IMHO yes. The hash id forces events to be "value objects" (no independent identity).
It's reasonable to assume that two different events, with the same data at the same readPoint and same millisecond, cannot be captured. But since a lot of fields are optional (including readPoint) while the granularity of eventTime is not mandated, this assumption is weakened.
So if the same event is received twice from some capture system, it should be treated as an identical event.
IMHO yes. The hash id forces events to be "value objects" (no independent identity).