None of the following is based strictly speaking on any actual research. On that basis, it is purely speculation on my behalf. I will defend myself however by writing that I am taking research results from different areas and putting them together, so I am not just making stuff up. But it is just a collection of thoughts that might not in reality fit together.
Blocked practice has been shown, as reported here for example, to produce some positive short term learning effect but overall less retention of the learned activity than random / distributed practice. For learning it is therefore clear that random practice is essential. But at different moments during the course of the season, the goal of the coach may not necessarily be learning. Sometimes the goal of the coach is a short term improvement in some particular area for a specific match. In this situation blocked practice might be a perfect solution.
As I wrote previously the content of practice is in itself a form of communication with the team. By choosing to practice a particular area in any way at all, even in a way that does not directly lead to long term learning, informs the team of its importance and draws their attention to it. In this case blocked practice can lead to on court success by priming the team for certain skills and situations.
If I take the train of thought further and add a few more speculations on top…
Although blocked practice does not lead to long term learning, coaches persist with using it. Obviously no coach wants his team to be worse, so one can only conclude that the coach ‘sees’ improvement in his team after blocked practice, particularly intra session improvement. In seems likely / possible that what he is actually seeing is just the priming effect of the drill and not actual learning. Hence the confusion.
It’s just a thought.