Compare experiments

Choose up to three experiments: the same nation under different methods, scenarios or horizons, or the same configuration in two cycles of the improvement loop. Every method faces the same draws for the same seed, so a difference within a batch is the configuration's, not luck's.

Compare two tree-search experiments, with their regret curves, decision strips and radars; compare two CMA-ES experiments, with their regret curves, convergence and radars.