Similar to the related conversation, (Tap button blinking irregularly) my conclusion is the "less important " functions of the remote (and toaster/rack) are given a low priority over the audio-processing of the main unit, quite rightly so and Visual feedback is probably lowest in terms of importance.
Im guessing the "polling" of buttons is limited to occasional cpu cycles rather than a priority task and more often than not it will (eventually) get the tempo correct from an average measurement of tempo-button taps (main unit and remote alike?) based on a few bars worth of taps, *after* its done all its other business, ie, a new performance load seems to consume processing resources for a second or two after its been executed.
Be interesting to hear the K-team's response to this to confirm if my theory is correct or perhaps there is another reason...