One of the greatest uses of WSPR is to get accurate, real-world comparative reports on antennas. When comparing antennas in the past I have used two computers and two radios on two different frequencies using the same power and manually set the WSPR programs to both transmit on the next cycle so that they send their beacon signals simultaneously. During the transmission I set both programs back to "no transmit". This way I get directly comparable reports from both antennas at the same time and I don't have to worry about one transmission being received and reported by the other system.
Now, this is no longer possible. The WSPRnet website has implemented a "data scrubber" which will filter out invalid spots including bad calls (bogus decodes), bad timestamps, wrong band reports and duplicate reports. The latter are apparently the most common type of error, accounting for 1.7% of spots. I wasn't aware this was that common, though it can happen due to reciprocal mixing in the receiver or the computer sound card. But one result of this is that you can no longer intentionally transmit two beacons at the same time using the same call, locator and power level, because if you do all but the strongest report will be filtered out.
I think this is an ill thought out move that will result in deleting valid spots just because they can't be distinguished from invalid ones. The WSPR software does report the transmit frequency being used by each instance of the software so it should be possible in some cases to determine whether duplicate spots in the same timeslot are genuine or not. This won't help those using hardware based WSPR transmitters that don't connect to the website though.