Use of the brightest stars in galaxies as a distance indicator and claims that the method supports the `short' distance scale are examined. Data from several different observational programmes are brought together for the first time with a procedure for the careful accounting of errors. The true uncertainties of the method are found to be much larger than claimed by its advocates. The method is incapable of distinguishing between the `long' and `short' distance scales.