THE CHRONICLE

Each study is a moment: December 1961, October 1962, October 1973, November 1983, September 1980, the present. Between the moments is the history the studies assume: how the stockpiles grew and fell, how the forces moved from bombers on runways to missiles in silos to boats at sea, what the plans were for (massive retaliation, assured destruction, flexible response, counterforce, launch under attack, damage limitation), which treaties bounded what, and where the weapons stood in relation to the crises of the day (Thors in Britain, Jupiters in Turkey and Italy, R-12s in Cuba, SS-20s and Pershing IIs in Europe, the boats in the bastions). The studies show the machinery at an instant. The chronicle would show the machinery changing, and why.

The manifesto's tension runs through this history more plainly than through any single plan: the stockpile curve is the calmest chart in the world, and it counts sixty thousand warheads at its peak.

• Visual first. The chart or the map state is the chapter; the prose is a caption, never an essay. A chapter that needs more than a screen of words has not found its picture.
• Sources and methods, one click away and never in the way. Every chart and map state carries a disclosure, collapsed by default, that names the series, the document and the method with page or figure references, and links out to the source and to the validation notes. The reader who wants the scholarship finds all of it; the reader who does not is never made to read it. The lab's Sources and methods panel is the pattern.
• The register holds. Calm, terminal type, the evidence grammar on every mark. The chronicle is a strange work of scholarship and looks like one.

A third form beside the studies and the labs, reached from the front page: chapters that read as text with the map as their spine. Each chapter is a few hundred words, one or two charts drawn from a cited series, and one or more map states, which are the engine's own entities at a date, opened in the same view the studies use and scrubbed across time where the data allows. The reader moves between the chapter and the map; the map is never illustration, it is the argument's evidence.

The map states already exist for five epochs, because the studies built their orders of battle by script: 1961 (SIOP//62), 1962 (Cuba), 1973 (DEFCON 3), 1983 (Able Archer, theatre only), 2024 (seventy-two minutes). A posture atlas that scrubs a date across those epochs, sites appearing and disappearing with their strengths, is the first thing the chronicle can show and needs no new data beyond the gaps between them. Filling the gaps (1950s bombers, the 1990s drawdown) is the chronicle's own work, epoch by epoch, each with its Nuclear Notebook or Databook citation.

1. The stockpiles, 1945 to the present. The Kristensen and Korda inventory series for every state, as a chart, with the moments of the studies marked on it. The chart is the chronicle's cover: the calm line and its peak.
2. Posture. From bombers on ground alert and airborne alert to the triad, the alert rates, the bastions; the fraction of the force that could fire within fifteen minutes at each epoch, which the readiness lab already computes for 1961. Launch on warning and launch under attack as the doctrine the present scenario rests on.
3. Doctrine as arithmetic. Massive retaliation, assured destruction's "20 to 25 per cent of population and 50 per cent of industry" (McNamara, 1965), flexible response, NUWEP-74's categories, the countervailing strategy of PD-59, the present. Each named with its document and shown as the target rule it implies, which the studies already enact (SIOP-62's task order, NUWEP-74's categories, OPLAN 8010 as a rule).
4. Geopolitics as basing. The forward bases and what they did to the crises: Thor, Jupiter, the Cuban regiments, the Euromissiles, the Pacific boats. Map states with the crisis studies opened from them.
5. Treaties as bounds. LTBT, NPT, SALT I and the ABM treaty, SALT II, INF, START, SORT, New START and its expiry in February 2026; each as a line on the stockpile chart (built as a toggle on chapter 1, 10 September 2026) and, still to do, a change on the map (the INF zero, the START drawdown).
6. Missile defence and the shot exchange. Nike-Zeus, Sentinel, Safeguard's one site at Grand Forks (1975, closed within months), the ABM treaty, Moscow's A-35 and A-135, SDI, Brilliant Pebbles, GMD's forty-four, and the present decade's Golden Dome. The chapter's argument is arithmetic and belongs in a lab (below): the defender must buy shots faster than the attacker buys warheads and decoys, and no system since 1957 has done so.

The shot-exchange lab (built 10 September 2026 at #/lab/defence; model in src/models/defence.ts). An attacker with N warheads, D decoys per warhead and a MIRV multiplier against a defender with M interceptors of single-shot kill probability p fired in salvos of k. Outputs: leakage (warheads through) as a function of M, the cost-exchange ratio at stated unit costs, and the point at which the attacker's cheapest countermeasure beats the defender's dearest interceptor. Reference cases from the published record: Safeguard's 100 interceptors against the 1975 Soviet force; the APS 1987 directed-energy study's numbers for SDI; the UCS 2000 countermeasures report; the NAS 2012 assessment of GMD; the Congressional Budget Office's 2025 estimate for a space-based interceptor layer. The lab runs against the seventy-two-minute act: what forty-four interceptors at the test record do to one missile, and to fifty. This is the "folly" made visible as a curve, in the same calm as everything else.

The yield-and-accuracy lab (built 10 September 2026 at #/lab/accuracy; model in src/models/lethality.ts). The user's point: counterforce against countervalue was a technically mediated choice before it was a doctrinal one. A city fails at 5 psi and a silo at 2,000; the radius for the first is kilometres and for the second hundreds of metres; with a CEP of miles the city is a target and the silo is not. The lab plots single-shot kill against a hardened silo by year of service for eighteen systems from Atlas D to Trident II, with the even-chance line, a matrix of every system against every target hardness, and editable yield, CEP, hardness, shots and reliability. The crossing comes with Minuteman III's guidance refit, Peacekeeper and Trident II, by accuracy rather than yield, and the MIRV bus is what made the accurate warhead cheap enough to spend on silos. That is the shift from the large unitary warhead to the accurate MIRV, and it is the chapter on doctrine as arithmetic in one chart.

The guidance error-budget lab (built 10 September 2026 at #/lab/guidance; model in src/models/guidance.ts; the text is MacKenzie's Inventing Accuracy, which the user pointed at on archive.org, a lending copy whose pages could not be read for the build, so the term magnitudes are illustrations that reproduce the published CEPs and are marked as such for correction from the book). The companion to the lethality lab: where the CEP comes from. A missile's miss distance is the root sum of squares of independent error sources, and for a submarine missile the largest of them in 1960 was not the missile's at all but the boat's: its own position and velocity as its inertial navigator carried them since the last fix, and the gravity it had not measured. A lab that composes a CEP from its terms (initial position and heading, velocity at launch, gyro drift over the flight, accelerometer bias, the gravity model, reentry dispersion), lets the reader scrub the time since the last fix and watch the boat's contribution grow, and then adds the star sight of Trident to collapse the initial-condition terms, would show the problem of unknowns and cumulative error the way the Polaris guidance history tells it. The figures need a source: MacKenzie's Inventing Accuracy gives the structure and some magnitudes, and the text the user has in mind should be named before the lab is built so its numbers, not illustrations, drive it.

The stockpile chart (built 10 September 2026 at #/chronicle, chapter 1). The Federation of American Scientists' inventory series as republished by Our World in Data under CC BY, all nine states, linear and log, the studies marked on their years, the world's peak labelled (64,452 in 1986), a hover readout by year, and the sources collapsed beneath. The first chapter of the chronicle and the pattern for the rest.

The posture atlas (built 10 September 2026 at #/chronicle/posture, chapter 2). scripts/build-posture-epochs.py normalises the five orders of battle into one shape, a site with a side, a kind, a position and the weapons its study counted by that study's rules, keeping each site's tier and source. The atlas draws them as discs whose area is the weapons, scrubbed by a slider over 1961, 1962, 1973, 1983 and 2024, with totals by side and kind and the selected site's provenance. The theatre epochs say they are not comparable with the strategic ones. Radars draw as a glyph, and a click lays the radar's field of view on the map as a light fan, the faces' centre bearing and width from the public descriptions and the range of the class, marked inferred; a button shows every fan of the epoch. BMEWS stands on the 1961, 1967, 1973 and 1991 epochs, PAVE PAWS and the Soviet Dnestr and Daryal sites on 1973 and 1991, and the 2024 order of battle's radars on 2024. The 1967 epoch, the peak of the American stockpile, was added too. The 1991 epoch, the start of the drawdown, was added the same day from the Nuclear Notebook's end-of-year tables by scripts/build-order-of-battle-1991.py, with no study behind it and every site inferred or reconstructed; the atlas says so. The 1950s and the years between remain to fill.

#/chronicle carries the six chapters on one page in the agreed form: chapter 1 the stockpile curve with the studies and the treaties as lines; chapter 2 the weapons by side at the strategic epochs as bars, opening the posture atlas at #/chronicle/posture; chapter 3 the counterforce crossing chart from the accuracy lab and a table of doctrines as the target rules they imply, each linked to the study that enacts it; chapter 4 the forward deployments as a table linked to the crisis studies; chapter 5 the treaties; chapter 6 missile defence from Nike-Zeus to the space layer, linked to the defence lab and the salvo act. One caption per chapter; sources collapsed beneath. The nav bar was cut to the forms and the studies, with the labs shown as a group only while one is open and otherwise reached from the front page.

The shot-exchange lab first: self-contained, a week of scholarship already in the record, tied to the seventy-two-minute act, and the clearest demonstration the chronicle has of the manifesto's tension. Then the stockpile chart and the posture atlas over the existing epochs, which together are the chronicle's spine and need no new data. The chapters follow as the map states and charts exist to hang them on.

• Whether chapters live in the app or as briefs in docs/ with the app holding only the map states and charts. The app, probably: the front page already reads as a page, and the reader should not leave the instrument.
• How much text the register can bear. The studies say almost nothing in prose; the chronicle must say more, and it must keep the calm.
• The 1990s: the drawdown is the part of the history the studies never touch. The 1991 epoch now marks its start; the curve down to 2024 is still a jump on the atlas and a line only on the stockpile chart.

A table of eleven turns from Carte Blanche in 1955 to the presidential nuclear initiatives of 1991, each linking to the study or lab that draws it, under a head that states the argument: NATO could not match the conventional force against it and so promised nuclear use early, every strategic argument of the Cold War follows from that promise, and the documents describe a slope rather than a firebreak. Missile defence becomes chapter 7. The chapter's own brief is THEATRE.md.