Enhanced Scripts...

Kerozevok

If
condition
Then 
Response #100
action
End

No problem.

If
condition
Then 
Response #50
action A
Response #50
action B
End

Despite the 50/50, most of the time, the game always choose the 1st possibility, is it possible to improve this ?...

Ward

Change the first one to #49.



No to be honest I'm no help and would be interested to know more.

Kerozevok

We can write #10 in 1st and #90 in 2nd, the problem is the same, most of the time, the game choose the 1st possibility.

lansounet

How about

IF
RandomNum(2,1)
Trigger A
THEN
RESPONSE #100
Action A
END

IF
RandomNum(2,2)
Trigger A
THEN
RESPONSE #100
Action B
END

?

You can split the probabilities unevenly too with RandomNumGT() / RandomNumLT() (Instead of a lot of OR() RandomNum()), for example

IF
RandomNumLT(10,2)
Trigger A
THEN
RESPONSE #100
Action A
END

IF
RandomNumGT(10,1)
RandomNumLT(10,5)
Trigger A
THEN
RESPONSE #100
Action B
END

IF
RandomNumGT(10,4)
Trigger A
THEN
RESPONSE #100
Action C
END

Kerozevok

@lansounet I agree, but that increases the number of lines because the % are not respected by the engine.

lansounet

Well I've never seen people complain about BG2 Tweaks @CamDawg 's slow drow disintegration component.

It goes like this (pretty much):

IF
GlobalTimerExpired()
Triggers
THEN
RESPONSE #5
TakeItemReplace()
Continue()
RESPONSE #95
Continue()
END

and that seems to work fine for a 5% probability to break items everytime the timer is expired.

Kerozevok

However it's an old problem...

Scripting Guide by SimDing0 (writing in 2004)

You are not limited to simply using a single RESPONSE #100 in scripts. (...)
So, you could have, for example:

IF
// Triggers
THEN
RESPONSE #100
// Action 1
RESPONSE #100
// Action 2
END

This will theoretically execute either action 1 or action 2 with equal chance of each when the block is executed. However, in practice, there is a bias towards the first response (see this post). Truly random responses, if essential, can be achieved using the RandomNum trigger.

Moreover, I am not sure the game makes a difference between 5% or 300%.

lansounet

Yeh I knew about this when the values are equal or close. That's why I first mentionned RandomNum() as in that guide. However when a relatively low value is put first it appears to work somewhat fine.