Stefan
Level: Smitemaster

Rank Points: 2119
Registered: 05-25-2004
IP: Logged

File: Serpent behaviour incompatibility.hold (1.2 KB)
Downloaded 65 times.
License: Other
From: Unspecified

I've found a minor incompatibility issue regarding brained serpents (again... ). I haven't investigated it very much, but I believe that the bug is actually in the AE code, not in JtRH, because it seems as if it's caused by the pathmap not updating. Anyway, I thought I'd report it just in case.

I've attached a hold with a demo to illustrate the bug. Watch closely at how the serpent behaves when Beethro closes the door. (in AE it moves east, but in JtRH it moves south (as it IMO should))

____________________________
0.099³

Rabscuttle
Level: Smitemaster

Rank Points: 2539
Registered: 09-10-2004
IP: Logged

Serpents' heads don't count as walls (unlike their bodies).

I think what is happening in AE is that the pathmap is heading through the serpent's head (then the square NW of its head, and around the top of the room etc). Both squares E and S are the same distance from Beethro according to the brain and brained monsters favour E over S.

I don't know why it doesn't behave the same way in JtRH. My guess would be that a separate pathmap is calculated for each serpent (where its own head counts as a wall)

---

[Last edited by Rabscuttle at 08-21-2005 04:52 AM]

mrimer
Level: Legendary Smitemaster

Rank Points: 5438
Registered: 02-04-2003
IP: Logged

If I understand correctly, the inconsistency is that in AE, brain pathmaps can lead through serpent heads, and in JtRH they don't, right? I agree that I'm not sure whether this should be reverted to the 1.6 behavior in 2.0+.

Stefan
Level: Smitemaster

Rank Points: 2119
Registered: 05-25-2004
IP: Logged

mrimer wrote:
If I understand correctly, the inconsistency is that in AE, brain pathmaps can lead through serpent heads, and in JtRH they don't, right? I agree that I'm not sure whether this should be reverted to the 1.6 behavior in 2.0+.

This behaviour was already changed to conform with AE behaviour (during beta, IIRC). Brain pathmaps can lead through serpent heads in both AE and JtRH. This can be illustrated by the following situation:

W - Wall, S - Serpent head, s - Serpent body, R - roach, B - Beethro

 WWW
WWSWWWWWWWWWWW
  s      sss
  s      s s
B s R    s s
  s      s  
  ssssssss

For the first couple of moves, the roach will head for the serpent head, trying to go through it. This happens in both AE and JtRH.

I think the problem is an ordering issue. When Beethro strikes the orb in AE, the pathmap will not update immediately, making the serpent move as if it were unbrained. In JtRH the pathmap gets updated immediately, and the serpent moves in the correct direction. However, it doesn't quite work for other monsters than serpents, so it may be something completely different.
Edit: Scrap that. I don't know what I'm talking about (and no, I don't really have a better idea yet).

Anyway, if the above is what's causing it, then I think that it should be considered a bug in AE, not JtRH. I just wanted to let you know that the inconsistency exists.

____________________________
0.099³

---

[Last edited by Stefan at 09-06-2005 07:49 PM]

Doom
Level: Smitemaster

Rank Points: 3378
Registered: 07-05-2004
IP: Logged

Stefan wrote:
This behaviour was already changed to conform with AE behaviour (during beta, IIRC). Brain pathmaps can lead through serpent heads in both AE and JtRH. This can be illustrated by the following situation:

W - Wall, S - Serpent head, s - Serpent body, R - roach, B - Beethro

 WWW
WWSWWWWWWWWWWW
  s      sss
  s      s s
B s R    s s
  s      s  
  ssssssss

For the first couple of moves, the roach will head for the serpent head, trying to go through it. This happens in both AE and JtRH.

Is this really the best possible behaviour? Replace the serpent in Stefan's diagram with a rattlesnake. The roach will try to move towards the head all the time. I think that it would be better if the head counted as wall too. Why head for the head anyway? That's the part of the snake where the roach is going to get least likely through. Tail is the part that shortens for both serpents and rattlesnakes.

Cascade
Level: Delver
Rank Points: 46
Registered: 05-28-2005
IP: Logged

Doom wrote:
Is this really the best possible behaviour? Replace the serpent in Stefan's diagram with a rattlesnake. The roach will try to move towards the head all the time. I think that it would be better if the head counted as wall too. Why head for the head anyway? That's the part of the snake where the roach is going to get least likely through. Tail is the part that shortens for both serpents and rattlesnakes.

I'm guessing the reason pathmaps still lead through heads is less about "Let's make brained roaches as effective as possible" and more about "Let's keep the behaviour of the monsters the same as it was in AE so as not to break backwards-compatibility".

____________________________
"Does a cow have Buddha nature?"

mrimer
Level: Legendary Smitemaster

Rank Points: 5438
Registered: 02-04-2003
IP: Logged

Cascade wrote:

Doom wrote:
Is this really the best possible behaviour? Replace the serpent in Stefan's diagram with a rattlesnake. The roach will try to move towards the head all the time. I think that it would be better if the head counted as wall too. Why head for the head anyway? That's the part of the snake where the roach is going to get least likely through. Tail is the part that shortens for both serpents and rattlesnakes.

I'm guessing the reason pathmaps still lead through heads is less about "Let's make brained roaches as effective as possible" and more about "Let's keep the behaviour of the monsters the same as it was in AE so as not to break backwards-compatibility".

That's correct. Brains, of course, aren't perfect -- they might not try to lead monsters through other monsters if they were smarter, for instance. But that's how it has worked since Webfoot DROD, and it will stay that way unless a different kind of brain is introduced in a future version.