I thought it was more like: Set a team of experts to find bugs independently for some specified amount of time. Then look at how many of the same bugs were fou

I thought it was more like: Set a team of experts to find bugs independently for some specified amount of time. Then look at how many of the same bugs were found by multiple experts. If most of the bugs were found by multiple experts, then there are probably not that many more bugs than the total number that they found. If most of the bugs were found by only one expert, then there are probably a lot more than the total that they found. With some math you can pin down the 'probablies' to numeric ranges.