Calibration of flicker against asteroseismic ﹩\mathrm{log}\;g﹩. Top: 8 hr flicker vs. asteroseismic ﹩\mathrm{log}\;g﹩ for the calibration samples used in this work. Red points are from Chaplin et al. (2014), cyan points are from Bruntt et al. (2012), and magenta points come from Thygesen et al. (2012). Data from Huber et al. (2013) are shown in green, and we plot the data from Stello et al. (2013) in blue; this last data set contains a mixture of red giants and red clump stars. The star symbol to the lower left represents the Sun, through which we force our fit. Solid points are those we used in the initial fit; open circles are “doubled-back” stars that we exclude from the fit (see text). The curves show our initial fits to the data; the residuals (second from top) show significant structure, caused by the doubled-back stars pulling the fits. The rms scatter about these residuals is 0.1 dex, and the median absolute deviation is 0.09 dex. Third from top: to reduce the impact of the doubled-back stars on our fit, we subsequently iteratively reject outliers lying 0.1 dex above and 0.2 dex below the polynomial fit, achieving convergence after three trials. The dashed curve is our initial best fit, and the solid curve is the result of fitting after outlier rejection. Open circles are points that were ultimately rejected from the fit while solid ones were retained. This process not only removes the doubled-back stars but also removes red clump stars, which show a slightly different F₈–﹩\mathrm{log}\;g﹩ dependence (see Section 3.2). The residuals, shown in the bottom panel, reveal significantly less scatter about the final fit for the stars used in the fit: the rms is 0.05 dex and the median absolute deviation is 0.04 dex; however, we use the rms and median absolute deviation from our initial fit as our formal uncertainties for the F₈-based gravities.