How accurate are indices generated by BJ-strat program comapred to the
one by a simulator?
Linear approximations like the ones bj-strat makes are actually pretty good. Given that the gain from playing strategy is so small in shoe games, and it is somewhat subjective what the "best" indice is for a given play, its rare we would be talking about more than hundredths of %.
Risk-averse indices do gain rather more in terms of bankroll growth for no extra effort, but that is really a separate issue.
