ABLE ARCHER//83

SIOP//62 enacts a plan that was written and never executed. Able Archer 83 is the opposite object: an exercise that was executed and never became a plan, during which the other side prepared for the real thing. The record is unusually good on both sides and unusually candid about its own uncertainty, which makes it the right second study for an engine whose visual language is built on evidence tiers.

The exercise ran from 7 to 11 November 1983 in a scenario in which Warsaw Pact forces had opened hostilities across Allied Command Europe on 4 November and used chemical weapons by the 6th. NATO simulated escalation to DEFCON 1, requested nuclear release on the evening of 8 November, employed weapons on the morning of the 9th, and executed a follow-on nuclear strike on the morning of the 11th. It did so with new features: a 170-flight radio-silent airlift of 19,000 soldiers, a move to the alternate war headquarters, new release procedures with consultations in Washington and London, and "slips of the tongue" calling B-52 sorties "strikes". On the other side, Soviet air units in East Germany and Poland went to heightened readiness, fighter-bombers of the Group of Soviet Forces in Germany to thirty-minute alert, and the Soviet 4th Air Army began "preparations for immediate use of nuclear weapons". The President's Foreign Intelligence Advisory Board concluded in 1990 that "at least some Soviet forces were preparing to preempt or counterattack a NATO strike launched under cover of Able Archer", that the United States "may have inadvertently placed our relations with the Soviet Union on a hair trigger", and that the one American general who declined to answer alert with alert "acted correctly out of instinct, not informed guidance".

The study enacts the branch that instinct avoided: Soviet frontal aviation and theatre missiles strike first, on the last morning of the exercise, against NATO's nuclear delivery means and the bases they fly from, and NATO's surviving theatre forces answer. It is a European war of hours, not the SIOP's fifteen, and its population grid is 1983.

Everything on the readout carries a tier, as in SIOP//62, and the withheld tier will be busier here than there.

Primary and archival sources

The timeline as the documents give it

The record disagrees with itself in the open. The 1984 estimate and the 1990 board reach opposite conclusions from the same indicators. The study keeps both on the readout, as SIOP//62 keeps its 14 and 16 options.

The branch, in the register of the evidence tiers:

1. The exercise as documented, played on the clock from 4 November: the scenario's fictional war as a ghost layer, the real airlift and headquarters move as documented movement, the release simulation on the 8th, 9th, 10th and 11th as documented events with no geometry, because nothing left the ground.
2. The Soviet readiness as documented: fighter-bomber regiments in East Germany and Poland at thirty-minute alert, the 4th Air Army loading, drawn at reconstructed bases with the units the record names and the units it does not badged accordingly.
3. The inferred order: on the morning of 11 November, when the exercise executes its follow-on strike, the Soviet command reads it as real. Frontal aviation and the theatre missiles of the Group of Soviet Forces in Germany strike NATO's nuclear delivery means: the quick-reaction-alert airfields, the Pershing and cruise-missile sites then arriving in Britain and West Germany, the nuclear artillery and Lance units, and the headquarters. The targeting is inferred; the readout says so on every mark.
4. NATO's answer from what survives, under the release procedures the exercise had just rehearsed, against the Soviet forward airfields and missile brigades. Also inferred.
5. The outcome over HYDE 1983 for West Germany, East Germany, Poland, the Low Countries and Britain, by the same two-number method, with the same blast-only and fire bounds, and the plume lab under November winds for a theatre whose fallout falls on the people fighting in it.

The finding the study is built to reach is the PFIAB's own: the difference between the branch and the record is one order, and one general's refusal to give the matching one.

#/study/able-archer-83 plays the week and the branch on one clock of hours from the morning of 11 November. The order of battle is built by scripts/build-order-of-battle-1983.py: fifteen NATO strike wings and squadrons, the three Pershing 1a battalions, Greenham Common drawn withheld with its sixteen missiles three days away, eight headquarters and ports; on the other side nine Soviet regiments in East Germany and Poland, four missile brigades, five SS-20 garrisons standing for the western share of 405 launchers, and two headquarters. The exercise's ladder is data/able-archer/exercise-1983.json, ten steps from the scenario's fictional 4 November to the record's verdicts, each with a source; a dozen of the 172 radio-silent airlift missions are drawn from the east-coast ports to Rhein-Main.

The sums are over GHSL's 30-arc-second grid for 1985, which resolves the towns beside the airfields that HYDE's five-arc-minute cells smeared; the first run on HYDE 1983 read 500,000 and 1.2 million on NATO territory and 140,000 and 300,000 on the Pact's. NATO's answer grew when the generated aircraft came to carry two weapons each as one sortie, and when the SS-20 regiments were given three warheads to a missile.

The finding. The first twenty minutes decide the answer. The Scud brigades fire at H+15 and land four minutes later; the SS-20s fire at H+10 and land at H+20; NATO's fifteen-minute alert forces leave at H+20 on warning of the first launches. Every Pershing garrison and most airfields are struck before that, so the answer is a dozen weapons from the two or three bases the first wave did not reach, and the generated force, four hours later, never leaves. The Warsaw Pact's own strike is small by the list, not the force: 71 weapons cover NATO's nuclear delivery means and headquarters three deep, and 868 stay unassigned because the rule gives them no target. The Soviet regiments and garrisons are reconstructed from the divisions the unit histories name, and the readout says so on each.

• Which units, exactly, did the intelligence record see readying? The public documents name the 4th Air Army and GSFG fighter-bombers; the regiment-level picture is in the still-redacted material and in Soviet sources not yet read.
• What did the Soviet General Staff's own plan for a preemptive theatre strike look like in 1983? The study can use published Warsaw Pact exercise material (the Seven Days to the River Rhine plan of 1979 is the nearest published analogue) as a reconstructed template, labelled as such.
• How far should the study follow escalation past the theatre? The brief stops at the theatre exchange. Anything beyond it is the SIOP's successor, SIOP-6, and nothing of it is public.
• Register. This is the study in which the people preparing to strike were afraid, and the readout should carry that from RYaN's indicator lists onward: churches, banks and blood banks watched for a war that was not coming.

• National Security Archive, Able Archer 83 Sourcebook and EBBs 426, 427, 428, 2021-02-17 and 2025-11-14, linked above.
• President's Foreign Intelligence Advisory Board, The Soviet "War Scare", 15 February 1990, released 2015.
• Jones, Nate. Able Archer 83: The Secret History of the NATO Exercise That Almost Triggered Nuclear War. New Press, 2016.
• Foreign Relations of the United States, 1981–1988, Volume IV, Soviet Union, January 1983–March 1985.