Please explain how multizap bounces affect range

[tedric]

This commit changed the multizap targeter. Some positions that used to show multizap potential no longer do so; and from playing lots of AEs recently, it does seem that my zap passes through those formerly-multizap-indicated tiles only once. I guess I never paid enough attention to realize that these shots never actually delivered a multizap.

I'm grateful that the targeter isn't lying to me anymore, but I'm also deeply confused now about the way bounces affect the range of a zap. The Shock spell is full-LOS (range 7), but in some bouncing positions it seems to "fizzle" before 7 tiles worth of range have been consumed.

Example A

  Code:

#@#
#1#
#2#
#3#
###

Before and after, firing a zap straight down the corridor seems to multizap 1, 2 and 3. This makes sense to me, because I assume the bounce at the end of the corridor consumes 1 range (7-1=6, 6/3=2 multizaps per tile).

----

Example B

  Code:

@.#
#1#
#2#

Before: By aiming at the wall to the right of 1, the targeter indicated multizaps at 1 and 2. I could make sense of this by counting 3 bounces (to the right and left of 1, to the right of 2) and two passes each through 1 and 2 (running out of range the second time it passes through 2).
After: 1 multizaps, but 2 does not. Where did we lose 1 range?

----

Example C

  Code:

#########
@.....1#
#########

Before: Firing straight down the corridor indicated a multizap at 1. This shouldn't have made sense before, because we learned in Example A that bouncing straight back from a wall consumes 1 range and the zap had already traveled 6 tiles before hitting the wall. But it looked intuitive, so I never questioned it (and I could swear it's always multizapped).
After: Firing straight down the corridor ends at 1 without bouncing, as the math indicates it should.

----

Example D

  Code:

##@
#1#
.##

Before & After: Aiming at the wall to the left (or below) 1 indicates that the beam will only pass through 1, and will do so at least twice. If every "wall strike" consumes 1 range, this should result in 4 multizaps. But it only seems to cause 2.

----

I could go on and on like this, and the more complicated the corners get the less consistent the results seem to be. It almost seems like every zap pattern calculates range differently... but I suspect what's actually happening is that there is one non-intuitive rule I don't know about, which all of the weird situations are triggering. If somebody could explain it to me, that'd be great.

[Siegurt]

Well, all multizaps are capped at 2, so the last case you can ignore.

[Midn8]

Each bounce costs 1 range, and as stated above bounces are capped at 2. And in the second example there are some "hidden bounces" that triple-zap a square, but you can't really see that because the targeter only shows the 2 and the reduced range.

[Hurkyl]

Example B definitely used to multizap both locations -- have the actual mechanics changed?