The full story
A changing river. A steadier map.
At dawn, Mira the marsh-cartographer measured the river by dropping a pebble and listening for the splash. Fog made each splash unreliable: some echoes were late, some were swallowed. So Mira kept two ledgers. One held her prior map, drawn from yesterday's tides. The other held today's noisy readings. Each morning she blended them, trusting the old map when the fog was thick and the new splash when the water was clear.
Soon the villagers noticed her route was never perfect, but it was always steadier than any single guess. When the storm changed the river overnight, Mira did not cling to the old map; she updated it just enough to follow reality without chasing every ripple.
That is the Kalman filter: a recursive estimator that fuses a prior prediction with a noisy observation, weighting each by uncertainty so the next belief is sharper than either input alone.
Moral: in a changing world, the best compass is prediction corrected by evidence.