A platform expected to enter a space saturated with radars and missiles cannot be designed arbitrarily. The mission demands properties; those properties demand infrastructure, training and trade-offs.

In the questions after the talk, Clancy says everything he knew about the stealthDesign that reduces a platform’s detectability by sensors without making it literally invisible. aircraft came from four press paragraphs: two in a Baltimore newspaper and one each in two Air Force Magazine articles. The report that the aircraft could be carried in a C-5 from Burbank to Nellis suggested a fighter-sized machine.

The next step is mission. Why build a small, difficult-to-detect aircraft capable of going where other aircraft face excessive risk? Clancy selects the Wild WeaselA mission to detect and suppress enemy air defences, especially surface-to-air radars and missiles. mission, hunting and suppressing surface-to-air radars and missiles. From that he constructs the “Frisbees of Dreamland” in Red Storm Rising. A consultant to an aerospace company asks for the chapter and makes dozens of copies. Clancy treats the reaction as evidence that the hypothesis is interesting, not as confirmation of the actual design.

The question “what must be true for the observed thing to work?” is one of the most productive forms of analysis. A long-range system needs sensors, data links or external targeting. An operation executed very quickly implies delegated authority or a sequence prepared in advance. A unit that remains in the field for months needs a solution for maintenance, rotation, supplies and evacuation.

Reasoning backwards also produces indicators. If the hypothesis is right, facilities, contracts, specialisations, exercises, logistical patterns or doctrinal changes should appear. The conclusion becomes testable: the story that sounds good acquires an obligation to predict the traces that should be found.

The method is powerful because missions narrow the possibility space. It is risky because several solutions may satisfy the same constraint. Every “must” requires an alternative: what other capability, doctrine or mission would create the same public clues?

In the questions, he applies the same logic to Warsaw Pact forces. A division can exist in the order of battleThe list of units and capabilities declared to exist in a force. It does not prove that they are politically or operationally available for a mission. and still be politically unavailable for the mission being imagined. Clancy asks how much trust the Soviets could place in allies that had rebelled or been invaded by Moscow. His tone is sweeping, but the variable is real: nominal force must be discounted for political reliability, cohesion, command authority and willingness to bear the cost. Inventory is not the same thing as usable power.

A technical sheet can be impeccable and almost useless to a reader. Range: 400 kilometres. Speed: Mach 3. Latency: 30 seconds. Without a situation, figures do not say what becomes possible, which window disappears or who is forced to decide with incomplete information.

Clancy introduced technology at the moment a character needed it. A property of sonar appeared when a contact had to be detected. Geometry became relevant when a submarine could obtain a better solution only by closing the distance and exposing itself. The reader learned the mechanism because the mechanism had become the obstacle between actor and objective.

Range, speed, frequency and warhead weight are properties. They become analysis when they alter a situation: warning falls from eight minutes to ninety seconds; a commander must turn on a radar and reveal position; ammunition permits only one attempt; a sensor sees farther only in a particular environment.

In the Harpoon manual, Clancy encountered the full chain: environment, detection, classification, decision, engagement, effect, response. An electronic-support measureElectronic-support measures: detecting and analysing the electromagnetic emissions of other systems, without necessarily emitting a signal of one’s own. was not merely a box of specifications; it could detect an emission and turn the searching adversary into a target. A convergence zone was not an oceanography lesson; it changed where a vessel could be heard.

Technical explanation therefore works best at the instant an actor needs it. The reader learns sonar when a contact must be classified before disappearing. It learns latency when the picture arrives after the engagement window closes. It learns production when losses exceed replacement capacity.

The writing unit is actor, objective, obstacle, constraint, choice and consequence. If a detail changes none of these, and does not change confidence in the judgement, it may remain in the annex. Clarity does not require emptying the system. It requires choosing the mechanism that carries the effect.

The premise of The Hunt for Red October grew from a real mechanism: the revolt aboard Storozhevoy. Clancy moved the mechanism to a strategic submarine and changed the stakes. The result tested the consequences of a known form of insubordination inside a system of much greater military and political value.

The analytical operation can be written precisely: identify the real mechanism, change one or two structural variables and propagate the consequences. Then count the assumptions, consult the base rate, build the simpler alternative and state the indicators that would strengthen or weaken the scenario.

Larry Bond described his game Convoy as a way to observe variables that return across repeated runs. The game did not prove the outcome of a confrontation. Repetition exposed warning, logistics, sensors or escort tempo when they continued to control decisions.

When Bond comes to Clancy’s son’s birthday and says he has been playing with an idea, the conversation becomes Red Storm Rising. They exchange premises, play situations and follow operational rhythm. A wargameA structured simulation of conflict in which participants make decisions under defined conditions and rules. It is useful for relationships and dependencies, not as prophecy. does not produce the future. It produces conditional consequences: if warning is short, if a sensor fails, if escorts are insufficient, then options change in this way.

Repetition exposes structure. Vary readiness, weather, ammunition, command latencyThe delay between information being produced, transmitted and becoming usable in a decision., rules of engagement and adversary competence. Do not merely collect results. Record which dependency dominates in several configurations. This is sensitivity analysis performed through human decisions.

Clancy was useful as the adversary because he did not respect an object’s official role. A weapon might become a decoy; an emission might create a false picture; a platform might be sacrificed for information. Red-teamingAn exercise in which a person or team thinks and acts from an adversary’s perspective to test assumptions, plans and vulnerabilities. fails when the adversary is one’s own organisation wearing different insignia.

The game retains a design trap. A commercial product must remain playable. It may balance forces, compress geography, simplify logistics and give both sides comparable definitions of victory. Reality owes no such balance. Before interpreting results, the analyst must identify what was introduced for play and what claims to represent the world.

After Naval Institute sends the manuscript to submariners, one reviewer recommends that it not be published for security reasons. Clancy invites him to the insurance office and shows his sources. The concern centres on “Crazy Ivan”. Clancy explains that he heard the expression from a client working at the nuclear plant.

Eventually the author offers the obvious solution: tell me what must come out, and it will come out. The reviewer replies that he cannot say, because the information is classified. The anecdote is funny, and the problem is real. A precise correction may confirm that an outsider has inferred something important. A visible reaction may point to the passage. Silence may leave a dangerous detail intact.

Clancy adds an observation he developed after visiting the CIA: “there is no such thing as a classified question; there can be a classified answer”. Beyond the logical joke, the sentence separates curiosity from access. A journalist or analyst may ask anything. The holder of information must know where the answer stops.

Mature review distinguishes four stages: information already public, an unconfirmed public inference, official confirmation and additional operational detail. “It is online” does not make them equivalent. Nor does the label “sensitive” explain the harm by itself.

Later, asked directly, Clancy says he did not need classified information and that, if someone gave it to him, he would call the FBI or destroy it. The statement matters beyond personal ethics. Receiving a secret adds more than knowledge; it adds obligations, distribution limits and the risk of compromising a source. Open research allowed him to show the route to his conclusion and keep the product public. For an analyst, the prestige of classified material is irrelevant. What matters is how much it changes the judgement and whether that difference justifies the cost of using it.

Clancy says he read an article about a British security officer and vulnerabilities at Cheltenham. One paragraph mentioned a joint Anglo-American programme for tracking Soviet ballistic-missile submarines through their daily radio transmissions to base. The article then moved on.

To an ordinary reader, the detail may have appeared technical and inert. To Clancy, it overturned an operational assumption: American strategic submarines tried to disappear inside enormous areas, while a daily radio routine would create a collection and location opportunity. He calls an acquaintance in the field, repeats the sentence and notices the reaction. The question about date and page tells him that the passage mattered.

The example shows why public access is not synonymous with absence of sensitivity. Information has literal and relational value. A modest fact can become important when connected to doctrine, collection capabilities and adversary behaviour. The reporter saw the sentence; he did not see the model into which it fitted.

For the intelligence practitioner, this is also a communication obligation. A reader without context cannot protect, challenge or use information whose consequence remains invisible. Explanation must reveal the chain without publishing the details that make exploitation easy.

“If something is so important you cannot use it, it is not intelligence; it is data in somebody’s file.” Clancy’s formulation is brutal and incomplete in exactly the right way to open the problem. A lost source can cost lives and years of access. Perfectly protected information can leave the decision-maker choosing in darkness.

The professional answer is control of resolution and distribution. A product can separate source and sensitive detail from the judgement that must be used. A tear lineA division in a classified product: sensitive sources and details remain above it, while a sanitised judgement below can be disseminated more widely. allows a sanitised portion to circulate more widely. Other tools include generic source description, oral briefing, annexes restricted to authorised groups, delayed release and masking operational thresholds.

Clancy connects the problem to deterrenceInfluencing an adversary’s choice by making it believe an action will fail or impose unacceptable costs.. If the adversary does not know a capability exists or does not believe it will be used, that capability cannot alter the choice. Completely secret nuclear weapons do not deter. A wholly invisible attribution system does not warn. The state must show enough to create belief and conceal enough to retain advantage.

This is an information-design problem, not a binary choice between secret and public. The useful question is: what must the consumer or adversary believe, what evidence is required for credibility and which details would permit adaptation?

Discussing terrorism and Patriot Games, Clancy tells the NSA audience that intervention teams may be perfectly trained and still have nowhere to go. Somebody must identify the people, place and moment. To him, the intelligence community is the first line of defence because it enables the other lines to act.

The point becomes concrete through a recent encounter. A senior Royal Ulster Constabulary officer had told him that he was shot with an American M60 machine gun stolen from a National Guard armoury. For the unit in the United States, the incident might begin as lost inventory. At the end of the chain it had become terrorist capability in another country. Analysis has to follow the entire route: theft, brokerage, transport, concealment, use: because the strategic effect may appear far from the organisation that missed the first link.

He then invokes the Battle of Midway, where American cryptanalysis helped anticipate the Japanese operation, and Heinz Guderian, who connected armoured formations by radio. The comparison does not claim that information wins wars alone. It says an organisation without an observation–command–action circuit cannot turn resources into coordinated effect.

For an analyst, a product does not end when the sentence is correct. It must reach the right person while an option remains open and in a form that can be integrated with other flows. Truth delivered after the window closes is history. Truth delivered at a classification inaccessible to the consumer is inventory.

In the questions, his imagination moves to a sunken Soviet submarine. What would he recover first? His first choice is the cryptographic equipment. The reactor technology was already understood; the cipher machine and material might open the adversary’s communications. It is a small exercise in prioritising by information advantage. The object’s value is measured by the important uncertainty it can remove, rather than by how spectacular it looks.

This explains why writing interested the CIA. Translating complexity was not ornament. It was the final link in a chain beginning with observation and ending in decision.

The mind uses fluency as a clue. When a story is easy to simulate, it feels familiar; familiarity is mistaken for frequency. The analyst loses the base rateThe ordinary frequency of a class of events before the specific clues in the present case are considered. and begins to judge how well the scenes connect rather than how often the mechanism occurs in the world.

Source fusion can slide into apopheniaThe tendency to perceive a meaningful pattern in fragments that may be independent, random or connected differently from the proposed model.. Every fragment is genuine, while the link between them exists only in the author’s imagination. Correct detail about a torpedo, radar or procedure transfers authority to a strategic judgement that has not received the same testing. The expert who corrects a scene becomes the representative of an entire institution.

Storozhevoy is the perfect warning. The story Clancy tells at NSA has causality, geography and suspense. It helped produce one of the era’s most effective novels. Sablin’s intention was understood more fully only later. A narrative can be excellent as a thinking instrument and outdated as a factual account.

After drafting, the charm must be dismantled. Mark the epistemic statusThe status of a claim in relation to knowledge: observed, reported, corroborated, inferred, assumed or modelled. of every load-bearing claim. Count assumptions. Search for the simpler explanation. Write what would falsify the scenario and which indicators should appear if it becomes real. Keep story to make the chain intelligible; keep alternatives so story cannot close the file.

Used this way, Clancy is not a model of prediction. He is a teacher of curiosity, causal modelling and communication. His imagination produces hypotheses; analytical discipline then takes them one by one, strips away their rhythm and forces them to answer to evidence.