I posted this on Facebook a few years back. Thought I’d blogged it.

In 1981 a Soviet submarine ran aground in Sweden. The Swedish navy helpfully towed the errant vessel out to sea and set it on its way. Then the Swedish navy went a little nuts.

Convinced that the Russians were regularly running submarines in Swedish waters, the navy put hydrophones on buoys everywhere. And when a hydrophone detected the “typical sound” (of a Russian submarine) they sent helicopters out with hydrophones (and I assume magnetometers?) to locate the sub. When they located the “typical sound,” they’d drop a depth charge. And … nothing.

This went on for over a decade. They dropped countless bombs into the water and never once found an actual submarine.

In 1996 the Swedish military formed a group of scientists to investigate. They gathered everybody in a room and played for them the “typical sound” that the military was convinced was a Russian submarine. At least two of the scientists, experts in underwater sounds, were convinced that the sound was biological, not mechanical. But they didn’t know exactly what it was.

I’m a little sketchy on the details of how researchers came up with the answer. I do know that it involved first getting a dead herring and squeezing it, then testing a large one in a cage, and finally going out on a fishing boat with a hydrophone and recording a school of herring. Turns out that the top-secret classified “typical sound” that civilians and most members of the military were not allowed to hear was in fact millions of herring farts.

It turns out that the herring is a rather strange kind of fish. Most bony fish have what’s called a swim bladder: an organ that the fish can inflate or deflate to affect buoyancy. In most fish the swim bladder is essentially a closed system. But in the herring there is a canal from the swim bladder to the anus. Depressurization is essentially a fish fart. (Yes, I know that it’s not technically a fart, because it has nothing to do with digestion. Just go with it.)

When I first heard this story 15 or 20 years ago I tried to add the term “chasing fish farts” to my lexicon as a euphemism for a developer who is looking in the wrong place to find the source of a problem. Or a product team chasing a misleading metric. In general terms, mistakenly thinking that a particular type of event or sequence of events is the cause or the cure, I tried to refer to as “chasing fish farts.”

I couldn’t make it stick. People would look at me funny and I’d have to derail the conversation to tell the amusing story of the Swedish navy chasing Russian subs that turned out to be fish farts. Everybody would laugh and promptly forget the story. If I mentioned “chasing fish farts” a week later, half or more of the people I told the story to a week before wouldn’t get the reference.

But I still chuckle whenever I think of it.

What’s curious is that I’ve never been able to understand how in the world the Swedish navy got the idea that the sound they recorded was a Russian submarine. None of the articles I read at the time explained that, and a quick search today didn’t reveal anything new. Which is unfortunate because although failures are usually instructive and often funny, we learn the most by understanding the events and decisions that lead to the failure. There had to be a critical point when somebody said, “This is the sound that a Russian submarine makes!” I want to know how they came up with that.

The lesson to learn here is that when you repeatedly find the data pattern that your previous analysis indicated precedes the event you’re interested in, but the event never happens, it’s time to re-visit your analysis. Yeah, it’s embarrassing to admit that you discovered you were looking for the wrong thing. Better that, though, than to have somebody else point it out.

https://improbable.com/tag/herring/ will get you a few links to articles about the incident. Or just search for “swedish fish farts”.