The EMDrive is controversial. Sealed cans should not be able to move on their own accord. A number of theories have been proposed. Leading ones involve some kind of relativistic action with spacetime and pushing some form of virtual particle out of one end. At least one recent test by a DIY engineer (engineer building a test article as a hobby) in Germany has detected an unknown emission coming out of the drive opposite to the direction of expected thrust. The unknown emission interfered with the path of a laser between fixed points, possibly suggesting some kind of gravity ripple. If true this would indicate that the EMDrive is not a closed system and does not violate conservation of momentum (which makes it a hell of a lot more likely to actually work).
On Saturday a member of the team at NASA released comments on a pre-publication third party paper calling the NASA tests into question. The release indicates NASA is still detecting an anomalous thrust from the device and had already compensated for the proposed source of error. The team has sacled up the force up since the last public release of information and are attempting to scale if further.