Considerable efforts in recent multigrid research have concentrated on algebraic multigrid schemes. A vital aspect of this work is uncovering algebraically smooth error components in order to construct effective multigrid components. Adapative or self-correcting multigrid schemes expose algebraically smooth error, analyze the effectiveness of the resulting multigrid algorithm and adjust the cycling as needed in order to improve the rate toward convergence. This talk will discuss an adaptive multigrid method that uses relaxation and subcycling on complementary grids as an evaluative tool in correcting MG cycling. The particular implementation of this idea manages smooth error in a manner analogous to spectral AMGe. Numerical results will be included.