i was wondering about the reliability of this theory do you guys think it's true?
For a start, those laundering transactions are exactly the kinds of things intelligence services are good at figuring out. They can put supercomputers to work analysing the global transaction stream (all available in ONE place; no need to talk to lots of banks - or worry about infiltrating uncooperative foreign banks). Some value that goes into an account then buzzes through a self-contained pool of accounts for some time then zooms out to somewhere else can probably be traced through analysing the timings of transactions and the like; the pattern of automated laundering will be different from actual spending, if you have enough computer power to find the patterns. Imagine drawing a diagram with a blob for each address you know something about (eg, can tie to a person or organisation), and drawing arrows for all the transactions between them. Any single-use addresses can just be chained together as part of the same arrow. Any unknown addresses can be given small blobs on the diagram. Colour the arrows with the magnitude of the amount transferred, on a log scale. Arrange the diagram so the minimum of arrows overlap. Do this for the transactions in each day, and then make a movie of them changing over time. Take a given known-suspect transaction and treat it like a drop of dye, colouring it strongly, and mixing it with the light grey of other money flowing through the system as it dissipates, and see where that dye spreads to. Then get computers automating the analysis even further."