[How will they react ?] The Compiled Wall Breaker's AI Discussion Thread.

Quote:

---

Originally Posted by pinser

Because I'm nice, I am willing to sacrifice my base for the answer to this question.(will the wbs go to that wall when deployed? I'm pretty sure but I want to make sure)
http://i50.tinypic.com/11rvma0.jpg

---

If you're talking about the T, then yes, they will go to it when deployed in front. There needs to be a difference in elevation, ie move that T back towards the Builder's Hut, if you want the WBs to branch off to the left and right walls.

Quote:

---

Originally Posted by Pingfao

If you're talking about the T, then yes, they will go to it when deployed in front. There needs to be a difference in elevation, ie move that T back towards the Builder's Hut, if you want the WBs to branch off to the left and right walls.

---

I want them to attack it because, according to my calculations, there should be a single walls left and i want to see if the wb will attack it.

Here's a question worth considering:

How long can a wb lane be before the wb loses its heading? I know that lanes length 12 work. Anyone want to post longer paths? Any reason to think it's not an absolute number but depends on the walls it's bypassing?

Quote:

---

Originally Posted by Hannibal

Here's a question worth considering:

How long can a wb lane be before the wb loses its heading? I know that lanes length 12 work. Anyone want to post longer paths? Any reason to think it's not an absolute number but depends on the walls it's bypassing?

---

That is a great question, ill try and keep track as I play around.

By the way whoever was first to introduce the diagonal long spike wb path much thanks! That is such an awesome strategy.

Quote:

---

Originally Posted by Hannibal

Here's a question worth considering:

How long can a wb lane be before the wb loses its heading? I know that lanes length 12 work. Anyone want to post longer paths? Any reason to think it's not an absolute number but depends on the walls it's bypassing?

---

Quote:

---

Originally Posted by TapouT

First time posting, so I might as well be helpful.
According to my experience, WBs don't depend on the lane's length, but rather by the symmetry of the base, and how many layer of walls are there. Let me give an example:

123 321
| | | | | |
| | | | | |
| | | | | |

Say this are 6 walls. If you try to make a path with the WBs, they will go through the first 3, after which they will change their "heading". If you want to go through all 6, you will therefore have to attack from both sides.

Let's check another case:

12 3 21
| | | | |
| | | | |
| | | | |

Here we got 5 walls. In this case, whether you start from the right or the left, the WBs will break the first 3, and then change paths, therefore the mid wall can be blown up by either side.

Obviously the difference is that in the first case the walls have an even number, while the second case has an odd number of walls- this means that when the number of walls is ODD, the WBs overlap on the mid wall (if the walls were even, then each side will have its separate set of walls to break).

This is only a THEORY, and its based on my opinion and observation, so don't take it as a fact. Yet... :)
Hope I helped

---

Not quite sure what the 2 examples were trying to accomplish regarding symmetry TBH. It boiled down to the distance between the first layer and the 4th layer in both of those cases.

I agree with Hannibal that it's based on the lane's length.

I have to say that I also think path length is based on base symmetry. I've been recording paths over the last couple days and am seeing different path lengths on different base plans. My suspicion is that wall segments exhibit a "magnetic pull" on wbs, and that the longer the path gets, the weaker the "gravitational pull" of the lane you've established and the more likely they are to be pulled off it by arrangements of wall tiles that "pull more strongly."

I know that sounds like hocus pocus, but it seems most consistent with what I'm seeing.

Quote:

---

Originally Posted by Hannibal

I have to say that I also think path length is based on base symmetry. I've been recording paths over the last couple days and am seeing different path lengths on different base plans. My suspicion is that wall segments exhibit a "magnetic pull" on wbs, and that the longer the path gets, the weaker the "gravitational pull" of the lane you've established and the more likely they are to be pulled off it by arrangements of wall tiles that "pull more strongly."

I know that sounds like hocus pocus, but it seems most consistent with what I'm seeing.

---

As in, a more hollow base will give you longer paths; vs a normal base, with a lot of side walls/sideway "magnetic pulls," will give you shorter paths?

Something along those lines, yes. I have examples of paths that fail at surprisingly short lengths on designs with unusual wall patterns, as well as at least one 12 length path on a very symmetric base plan. But in fairness I want to record more paths before I commit myself to a theory. It's possible I'm interpreting the noise rather than the signal at this point.

Quote:

---

Originally Posted by Hannibal

Something along those lines, yes. I have examples of paths that fail at surprisingly short lengths on designs with unusual wall patterns, as well as at least one 12 length path on a very symmetric base plan. But in fairness I want to record more paths before I commit myself to a theory. It's possible I'm interpreting the noise rather than the signal at this point.

---

Interesting. Keep us posted. :o

From what I have been experiencing, after 3 separate layers, wbs get tired and don't wanna go to the next layer. :rolleyes:

Ok, I'll try to elaborate.
Let me first give my thesis:
Each base has a center, and WBs will not loose their path until they reach it. ok?

Now to prove it, I went to the single maps:

First case:

Attachment 3795

You made me measure distances on my screen with a ruler, thank you :| ...
The red line shows the two axes, the circled stuff are for the WB and its target (there are 2 WBs, and both are heading to the same target, but i wasn't fast enough the first time to screenshot it, lol).
you see now, the last wall the WB targets lies on the red line, therefore anything FURTHER was not targeted (and yes I tried).

It broke 4 walls, and left 2 on the opposite side, which it couldn''t target.

Second case (can't upload image, must be URL...):

http://prntscr.com/unp93

Here it went through 3 walls, and couldn't go past that. Till here I thought my thesis was looking good, and might be valid, but then sherbet came along and made me mad...

Third case - sherbet:

http://prntscr.com/unrr6

In this case the WB was attacking the FIFTH wall in its path. What threw me away was what happened afterwards... the next WBs lost their path, while according to my thesis they should've went for the last one which was next to the townhall (sixth layer) ...

I admit my thesis was wrong... But now I wonder why was it wrong? the only difference between the sherbet and the other cases was the direction from which I deployed the WB- I deployed them on the lower left side on sherbet, while I deployed the WB on the upper left side on playhouse and jump around...

Now I hope my idea is understandable, and you also understand why its wrong :D

Ideas please