one of the reasons I cited file length was because is the drift were say .0001%, then a larger files size would show drift sooner since .0001% of 4GB is a much larger number than .0001% of 0.6GB. And As also stated this was not a full band recording. Could I have picked it up if it were a recording of the full band. Quite possibly not. But since It was something with very sharp peaks and easily identifiable, it was not hard to identify. all of the above numbers are completely made up for the purpose of an example. I do not pretend to know how much of a percentage off anything was.
I would also like to add, that there is nothing in the world to say that two decks will not run the same. Case in point, the same R-44 I owned ran dead on with a marantz CDR that the sound guy owned. DEAD FUCKING ON, without any sync at all. So can your example of two R-44s running in sync be right, why yes it can. Can two R-44's run different, why yes they can. Can you depend on them running in sync? I would not and if it was a paying gig I would either rent gear with the number of tracks I need or borrow a USBPre2 to ensure that they are locked to a single clock source.
And Ziko, even on a stage lip set up, I find that the mere 4 or five feet to the lip causes a time shift which can be adjusted to bring the SBD track and the mic track into perfect sync. It's not much but it is there. Maybe I'm too particular or too anal but that tiny difference helps with phasing issues and comb filtering effects.