They Both Had Radar

▶️ Prefer to watch? Every story’s a short video → https://www.youtube.com/@GetSafetyIntel · 📣 Got a near-miss? Share it anonymously (QR + link at the bottom).

Two of the most modern ocean liners afloat, both with radar on the bridge, both with officers watching the screen — and on a foggy night off Nantucket they steered straight into each other. The technology didn’t fail that night. The people reading it did. And that’s a lesson with nothing to do with ships.

Late on the night of July 25, 1956, the Italian liner SS Andrea Doria — a showpiece of postwar engineering, radar-equipped, built with watertight compartments that were supposed to keep her afloat no matter what — was inbound to New York through heavy fog south of Nantucket. Coming the other way was the Swedish liner MS Stockholm, ice-strengthened bow, also on radar. Each ship saw the other on its scope from miles out. Then, over about fifteen minutes, the officers on both bridges misread what the radar was telling them — each convinced they’d pass the other safely — and maneuvered. Wrongly. The Stockholm turned to pass port-to-port; the Andrea Doria turned to pass starboard. They turned into each other.

At 11:10 p.m. the Stockholm’s reinforced bow drove about 30 feet into the Andrea Doria’s starboard side, straight into the passenger cabins. That impact killed most of the 46 who died on the Andrea Doria; five more died aboard the Stockholm — crew, in the crushed bow. The Doria took on a hard list within minutes, and that list jammed half her lifeboats uselessly against the hull. It was close to a miracle that the sea was calm and other ships reached her fast: 1,660 people were pulled off before, at 10:09 the next morning — July 26 — she rolled over and sank.

The investigators’ verdict became a phrase mariners still use: a “radar-assisted collision.” Both crews had the information. Both had the very tool that was supposed to make fog irrelevant. What they didn’t have was a correct reading of it — and the new technology had bred a confidence that let them keep their speed up in near-zero visibility. The tighter standards that followed, for how you use radar, plot and verify a contact’s course, and slow down in fog, exist because two competent crews trusted a screen they had misunderstood.

Here’s why it belongs in a brief that usually lives on 1910 and 1926: swap “radar” for whatever screen you’ve been told keeps your workplace safe. The four-gas monitor. The machine light curtain. The backup camera. The LEL sensor. The AI camera that’s supposed to catch the guy in the red zone. Not one of them keeps anybody alive on its own — they keep people alive only if the human system around them reads them right and acts. A gas monitor in alarm that everyone has quietly decided is “probably faulty” is the Andrea Doria’s radar. On the floor, that’s the entire logic behind atmospheric testing under 29 CFR 1910.146 for permit-required confined spaces: the rule doesn’t just say “have a meter,” it says test it, interpret it, and act on it before anyone goes in. The meter is not the safeguard. The decision you make when it reads is.

(S) The safety toll was 51 dead and a great ship on the bottom, though extraordinary seamanship and a calm sea saved 1,660. (H) Survivors carried the night for a lifetime, and the health toll didn’t stop in 1956: the wreck lies about 250 feet down and has killed or injured technical divers for decades — a hazard that keeps collecting. (E) On the environmental side I’ll be straight with you: this wasn’t a headline spill, but the Andrea Doria went down with her fuel oil and everything aboard, and for nearly seventy years that deteriorating wreck off Nantucket has been a slow, watched pollution source on the seabed. The same misread screen that killed 51 people also put a rusting hull and its oil into the Atlantic — S, H and E out of one wrong reading.

The ground is still the deadliest thing on a jobsite. Cave-ins kill roughly 23 workers a year in the U.S., and this summer drove it home: southeast Michigan had three trench collapses inside about 48 hours around June 30 to July 1, one of them fatal — a 23-year-old installing underground tanks in Handy Township. OSHA’s trenching National Emphasis Program is still running hot. So what: 29 CFR 1926.652 is not subtle — any trench 5 feet or deeper needs a protective system (slope it back, shore it, or set a trench box), a competent person has to inspect it, and there has to be a safe way out within 25 feet of every worker. A trench doesn’t give a second warning; the wall that’s holding is the wall that buries you.

EPA is easing off the chemical-accident rule — but not yet. On February 24, 2026, EPA published a proposed rule to roll back big pieces of the 2024 Safer Communities by Chemical Accident Prevention (SCCAP) update to the Risk Management Program — including the Safer Technology and Alternatives Analysis requirement — framing it as realignment with OSHA’s PSM standard; the comment period closed in April. So what for EHS leaders: a proposal is not a repeal. The current SCCAP compliance dates (May 2027) still stand until a final rule lands, so if you run an RMP-covered process, keep building the program — don’t stand down on a headline.

A confined-space crew was prepping to enter a below-grade vault. The four-gas meter alarmed on low oxygen right at the opening. The attendant’s read was honestly a reasonable one given his week: that meter had thrown two “bad” alarms in the past few days and both times the space checked out fine, so the working theory was a flaky sensor. They were one decision away from bump-testing it and calling it good enough to go in. What stopped them was the entry supervisor, who treated the alarm as true until proven otherwise, pulled a second calibrated meter, and got the same low-oxygen reading. The vault really was oxygen-deficient. Nobody went in. Nobody got hurt.

Blameless read, because it matters: the crew wasn’t reckless — the system had taught them to distrust that meter with two false alarms, the same way radar taught two ships’ crews that fog was handled. Alarm fatigue is a system failure, not a character flaw. The fix isn’t “trust the meter more.” It’s fix the meters that cry wolf, and make “the alarm is true until a second source says otherwise” the standing rule — so the one real alarm in a hundred still stops the job.

Got a fail or a near-miss? Hit reply, or scan the QR below — we’ll feature it anonymously — no names, no company, no blame. Just the lesson, so the next crew doesn’t learn it the hard way.

Pick one “screen” your crew has quietly started trusting to keep them safe — a gas monitor, an interlock, a camera, a proximity alarm — and ask two questions out loud today: when it alarms, what exactly do we do, and has anyone started treating its alarms as noise? If the honest answer is that a real alarm would get waved off as “probably the sensor,” you’ve found your Andrea Doria. Fix the false alarms, and make the real one impossible to ignore.

🎬 Rather watch than read? Every Safety Intel story is a short video on YouTube → https://www.youtube.com/@GetSafetyIntel — subscribe so you never miss one. 📸 Share YOUR near-miss (100% anonymous): https://scantoconfess.com — or just scan the QR above. 📤 Forward this to a crew member who needs it today.

Please stay Safe & Hydrated!!!