Expectation, Variance & Moments

the mode

When you ask 'what is the most common outcome?', you are asking for the mode. It is the third classic notion of a centre, alongside the mean and the median — but a different question altogether. The mean is the balance point, the median is the middle by rank, and the mode is simply the most likely value: the tallest bar in a histogram, the highest point of the curve.

For a discrete variable, the mode is the value x with the largest probability P(X = x). For a continuous variable, it is the point where the density f(x) is highest — the peak of the curve. A distribution can have more than one mode: if two values tie or two separate humps appear, it is called bimodal (or multimodal), which is often a clue that the data secretly mixes two different populations. And the mode need not match the mean or median at all — in a right-skewed distribution the usual ordering is mode < median < mean.

The mode is the natural centre when you care about the single most typical case rather than an average — the most common shoe size to stock, the most frequent diagnosis, the peak of a probability density. It has a special robustness: it is unaffected by extreme values in the tails. But it has weaknesses too: for continuous data it depends on how you bin or smooth, it can be unstable, and for some distributions (a perfectly flat uniform) every value is equally a mode, so the concept tells you nothing.

Roll two dice and sum them: the sum 7 occurs most often (6 of 36 ways), so 7 is the mode. Here mode, median, and mean all happen to equal 7 because the distribution is symmetric — but for a skewed distribution they would spread apart.

The mode is the most likely value — the tallest bar or the peak of the density.

A distribution can have several modes (bimodal/multimodal), often signalling a mixture of populations; and for a flat uniform distribution every value is a mode, so the concept gives no information.

Also called
most likely valuepeak最常見值峰值