Mauro,
> This argument is overrated: it indeed applies to 90% of the most
> common conventions.
I don't think so. Conventions that appear in mainstream textbooks usually have sound follow-ups. J2NT doesn't.
> Think of it, for an occasional pair, even the followup after a negative
> double is not really defined according to a standard...
> Just to mention a few other instances, this applies for most doubles,
> 2 suited overcalls, inverted minors, and even 1NT forcing sequences,
> and many more occasions...
Sorry, I disagree. Most of these are reasonably well covered in textbooks. (I will agree on the inverted minors though.. the standard follow-ups became mainstream, but are unsound in my opinion.)
> This does not mean we should play without conventions (...)
>
it just means we should know our stuff.
Yes, we should. But most don't. So it's better not to play conventions.
> There is a urgent need for a balance-of-power raise that
> IMMEDIATELY signals the presence of a 9 card fit.
I disagree again. There would be such a need if you antecipate opps are going to butt-in. But since RHO passed and we have a fairly strong hand, there is a good chance that LHO won't be bothering our side if we fail to convey our good support immediately. There is, of course, some risk in not doing it, but that's overrated.
Suppose you're playing 1M-2NT as something else, and are "forced" to show your 12+ 4-card support by temporizing with a 2/1 bid. Now compare
1
♥ (pass) ..?
with
1
♥ (1
♠) ..?
In the 1st case, if you bid, say 2
♣, LHO won't even know you have an undisclosed fit. After 2
♣ we will only bid what's in front of him.
In the 2nd case LHO won't be quiet, so you obviously can't insist on bidding a forcing 2
♣. You have to make some support bid right away. Big difference.
This isn't just academic. This style works, and the proof is the french team of Chemla et al. beating the USA nickell team in the 1997 bermuda bowl.