I am by no means an expert on surveillance - I just respect Inmarsat's capabilities, and what they say. Their data-set and analysis, I think, is most credible. They also ran this 'new type of analysis' , by a peer review group, and have also published some of their 'new' methodology, which will be undoubtedly critiqued by more experts.
The malaysian PM bases his statement seemingly on inmarsat data and analysis, and so is going out on a limb a bit. There may be other data given to him in confidence, which they don't want to release, and that makes them sure of the south route. The 'idiot artifact' in this case is going to be the actual debris field.
The parallel with AE is, at least, the lessons that the current searchers could follow from the AE search. In hindsight, things always become clearer. When they were searching for AE, like here, time was not on their side, and resources were far more limited. They did have to make certain assumptions in the heat of the moment, just because, like here, with limited resources, they had to concentrate on the most likely places to search. Similarly, in this search, they have to necessarily make assumptions/speculate, maybe on incomplete info, so that the planes and ships etc.,they do have are used most effectively.
So, when I was using words like 'improbable', I was trying to follow why the searchers were concentrating on the south route, which seemed logical. The northern route clearly went over central india, and if the indians had missed it, it went over northern afg-pak borders areas, where we would prolly have caught it, as we own the airspace over afghanistan.
In hindsight, if we could turn the clock back, and know what we know today of the Niku hypothesis, esp the radio signals which have been strongly shown to be credible, we would say "forget everything else - send all the planes and ships to niku". From what I can remember on the AE search, if Lambrecht, by design or happy coincidence, had overflown Niku 2-3 days earlier, and/or the tide had been low, the plane would have been out in the open. But in the pressure cooker environment they were in, like here, they had to speculate and reason a bit, from the data they did have. Time was running out -like here it is still running out to find the blackbox.