Not really, his stats (against Test-class teams) of the last 4 years have been distinctly average.
Again, it becomes a question of which statistic you wish to see. If you wish to see his stats against Test-class teams of the last 4 years to determine if he's the better batsmen, surely it shouldn't be too much of a hassle to filter by matches played for a country, rather than a mixed-nationality opposition. Tendulkar still leads aggregate runs in ODIs and Tests by a fair margin, but anyone who has watched cricket recently knows that he has not been in the top 3 batsmen in the world in recent times. The statistics provide the proof for that, but should not provide the basis.
Why even bother having the classification at all, then? Simple. Because it matters from a statistical POV. Most cricket followers love stats, and you need a status feudal system (ODI, List-A-limited-overs, and I'd have a third one - List-B-limited-overs) to generate that.
I'd definitely agree that you need to separate ODI, List-A and List-B. But this tournament was devised as one that pitted international players against each other, which is definitely of the same league as (if not higher than) international cricket. That the players decided to opt out is another thing altogether, and one that neither the ACC, ACA or ICC foresaw (although they should have when the latter devised the FTP).
Classifying Bangladesh vs Sri Lanka as the same as South Africa vs Pakistan is ludicrous.
I'm not sure I understand this... In the long run, the statistical importance of the games against 'minnows' has a lesser effect. Also, by this system you will have to keep shuffling teams around and that will cause a larger headache for statisticians than the introduction of mixed-nationality teams has caused. For example, West Indies, by all tokens, would qualify to the highest rung in the 70's and 80's but now they would have to strive to reach the middle rung.
I'm a statistics whore myself, and I agree that most cricket fans are, to a certain degree. But statistics never tell the whole story. In short, statistics are only important for the statistics themselves, and not the actual cricket being played.