A Funny Feeling in My Gut

Anyone who has seen the film WarGames remembers the big screen at the end, with missile tracks crawling over a map and a computer insisting that the attack is real. The Soviet Union had a room like that. Just after midnight on 26 September 1983, at a secret bunker called Serpukhov-15 south of Moscow, a red panel in it lit up with one word: "Start".

The man in the commander's chair that night was Lieutenant Colonel Stanislav Petrov, aged 44. The satellites he was watching had just reported that a Minuteman missile had left a base in the United States.

The system he was watching

The Soviet space-based warning network was called Oko, which means eye. It had been in development since the early 1970s and was formally accepted into service in December 1982, nine months before that night. By September 1983 it had seven satellites in long, looping orbits, placed so they could look down at American missile fields and catch the heat of a rocket leaving its silo.

The data came down to Serpukhov-15 and was processed by computer, which decided on its own whether a launch had happened. The officers could also look at the raw picture the sensors were sending. Petrov's job, as the nuclear historian Pavel Podvig describes the chain, was to assess what the computer said and pass it up to the command post of the early-warning army, which passed it to air defence headquarters, which passed it to the General Staff.

Petrov had worked on the satellite system since its beginnings. He told the Washington Post in 1999 that it had been rushed into service and was "raw".

Five missiles

The first alarm reported one launch. Then another, and another. "A minute later the siren went off again," he told the BBC's Russian Service in 2013. "The second missile was launched. Then the third, and the fourth, and the fifth. Computers changed their alerts from 'launch' to 'missile strike'." The system rated the reliability of the warning at its highest level.

"The siren howled, but I just sat there for a few seconds, staring at the big, back-lit, red screen with the word 'launch' on it," he said.

He had a phone in one hand and an intercom in the other, he told the Post, and another officer was shouting down the line at him to stay calm and do his job. Because the reports had come in so fast, an alert had already gone to the General Staff automatically, before he had judged anything. In less than five minutes he made his call. He reported a system malfunction.

"I had a funny feeling in my gut," he said. "I didn't want to make a mistake. I made a decision, and that was it."

Why he didn't believe it

Part of it was a guess, he admitted. Everything he had been taught said an American first strike would be massive, hundreds of missiles at once, built to overwhelm the Soviet Union before it could answer. "When people start a war, they don't start it with only five missiles," he remembered thinking.

Part of it was that the ground radars showed nothing. They couldn't yet. Radar sees only to the horizon, so it would not have picked up real missiles until some minutes after the satellites. Their silence wasn't proof, only a lack of confirmation, and a group of satellite operators looking at the visual feed told him they could see no missiles either.

And part of it was that the alert was too clean. "There were 28 or 29 security levels," he told the BBC. "After the target was identified, it had to pass all of those 'checkpoints'. I was not quite sure it was possible, under those circumstances."

Then he waited. "Twenty-three minutes later I realised that nothing had happened. If there had been a real strike, then I would already know about it. It was such a relief."

What the satellite saw

The cause turned out to be the sun. A satellite had picked up sunlight reflecting off the tops of high clouds over the United States, and the software meant to screen out such false signals hadn't. An investigation reporting to the defence minister concluded that the system had misprocessed the signal in difficult lighting conditions, and the program was rewritten.

Petrov did not come out of the investigation well. He was praised at first, then questioned hard. Why hadn't he written everything down that night? "Because I had a phone in one hand and the intercom in the other, and I don't have a third hand," he told them. He was reprimanded over the logbook, he said, not over his decision. There was no reward.

The year it happened in

What made the night frightening in hindsight was its timing. In 1981 the Soviet leadership, convinced the United States might be planning a surprise attack, had ordered the KGB and military intelligence to look for evidence of one worldwide, an effort called Operation RYAN that the CIA historian Benjamin Fischer later reconstructed in detail.

Three weeks before Petrov's shift, on 1 September 1983, a Soviet interceptor shot down Korean Air Lines Flight 007 after it strayed over Soviet territory, killing all 269 people aboard. Six weeks after it, NATO ran its annual command exercise, Able Archer 83, which that year included a practice run through a full simulated release of nuclear weapons. It was about the worst autumn of the 1980s in which to have an early-warning satellite report an American launch.

How close it came

Here the legend needs some care. Petrov was the officer who assessed the data, not the man with his finger on anything. The decision to retaliate belonged to the General Staff and, above it, the Soviet leader, Yuri Andropov.

Podvig, who has spent decades studying the Soviet nuclear forces, argues in a 2025 working paper that the danger has been somewhat exaggerated. Alarms from the satellite system reached the General Staff's central command post with some regularity, he notes, and there is no evidence that anyone there was preparing to treat this one as a real attack. Even if they had, the most they were likely to do in the minutes available was issue a preliminary command, a step short of launching anything. Petrov's report helped the right reading reach the top. It is likely, Podvig concludes, that the alarm would have been recognised as false anyway.

That doesn't make Petrov wrong or the night safe. It moves the lesson from one man to the design. A system designed to report launches automatically fed an unconfirmed alarm straight into the top of the command chain, during a war scare, and the only correction available in the first minutes was a human being deciding not to believe a machine he had been trained to trust. Petrov himself put the odds at 50-50 and said he was never completely sure.

He also thought it mattered who was on shift. He was the only officer on his team with a civilian education, he told the BBC. "My colleagues were all professional soldiers, they were taught to give and obey orders."

Afterwards

He kept quiet for ten years. "I thought it was shameful for the Soviet army that our system failed in this way," he said. The story surfaced after the Soviet Union collapsed, and he spent his last decades in a small flat in Fryazino, outside Moscow, collecting awards abroad that he never quite seemed to want. In February 2013 he went to Dresden to receive its peace prize at the opera house, the photograph at the top of this page.

He died in May 2017, and the outside world did not hear of it until September.

"That was my job," he told the BBC, of the night in the bunker. "But they were lucky it was me on shift that night."

The Cuban missile crisis is the Cold War's best-known near miss. Petrov's was quieter, and it came out of a computer, which is the kind more likely to happen again.

Sources

• David Hoffman, "'I Had A Funny Feeling in My Gut'", The Washington Post, 10 February 1999 (Serpukhov-15, the "Start" panel and the time just past midnight; Petrov's age and rank; one launch then five Minuteman missiles; the automatic alert to the General Staff; the decision in under five minutes; "a funny feeling in my gut" and "only five missiles"; the ground radars; the system being "raw"; the interrogation and the third hand; the sunlight off cloud tops and the rewritten software; no reward; KAL 007 and Able Archer as context).
• BBC News, "Stanislav Petrov: The man who may have saved the world", 26 September 2013 (the siren and the red screen; the second to fifth launches; reliability rated highest; the satellite radar operators seeing no missiles; the 28 or 29 security levels; the 23 minutes; the 50-50 odds; his civilian education and his colleagues; the logbook reprimand; ten years of silence; "they were lucky it was me on shift").
• Pavel Podvig, "Soviet and Russian false alarms and nuclear weapons incidents", working paper for the Managing the Atom project, 27 February 2025 (Oko's development and acceptance in December 1982; seven satellites in highly elliptical orbits; automatic processing and the visual feed; the reporting chain to the General Staff; regular alarms reaching the central command post; the investigative commission and the lighting conditions; no evidence the General Staff was prepared to act; the preliminary command; the alarm likely recognised as false anyway).
• Pavel Podvig, "Did Stanislav Petrov save the world in 1983? It's complicated", Russian Strategic Nuclear Forces, October 2022 (the argument that the danger of the incident has been exaggerated).
• Benjamin B. Fischer, "A Cold War Conundrum: The 1983 Soviet War Scare", CIA Center for the Study of Intelligence (Operation RYAN from 1981; the shooting down of KAL 007 on 1 September 1983 with 269 killed; Able Archer 83 and its simulated nuclear release).
• The New York Times, "Stanislav Petrov, Soviet Officer Who Helped Avert Nuclear War, Is Dead at 77", 18 September 2017 (his death in May 2017 and the news reaching the outside world only in September).
• Wikimedia Commons, "Stanislaw Petrow in der Kuche seiner Wohnung in Frjasino" (his home in Fryazino in 2016).