Okay, Threat from the East states that:

When revealed, you choose one:

Each player draws 3 cards and then discard 3 random cards from hand.
Each player discards 3 random cards from hand and then draws 3 cards.

Now, how does the latter choice interact with Knight of Harrenhal:

If a card would be discarded from your hand at random, you may choose the card to be discarded instead.

Mainly, if we go with the second option (discard 3 at random), and a player decides to use Knight of Harrenhal's ability (note the word may, so it's not obligatory) to choose the cards that he/she discards, then does the then portion of Threat From The East interpret the first portion as not having completely resolved succesfully or not? I.e. how does the replacement effect (replacing random with choice) in Knight of Harrenhal interact with the fact that Threat from the East requires all players to have discarded 3 cards at random, for the second portion of drawing 3 cards to happen?

Pretty sure it would still trigger, replacement effects, afaik still can be used for triggers, stay of execution dosent actually stop the kill effec as its not a save or cancel. Instead it replaces the moribund state with a new destination but i believe that you could trigger appropriate when killed etc responses accordingly.

Or at lest i seem to remember that from somewhere, i could be conjuring it from my the dark addled corners of my mind so i would wait for ktom or someone to riddle this with poor game understanding before providing the correct answer.

Correct. The replacement effect - making the discarded cards intentional instead of random - still counts as a successful discard of "random" cards. The important thing to remember is that even though the Knight lets you choose the cards, it is still the plot that discards them from your hand.

That's all "success" is measured on - did the plot make 3 cards leave your hand. If "random" was a strict measure of success, then would the plot succeed in discarding 3 random cards from your hand when you have exactly 3 cards in your hand?

ktom said:

Ok, good to know. Thanks.

ktom said:

Well, in computational sciences 'random' numbers are often created using an equation that goes through all of the numbers (of a desired interval) in a 'seemingly' random order, creating a completely even distribution. Thus the end result is not actually completely random (rather, it's each value from the interval and the distribution is much more even than a really random distribution would be), but the 'sampling' is still thought to be sufficiently random. Similarly here, the order in which the 3 cards would be discarded from hand would be random, even though all of the 3 cards are discarded (creating a non-random end result).

~So, the answer would be, yes.

WWDrakey said:

~Because a computational sciences logic tree always accurately models the application of rules in AGoT….

ktom said:

ktom said:

Well… Both are based on the same core fundamentals of logic, formal languages and discrete mathematics. Although, like you said, the actual rules of parsing the formal language can be quite different (although, they are quite varied for different programming languages also, within the scope of computational sciences). Now, since the formal language in AGoT says nothing about random (the word isn't even mentioned in the whole FAQ), we need to look to other 'common' sources for the definitions of random. Similarly to looking to discrete mathematics for how '1+1-1' is evaluated, really. Computational sciences (or actually, Applied Discrete Statistical Mathematics, which are a kind-of subportion thereof) are just the only 'common' field, where I could hope to look for a definition on how to interpret if something is random, and at least it relies on the same core fundamentals as the AGoT rules. If you have another field in mind, do tell.

It's actually quite interesting to note how much the AGoT rules really resemble classic formal programming languages with their structures and parsing. Let's take an example:

"If X then do Y" <--- A common structure in programming languages, if X does not evaluate to be true, then Y is never even looked at.

"X, then Y" <-- A common AGoT structure, which funnily enough, is evaluated oddly similarly.

Talk about off-topic rants… Anyway, the ruling seems logical to me. The end result is similar to replacement effects when a character is killed (while the physical action changes, the game still sees the original as having happened), which seems consistent. Just being my usual guarrelsome self.