Blue is the colour between violet and green on the optical spectrum of visible light. Human eyes perceive blue when observing light with a wavelength between 450 and 495 nanometres. Blues with a higher frequency and thus a shorter wavelength gradually look more violet, while those with a lower frequency and a longer wavelength gradually appear more green. Pure blue, in the middle, has a wavelength of 470 nanometres. In painting and traditional colour theory, blue is one of the three primary colours of pigments, along with red and yellow, which can be mixed to form a wide gamut of colours. Red and blue mixed together form violet, blue and yellow together form green. Blue is also a primary colour in the RGB colour model, used to create all the colours on the screen of a television or computer monitor.

In statistics, the Gauss–Markov theorem, named after Carl Friedrich Gauss and Andrey Markov, states that in a linear regression model in which the errors have expectation zero and are uncorrelated and have equal variances, the best linear unbiased estimator (BLUE) of the coefficients is given by the ordinary least squares (OLS) estimator. Here "best" means giving the lowest variance of the estimate, as compared to other unbiased, linear estimators. The errors do not need to be normal, nor do they need to be independent and identically distributed (only uncorrelated with mean zero and homoscedastic with finite variance). The requirement that the estimator be unbiased cannot be dropped, since biased estimators exist with lower variance. See, for example, the James–Stein estimator (which also drops linearity) or ridge regression.