In black: power spectrum of pure simulated signal at 143 GHz projected on 1.5 arcmin rings. In green: cross-power spectrum between (1) pure signal minus recovered signal and (2) pure signal. Solid lines join positive values and dotted lines negative values. In red: cross-power spectrum between (1) pure signal minus pure signal flags applied; and (2) pure signal. We see no apparent bias in signal estimation from either template subtraction or flagging.