I hit three blocks in a row again, with less than 300kh/s .. This shouldn't be possible at the current difficulty...
909839 49 to go 1054658909 eaf7aabb143fe3aef10bdfee8d8ecb7eabc7c32f479d17361493ec90bc8b936e 1/12/2016, 11:53:48 PM 3%
909837 47 to go 1057789727 d52b7b1afb24a093e8d3d642014062a0910d232197ffaa8221f6588ab580ea17 1/12/2016, 11:51:58 PM 10%
909833 43 to go 1060389309 3d7bf46d75c5847a1deed22dfa19da42ad0eb0eb8895dba2fa1990732f0e3514 1/12/2016, 11:45:31 PM 2%
909832 42 to go 1058626345 7f967bbc5076d0cfddc297565acd253dc6b08caddf3f573b321f84cae68b899b 1/12/2016, 11:44:22 PM 3%
909831 41 to go 1057809920 0b355e2c882eff0c0392d2ff49ffb8a06ddfe733e63a761ae966c9d8fe5d5ac0 1/12/2016, 11:42:08 PM 29%
Seems like block propagation is the issue here, once you hit a block probability you get the next one is higher than it should be..
EDIT : This needs to be looked at, pools or anyone with high hash rate might find a way to exploit this..
It's not really "higher than it should be", there is no "what it should be". Yes probability to find a block is slightly higher if you know the previous one before anyone else, and it's not an exploit, just a fact due to propagation times.
I doubt this is having a meaningful (visible) impact with 1mn time blocks, but anyway the network will switch to 2mn blocks next hard fork (March).