Picture a pendulum hanging in a room with drafts. Sometimes a gust pushes it harder, sometimes softer. If the pendulum barely loses energy, it will swing long and steady. But if friction is strong, each push quickly fades, and the motion becomes jerky: a sharp swing, then a long sway. That's exactly how the tropical Pacific behaves when the heat of El Niño gives way to the cold of La Niña.

For a long time, climate scientists debated which pendulum best describes this phenomenon: a self-sustaining one that swings without external pushes, or a damped one that needs constant gusts. The answer matters for forecasts: how the ocean shifts from one phase to another determines when to expect droughts or floods. The authors of the new study didn't force the model to fit the data as was done before. Instead, they tried thousands of parameter combinations and ran each one in a computer simulation spanning 30,000 years.

It turned out that a self-sustaining pendulum doesn't work: it gives too "flat" a temperature distribution, without sharp spikes. But a heavily damped pendulum, which loses almost all its energy within 20 months, reproduces everything as it should. It is rocked by random wind gusts, and these gusts get stronger when the ocean is already warm. It's as if the draft blew harder exactly when the pendulum was displaced to one side.

But friction and random kicks alone aren't enough. To make the pendulum behave like the real ocean, you need two small "tilts." One makes the warm phase shorter but stronger, and the cold phase longer but weaker. The second tilt works like a spring that, after a strong swing, slowly pulls the pendulum back, keeping it in the cold phase. This second tilt was long overlooked, and without it the model couldn't explain why La Niña sometimes lasts two years in a row.

The result is a simple model that, for the first time, reproduces all the major quirks of El Niño and La Niña: their differing strength, their differing duration, and even how often one phase replaces the other. Now this model can serve as a benchmark for testing more complex climate models. If a big model doesn't reproduce the behavior of the simple pendulum, something is wrong with its description of the ocean.

The pendulum in a drafty room turned out to be a surprisingly accurate metaphor for a climate phenomenon that affects weather worldwide. And now, when a gust of wind rocks it too hard, we have a better idea of what to expect from the coming season.