Results being independently reproducible is the foundation of science. Meyers-Briggs fails that simple metric.
But people take the test once, and then use the results for a long period of time. Guess what? At night, blood sugar level goes down. it's no longer X. And likewise with this test.
Same thing, here. I wouldn't put much stock into a single result on a single test, but I think there's some broader value to be gained by examining the trend, the fluctuations, and what you wind up as more often than not.
The only assertion I have to disprove here is that the test is "worthless".