Pestilent Double Bluff > Promise Double Bluff

[Romain672]

I think you are wrong springiest since I believe it's a difference between convention's sets.
I will rewrite it nicely:

Played: r1
Bob: b1-x-x-x
Cathy: ?-?-?-?
Donald: x-r3-r1-x
Alice clue 3 to Donald.
Bob plays b1.
What should Cathy do?
What should assume Donald if Cathy doesn't play?
What should assume Donald if Cathy play?

[springiest]

both very easy questions!

if cathy does nothing, donald assumes he has r3 exactly.
if cathy plays (say a g1), donald assumes he has any 3 except r3 or g3.

[Blacktango]

Why Donald wouldn't assume he has b2-b3-x-x ?
It should look like a self-finesse for blue 1 + 2 + 3, from their point of view. (At least, on BGA.)

[springiest]

b1 doesnt connect to 3 (it only connects to b2), so alice's clue is a bluff rather than a finesse if possible. Cathy's Connecting Principle explains this idea a bit more clearly. for r3 a bluff is possible, for other 3s it is not possible (because no other 3 is a valid bluff target).
this is why, even if donald indeed has `b2 b[3] x x` cathy must STILL blind-play! otherwise donald just assumes he has r3 (which is the only valid bluff target) and does something else.

[Blacktango]

OK, this is what Romain explained.

You describe the H-group convention, which is not very played on BGA.

All of your examples make much more sense with the H-group convention.