If, as the database suggests, CT-100 cabinet numbers are restricted to sets of 1000 (0 to 999), and if we assume for a moment that folklore is correct and 5000 CT-100s were built, and then sold, contributed, bartered, or fell off a truck, etc., there will be five complete sets of cabinet serial numbers.
If we randomly select one number from the first set of 1000 (say, 605), that number will occur four more times -- once in each of the four remaining sets of cabinet serial numbers.
So, what's the chance of randomly finding a match to cabinet number 605 in the remaining 4999 CT-100 cabinets? It must be 4999/4, or virtually 1250 to one. Or to consider it another way, one must amass a CT-100 cabinet serial number database of 1250 before a duplicate number can be expected.
Clearly, this ain't gonna happen. Maybe what I need is the Atlantic City factor. I've gotta beat the odds. Well wait a minute, the "cabinet" serial number actually occurs three times -- on the lid, cabinet, and front-panel insert. Therefore, the number of actual pieces with the same serial number on it becomes (not four, but) three times four. Twelve divided by 4997 is about 417. Better odds, 417 to 1, but still a long shot. Maybe I'll get lucky. That's what I'm hoping anyway. Let's see, Atlantic City is about 65 miles south of here....