The only way to know for sure is to analyze the traces of a drivetest.
If a MS enters in no coverage mode, you will certainly notice it… I don’t understand why this is a part of your assumption. It either happened or it didn’t.
Other possible scenarios are:
1./ MS2 goes out of coverage, look for the strongest cell in your band >> BTS 3 is selected (not “re” selected)
2./ In the BA list of BTS 1, there is a list of BCCH frequencies that should belong to the neighbor cells.
Therefore it might be possible that the BTS 3 has the same BCCH as one of the neighbors of BTS 1. MS 2 is being fooled, in a way (well, a MS isn’t fooled, ever, it’s just a machine).
3./ Then the difference between MS1 and MS2 can be the sensitivity, but also the processing speed of the results. Two different MS might have noticeable difference in this department, especially in idle mode. Perhaps MS 2 takes a decision 10 seconds after MS 1, so the radio conditions around have changed. MS 1 is already on BTS 1, but MS 2 is out of coverage.. well it’s just an example.