Forums-->Queries and help-->
1|2|3
| Author | Dreadbane's attack |
|---|
If this is the matrix, as viewed by a player on his screen,
(S1)(S2)(S3)
(S4)(B)(S5)
(S6)(S7)(S8)
Then I guess it should hit S1, since it has the lowest coordinate. | | i think horizontal/vertical would be preferred over diagonal, so s2,4,5 or 7. | First of all, I must excuse myself beforehand as I've been away from the game for a long period and some things may have changed that I didn't know of.
If it's too long, go to the bottom of this post.
There's a subtle difference we have to be aware of here:
The squares that gets picked may not be random, but nevertheless unpredictable. These 2 ideas are not the same.
For example, after the game generates a list of nearby stacks, there happens to be more than 1 such stack equidistance to our point of interest, then it simply picks the first one in that list. Then this will not be random. But if the way that the list have been generated is unknown to us AND it's not the same each time (for whatever reason), then it would be unpredictable. We will still have hope of course, if there's a pattern it follows, which is the topic of this discussion.
It might also be a good idea if we dig up some studies that have been done before such as this one on sprites and hydra attacks:
http://help.ordenmira.ru/information/articles/cerberus.htm
Although this may be unrelated, but there's another study that I think we should dig up;
- one where they studied where the the stack would appear in the next round in a conspiratory MG quest or ST when we try to block the entrance by simply putting all of your stacks around that area (they used demon since they can get a lot of stacks to cover large areas with gatings I think).
- one scattered discussions about where the gated stack (if there's some stack on top) or phantom will appear.
These questions are perhaps not about:
Where's the next dreadbane attack chain?
Where's the gated stack will go if it's blocked?
Where will the phantom appear?
Which stack will the berserker pick?
Where will the stack appear if I block the entrace in ST?
or even:
Which direction will my stack travel if there's more than 1 path to the target point?
etc.
But I think it's more about:
How does the game picks the nearest square (either empty square, square with enemy, etc.) from a point?
Gates, phantom, shortest path, they all pick something like this (IF my memory is correct or the game didn't change how this part works):
[x][2][x]
[1][C][4]
[x][3][x]
where C is the centre stack and I don't know about the diagonals and beyond.
Even when moving.
I.e. if you want to move your stack up the board, but there's a stack in the middle, your stack could go around it by going left or going right. But it will always pick left (if possible).
This is useful when dealing with firewall, traps, etc.
I also found out that the stack will also try to move vertically/horizontally >as far as possible<, before making any diagonal moves. I.e.
[_][_][_][_][X][X][X][X][_]
[C][X][X][X][B][_][_][_][X]
[_][_][_][_][_][_][_][_][_]
B is some obstacle, C the creature
right-most X the target tile to move to.
In this case, it will move horizontal until it gets to the obstacle, and only make the 2 diagonal moves when necessary (i.e. diagonal only to avoid the obstacle and to turn back down to reach the goal)
(it also picks up rather than down)
But this is just an aside.
If in doubt, we can always head to the demo and test how the stack moves anyway, that's how I worked these out, but dreadbane and other things are harder to test.
lowest coordinate
I have never heard of this before, but it also make sense with the scheme just above.
What randomr suspected is reasonable horizontal/vertical would be preferred over diagonal since it's known (unless again it's been changed) that horizontal/vertical costs 1 while diagonal costs root(2), hence a little further away.
But this is insufficient since we still don't know between:
S2, S4, S5, S7, ..
S2, S4, S7, S5, ..
S4, S2, S5, S7, .. or
S4, S2, S7, S5, ..
It might be that the game compare x coordinates first, then the y, hence
S4, S2, S7, S5 giving the same scheme as how the gates, phantoms, and other things work.
I have no experience with these ideas + large creature. | Sorry, it was too long
>>>> tl;dr
Note horizontal/vertical costs 1 while diagonal costs root(2)
Then also assume it compares x coordinate first, then y
This gives:
(again, using narutoayan's diagram on post 41)
S4, S2, S7, S5, S1, S6, S3, S8
And this can be extended to further distances away from the centre. | Oops I did one mistake.
(S1)(S2)(S3)
(S4)(B)(S5)
(S6)(S7)(S8)
The first criteria indeed is closest distance, and so yes, S2,S4,S7,S5 will be chosen first. After that, criteria (if equal threat level) must be lesser distance from 1:1, hence S2,S4 will be the choice. As suggested by Sylin, it's possible that lower x-coordinate is chosen first, and so S4 will be attacked.
However this is for beserkers only, not at all for dreadbanes. Their chain seems to follow the order of AI preference, like shooters/casters first etc. | Their chain seems to follow the order of AI preference, like shooters/casters first etc.
i doubt that, im still pretty sure its random.
@sylin: too lazy to read xD | Top-left of Large creature
I also just noticed just as guyb have said in post 18:
If you look carefully at chain lightning, the lighnting ball visual effect is on the top-left corner of the large creature whereas the dreadbane chain attack visual effect is a big lighnting in the middle of the large creature....
------
https://www.lordswm.com/war.php?warid=711191611<=1
First dreadbane attack, it picks the orc chief at the bottom rather than orgi magi at the top.
So I agree with narutoayan, it's not following the usual closest-tile scheme from the tile of the previous hit
https://www.lordswm.com/war.php?lt=-1&warid=711303759
First attack, chain goes up away from dreadbane instead of down.
So it doesn't seem to be picking the one closest to the dreadbane either.
I am also starting to think it's random
But it should still be tested with a carefully controlled combats in ways that will not break any of the rules. Just to see if the chain is different given that >>everything else<< is the same. | | I dont understand one thing. there sure r ppl who know how it works, such as admins, so why on earth arent they trying to answer this? | https://www.lordswm.com/war.php?lt=-1&warid=711213242
look at dreadbane second attack , chain went to recruits instead of wardens , so NO , It doesn't depend on Ai preferance.
https://www.lordswm.com/war.php?lt=-1&warid=711935445 , another
Above combat also shows that berserkers targeted lizards first nstead of dragons and other targeted termganants where lower left troop is ought to get attacked.
https://www.lordswm.com/war.php?lt=-1&warid=710285720
chain went to devils instead of more closer troop(cerberi)
Since you guys may even say that lower troops are not closest and won't get attacked , then-
https://www.lordswm.com/war.php?lt=-1&warid=695681350
https://www.lordswm.com/war.php?lt=-1&warid=711212322 , it also shows that chain simply is unpredictable along with madness attack.
one more- https://www.lordswm.com/war.php?lt=-1&warid=710451082
berserkers went for top troop.
you guys may even say that Berserkers picks troops by ai priority , -
https://www.lordswm.com/war.php?lt=-1&warid=711879072
Berserkers went for vermins and devils instead of blazing hound , As far I know Blazing Hounds have higher priority then these two. | My analysis -
https://www.lordswm.com/war.php?warid=711191611<=1
>>Orc chiefs are shooters, and more prior for AI.
https://www.lordswm.com/war.php?lt=-1&warid=711303759
>>According to my idea, the chain is chosen such that the second troop hit is the more prior one. So, temptresses were chosen over vermins.
https://www.lordswm.com/war.php?lt=-1&warid=711213242
>>Might seem crazy, but retribution rune I think works differently..and so, AI prefers to hit every troop first, and then repeat. (A completely random and probably wrong hypothesis)
https://www.lordswm.com/war.php?lt=-1&warid=711935445
>>Again, retribution rune, so I guess that works differently.
https://www.lordswm.com/war.php?lt=-1&warid=710285720
>>AI preference is not per troop but per stack. Eg. a stack of 5000 farmers is preferred over 1 rogue. So it was in accordance)
https://www.lordswm.com/war.php?lt=-1&warid=711879072
>>In retribution, I doubt if the order matters.
I'm confused about the middle two. So you may be right, unless we figure out another theory haha. But it's hard to believe that these 2 things can be "random". | https://www.lordswm.com/war.php?warid=711191611<=1
>>Orc chiefs are shooters, and more prior for AI.
There were no orc chiefs , there were wardens and recruits . If you look carefully , then you'll see that chain went to recruit instead of warden... | There were no orc chiefs , there were wardens and recruits . If you look carefully , then you'll see that chain went to recruit instead of warden...
this link is correct.
https://www.lordswm.com/war.php?lt=-1&warid=711213242>>AI preference is not per troop but per stack. Eg. a stack of 5000 farmers is preferred over 1 rogue. So it was in accordance)
I have alot of combats to proove that wrong too , But I am too lazy now. It'll require alot of digging up . I am out from this thread with an simple answer.
It's random | Well then let's leave it to Bheem.
Bheem, are you satisfied? :) | | Seems it's a story, you can write a book about.. |
This topic is long since last update and considered obsolete for further discussions.
1|2|3Back to topics list
|
