[blml] "Demonstrably" - practical meaning?

[Alain Gottcheiner]

Harald Skjæran a écrit :

(Nigel)
>> Take Alain's example 1S - 3S (limit) after hesitation. Opener now has a
>> borderline decision between passing 3S and bidding 4S. Suppose that:
>>
>> [A] AO is a notorious overbidder. If he tanks then bids 3S, his hand is
>> probably worth 2½S. Hence the BIT suggests that opener passes.
>>
>> [B] AO is a notorious underbidder. If he tanks then bids 3S, he probably
>> has a hand worth 3½S or more. Hence the BIT clearly suggests 4S.
>>
>>     

(Harald)
> Huh? I'd neither expect a notorious overbidder to tank before doing
> what he normally does, nor a notorious underbidder to tank before
> doing what he normally does.....
>
> I don't think you can reason like that.
>   
(AG)

And here I am again, insisting that, if two trained directors and clever 
guys don't agree, even knowing about the player's tendencies, what his 
hesitation could suggest, then neither could his partner know, which 
means no UI.
Whatever the exact sense of "demonstrably", surely you can't demonstrate 
two opposite things from the same starting point.

Best regards

    Alain