Infiniti2000 said:
Why do you "reset" the path for case A and not case B? If you have to draw a line all the way back to the origin for case B, why not for case A? In other words, do the same thing for the web by drawing the path from square to square. Any cover between intermediary squares? No. What you are doing is always comparing back to the origin for the web, but not for the wall.
No, one does the same for both. One starts at the origin and takes all paths. If one path has no cover, then the target gets no cover.
Example:
CCC......
FCC..XT
CCC......
Say we have a Widened Fireball. It has cover squares everywhere around it. X is total cover and T is the target. The target gets cover from the C cover squares, but does not get total cover from the X total cover square since there is a path from the origin that avoids that total cover.
In the case of Web it becomes:
CCCCX....
FCCCXXT
CCCCX....
There are no total cover squares that can be avoided if you determine this from the origin point. It only happens in your interpretation as:
CCCCX....
FFFFFXT
CCCCX....
The fire "removes" the cover as it advances (not literally removes, but removes because here is no cover 5 feet away).
But, that is your interpretation that assumes that the cover changes along the path. My interpretation assumes that when the Web spell states that there is total cover after 20 feet, that is what happens (as per the second diagram). The fact that it is a Spread does not change that because there are no corners without cover along the path.
The rules do not really have this concept of cover changing as a spread advances. That is a model you introduced.
The rules do, on the other hand, state that cover occurs after 5 feet and total cover occurs after 20 feet in a Web. They also state that Spreads can avoid cover if they can go around a corner and avoid cover by doing so. That is not what happens with Web. Web provides cover everywhere beyond a certain distance.