Roger,
I found your reply after posting my reply to myself (above) that says that B could also be 'better'. My answers were in separate posts to confuse other readers 8-). My reasons are based on my understanding of your explanation to Petra about the Test 1 answer:
A - the diagram shows only one earthquake per window (I think - but don't have it in front of me now), and the test statement says one earthquake was predicted for each window.
B - could be correct if number of days in a no-earthquake sequence is more significant than predicting only one earthquake per window. There are 6 no-earthquake days between some B earthquakes.

Mary C.