Time-delay measurements along with the 1σ errors for RX J1131−1231 obtained by our three standard techniques from the full nine-season-long light curves shown in Fig 4. The random and systematic error contributions are given in parentheses for each delay. The error bars represent the random plus systematic errors summed in quadrature. A positive AB-delay Δt_AB means that image B leads image A. We consider the measurements from the regression difference technique, which display the lowest bias and variance in our error analysis as the delays that should be used to constrain cosmology and lens models.