The simplest explanation and confirmation bias
The simplest explanation is that MH370 crashed into the sea near where it disappeared, and the rest of the shaky evidence and wild speculation is simply wrong (as much of it has already proven to be).
I'm working on an article to that effect with an associate who specializes in the psychology of investigators and the ways that investigations can go wrong (we don't expect the missing flight to be found before the article is published). This case is a classic, with all the signs. My associate doesn't know how to evaluate the technical details in the reporting, so I'm helping out as an engineer who knows a lot about air traffic control systems and satellite systems. While there is certainly some technical evidence of a flight diversion, I have to say that it has been very poorly reported, and I consider it at least as questionable as other evidence that has already proven false.
Most people don't seem to understand the need to evaluate individual bits of evidence independently in the early stages of an investigation to avoid "confirmation bias". One bit of questionable evidence leads to a theory, and suddenly everyone is trying to confirm that theory, adding more questionable evidence that isn't independent while ignoring other evidence that doesn't fit the theory.
There is certainly some evidence that the flight was diverted, and I wouldn't rule out that it might well prove to have been in the end. But extraordinary claims require extraordinary evidence, and so far there are only two bits of evidence that amount to anything in my view:
1. The wreckage hasn't been found in the area of disappearance. In every previous case of an airliner crashing at sea, floating debris and bodies were found within days. There are possible explanations, but I still think it's a key point.
2. The supposed InMarSat data. The problem with this is that the news reports are extremely vague and often inaccurate. It may consist of a single data point, and having never looked for such data before, InMarSat officials probably have no idea how accurate or error-prone it may be. Any number of possible errors could render it meaningless, such as a mis-identification of the transmitter number, incorrect conversion of the time stamp, or incomplete data stuck in a buffer being flushed out hours later.
The military radar data is nonsense.
It's not much on which to base a lengthy and extremely expensive search effort when common sense suggests that the original search area is more likely to be correct. Since it's no longer an emergency, the investigators should sit back, clear their heads, and go over the evidence again with fresh eyes and more in-depth analysis.