The reason why it's not that uncommon for two largish frame breakers to trip during the same event, especially in an uncoordinated system, is explained by the difference in the sensing of a fault and the actual interruption of it. Sensing, especially with electronic trips such as these, is fairly consistent. But interruption involves unlatch time, contact separation time and arc extinguishing time. Those time frames are almost never the same, and the larger the frame size, the longer it takes. So the same even is sensed simultaneously and initiates the trip in both of them, but although the smaller breaker opens first, the process had already started on the larger one and is irreversible once the unlatch has taken place, even though the fault may have already been cleared by the smaller breaker down stream. This is one of the main reasons why you do a coordination study. It looks at not only the trip settings, but st the total fault clearing times in order to know where the trip settings should be.
Now how you had two down stream breakers trip with the main is a different story. I like the cross connection theory on that one.