Erasing free space - why more than one pass for 'more' security?

When using the Disk Utility to erase free space on a disk, why does one need more than one erase pass to get extra security.

Saying that making multiple passes gives extra security implies that a '1 pass' erase doesn't do the job properly and that doesn't appear to make much sense.

As a simple example, if I were storing the data contents of a matrix in RAM, I could write a program which overwrote locations in RAM used to store one set of matrix values, with another bit pattern representing a new set of values (EG zeros) and it would be 100% successful in 1 run of the program. (It's been user tested)

Having used a program to transfer the RAM contents on disk, I could then use the same program to take the new data from RAM and overwrite the old stored data contents on disk. I would then expect that area to contain exactly and precisely the new data contents of the matrix (nothing more & nothing less) in exactly one run of the program. (A one pass overwrite) (All the usual caveats - same data volume/initial address etc)

Obviously disk storage and memory are different kettles of fish and disk free space is not necessarily contiguous but the same principles should apply to both.

(EG Free space has start & end addresses)

Explanations welcomed.

David