I've never seen this before, neither have a colleague and an ex-colleague, both with a reasonably strong background in smallish ISP operations.
What causes so many duplicate ICMP ECHO responses?
$ ping -n x.com
PING x.com (196.x.y.z) 56(84) bytes of data.
64 bytes from 196.x.y.z: icmp_req=1 ttl=120 time=51.8 ms
64 bytes from 196.x.y.z: icmp_req=1 ttl=120 time=51.8 ms (DUP!)
64 bytes from 196.x.y.z: icmp_req=1 ttl=120 time=52.3 ms (DUP!)
64 bytes from 196.x.y.z: icmp_req=1 ttl=120 time=52.7 ms (DUP!)
64 bytes from 196.x.y.z: icmp_req=1 ttl=120 time=53.6 ms (DUP!)
64 bytes from 196.x.y.z: icmp_req=2 ttl=120 time=92.2 ms
64 bytes from 196.x.y.z: icmp_req=2 ttl=120 time=92.6 ms (DUP!)
64 bytes from 196.x.y.z: icmp_req=2 ttl=120 time=93.0 ms (DUP!)
64 bytes from 196.x.y.z: icmp_req=2 ttl=120 time=93.4 ms (DUP!)
64 bytes from 196.x.y.z: icmp_req=2 ttl=120 time=93.8 ms (DUP!)
man ping
:It should be stated that public internet hosts usually are configured to discard echo requests to broadcast and multicast addresses to prevent smurf attacks - this is why you hardly ever see it happen.
It also might be the result of some kind of bizarre network looping where a packet gets copied in transit.