The availability of flexible radio interfaces capable of adapting their configuration to the time-varying operating environment is the key response to the demand encountered in modern wireless networks for high data rates under strict quality of service (QoS) constraints. To this end, this paper develops a novel link resource adaptation (LRA) scheme for soft-decoded multiantenna (MIMO) bit interleaved coded orthogonal frequency division multiplexing (BIC-OFDM) transmissions employing automatic repeat request (ARQ) mechanisms. As the first step, a simple link performance evaluation model based on the effective signal-to-noise ratio (SNR) mapping concept is derived in a closed-form expression that it is shown to yield better accuracy than previous techniques. Then, an effective LRA strategy is formulated taking advantage of that framework. The aim is maximizing the goodput (GP) metric, that is to say, the number of information bits delivered without error to the user by unit of time, over the available radio resources, such as the power distribution on the subchannels, coding rate, modulation order and the MIMO configuration. The numerical results demonstrate considerable performance gains compared with nonadaptive transmissions, while keeping the computational complexity at affordable levels in view of the specific structure of the GP objective function.