Strategies vs backward induction in dynamic decision-making: an experimental investigation