Yeah, it is fine that casual watchers of the game don't realise this yet but it will become apparent to even them in the next few years. T20 is as seperate a skillset as ODI is to Tests, and success in one is no guarantee of success in others. This is already accepted for bowlers, and batsmen to a degree but this will become more and more apparent. Especially as the likes of Smith play very little T20 vs guys who will be playing 20-30 games a year in it. You will get the occasional guy like Root who is so good he can adapt, but he will be the exception not the rule.
I agree with this. ODIs and T20Is actually have significantly different requirements - don't get fooled by the same jerseys and the fact that they are both limited overs cricket. ODI cricket has 3 'Acts' in an innings: Overs 1-15 (Start), 15-40 (Consolidation) and 40-50 (End). It is possible to lose early wickets in the first Act, have a good partnership in the second Act and you're still set for a big finish. Someone generally needs to score a 100, and honestly whether that's the opener or a middle-order batsman, there is not much of a difference. Big hitters are useful, either to get you off to a flyer in Act 1 (a la Jayasuriya / Gilchrist), or they should come in Act 3 when they just need to slog. Final team totals don't have that much variance. For example, a bad start / good recovery or a good start / poor finish can both lead to 270 runs.
Now T20s are just a single act. 20 overs which are highly momentum driven. An opening partnership can put on 100 and 9 times out of 10, they'll probably win the game. The simple fact is that there are only 120 deliveries to face. That doesn't give the batsmen too much time to settle down. And openers have a great deal in T20Is - their middle order colleagues need to be far more adaptable. A #4 batsmen in T20s for example should be adaptable enough to either come in the 2nd over or the 18th over. Here is where I think Raina has been struggling. During the WT20 he often came in early, and he just didn't look comfortable.
From a bowlers point of view, wickets are almost meaningless. That is not to say that getting wickets is not important in T20Is - it is extremely important. But my point is the bowler should concentrate on starving the batsman for runs and then getting the wicket. Giving a lot of runs away just to get a single wicket is pointless. A bowler has 24 balls - he should be ensuring a majority of them are dot balls. Mohammed Amir was devastating in the WT20 as batsmen just couldn't score off him. If Pakistan had better bowlers to support him, they could have got wickets by simply choking the opposition.
Overall, T20s can be a funny format. There is a huge variance in team totals. The same team that makes 180 in a game can get bundled out for 80 a few days later and it won't really mean much. It's time to carefully think about the team composition for a T20 and not assume it needs to be your ODI team with a few modifications. You definitely need batsmen like Kohli and Root who have great technique and can score 50s even when the rest of the team is in trouble. But you need your middle and lower order batsmen to be smart - know when to take the singles and when to go aerial. Of course, you do need a couple of guys who can hit 6s with brute power. For the bowlers, you need crafty guys who can outfox the batsmen. Bowlers who hold their nerve when they have to defend 10 runs off the last over, bowlers who can deceive the batsman and steal a wicket against the run of play. And I don't need to mention the importance for excellent fielding.
TL
R, it's going to be exciting to see teams come up with new T20 teams and strategies. This will be a third legitimate version of cricket.