My personal point of view: what matters much more than the distribution is the second-order statistics of the channel. In particular things like delay spread / coherence bandwidth and Doppler spread / coherence time. These parameters tell you a thing or two that you must take into account when estimating your channels:
The coherence bandwidth (roughly equal to the inverse of the delay spread) tells you for what bandwidth you can consider the channel approximately flat. If the bandwidth you want to transmit is below that, you can consider the channel a scalar quantity, which is simple to estimate. If your bandwidth is higher, you either need to go to frequency-domain techniques like OFDM and then estimate the channel on several frequencies (spaced no more than the coherence bandwidth) or estimate channel taps in time domain, e.g., when using a Rake receiver.
The coherence time (roughly equal to the inverse Doppler spread) tells you how quickly the channel changes in time, i.e., for how long you can consider it approximately constant. After the coherence time has passed, your channel estimate is no longer valid and you need to train again (or employ tracking if you can).
That said, the channel conditions (like LOS/NLOS) do influence these quantities as well. If you have a strong LOS, chances are your coherence time is higher since the LOS will not change rapidly in time.