In the link layer frame retransmission strategies, the short-term variation of physical channel properties has not been fully exploited. To combat the negative impact of fading, channel memory can be used to make intelligent predictions about the upcoming channel state and frame retransmission timings can be adapted based on these predictions. In this study, two variants of simple yet efficient channel-adaptive stop-and-wait (SW) retransmission schemes are proposed for short-range wireless networks, wherein the transmitter node learns about the channel condition from the immediate past response of the receiver to adapt the next frame retransmission timing. Using Markov channel model, performance of the competitive SW protocols are captured. It is shown that, with a suitably chosen frame retransmission timing, the proposed schemes fair better than the existing SW retransmission schemes in terms of reduced energy consumption with a nominal data delivery rate trade-off. The performance gain is significant particularly in the harsh fading conditions.