How Moving Averages Smooth Out Noise
Launch library · evergreen read
Raw weekly figures can jump around considerably even when nothing meaningful has actually changed, since ordinary random variation alone is enough to produce a visible wobble from one week to the next. A moving average tackles this by replacing each single point with the average of several recent points, softening the line considerably in the process.
That smoothing comes at a real cost, though, since a moving average always lags slightly behind the very latest genuine shift, blending fresh developments together with older, already familiar data for several weeks before the picture fully catches up to reality. Choosing the right window length becomes a genuine balancing act for whoever builds the chart.
A short window stays responsive but noisy, while a long window stays smooth but sluggish to react. Compilers pick a length that suits their particular audience, favouring quick reaction for fast moving charts and steadier calm for charts meant to be read over many months.