Trap/door feature branch

[calris]

This thread is garbage, I haven't heard a single person suggest rocket traps.

Or a 'seeker' - the trap sets of a magical orb that makes a beeline for the player - if it hits the player it does massive damage. The player needs to get into a position where the seeker will hit a wall (or a monster) first. I like the idea of 'traps' that can actually be useful if the player think about them and can use them to their advantage.

[debo]

This thread is garbage, I haven't heard a single person suggest rocket traps.

[Nomad]

Edit: forum removed white space, any ideas how to add it?

Put your diagram inside code tags. (<code>Diagram</code>, only with square brackets.)

[Nick]

OK, my plan then is to put the searching skill/visibility question aside for now, do new traps and better placement of traps, and then see how we're looking.

[TJS]

There's a monster on the other side. You can take the direct route and get into melee immediately, but that requires stepping on a trap. Or you could go around (or try to disarm the trap), but that takes longer. Or you could use ranged attacks, which may not be as strong. Some characters may have good trap avoidance and would prefer the straight route. Others may be able to simply "soak" whatever effect the trap has due to their other attributes. But most will have a tricky decision to make.

I can't really understand how this works to be honest. Why would anyone choose to step on a trap just to melee a monster a turn earlier rather than waiting for the monster to come to him? Or just walk towards the monster diagonally avoiding the trap.

I can't imagine any scenario where the player would choose to step on a trap, apart from maybe when it's in a corridor and the player is being chased from the other direction by a monster he can't handle (in which case there's no decision as the player has no choice).

Add to that monsters whose AI knows to take advantage of traps (that they are not affected by, presumably) and you can get some very interesting scenarios indeed.

Hmm would it not be a good idea to get the AI working well without complex trap logic first?

Remember that, if nothing else, traps must be placed as part of an overarching dungeon design. The current haphazard placement (outside of vaults) is dumb even if you like the old trap-detection system.

Maybe give an example of how this works with a few diagrams?

turn 1

###########
#
# @ ^ T
#
###########

etc.

Edit: forum removed white space, any ideas how to add it?

[nikheizen]

When DCSS removed trainable Traps&Doors, they made the chance of detecting traps strictly tied to XL, which I think is a very Angband way of doing things. Angband could have differences between classes/races a la the Searching skill.

If you want a way of scaling the "unpredictability" of traps, you could have the searching skill identify traps. Where a low level character might see a trap: ^, a mid-level character(or low level rogue) might see a rune: ^, and a higher level character(or mid-level rogue) might see a teleportation trap: ^.

That said, I am totally in favour of making traps visible 100% of the time. Letting them affect monsters too is a simple way of making traps incredibly tactically significant(with intelligent "D" "o" "p" "h" "A" etc.... avoiding stepping on them unless they simply don't care). Have fire resist while getting hassled by an orc pack? Quaff !Resist Heat and jump right onto that fireball trap: ^!

Other ideas for traps:

• Directed Teleport: Places you on a specific tile, mainly for use in vaults. (think someone already posted this idea or similar)
• Cave-in: Not a simple earthquake! Appears only in corridors near the exit to the next room and blocks off the path behind you. Great for running away in the corridor, terrible when the rumbling wakes up the D in the room ahead.
• Poison Gas Trigger: Has vents located in the vicinity, stepping on the trigger causes BA_POISON effects around the matching vents.

[Derakon]

I can't see how walking around a trap in the middle of a room can ever be interesting.

There's a monster on the other side. You can take the direct route and get into melee immediately, but that requires stepping on a trap. Or you could go around (or try to disarm the trap), but that takes longer. Or you could use ranged attacks, which may not be as strong. Some characters may have good trap avoidance and would prefer the straight route. Others may be able to simply "soak" whatever effect the trap has due to their other attributes. But most will have a tricky decision to make.

Add to that monsters whose AI knows to take advantage of traps (that they are not affected by, presumably) and you can get some very interesting scenarios indeed.

Remember that, if nothing else, traps must be placed as part of an overarching dungeon design. The current haphazard placement (outside of vaults) is dumb even if you like the old trap-detection system.

[TJS]

The trap is not pointless; it becomes a question of whether you want to try to move through the tile the trap occupies (and either take the time to disarm it or risk being affected by it), or take the time/different route to go around. With current traps this is not a hard decision, because they are small and placed haphazardly, which makes them easy to avoid. But there is no reason why this has to remain the case.

"You have one chance to detect each trap" solves the tedium induced by continuous detection / searching, but involves randomly punishing the player who fails to detect traps. I personally don't like that punishment as it can be completely out of the blue; it effectively devolves into "characters who are bad at searching will randomly have bad stuff happen to them", which doesn't strike me as an interesting game mechanic. Hence why (in the thread Nick linked) I tend to favor traps that are always visible, but hard to avoid -- the goal is to have the player have to make an informed but difficult decision.

Traps should have effects that are interesting challenges to overcome by the player, rather than be considered a "punishment". It's like saying that the player is punished by having to fight monsters on each level.

And having some characters better at dealing with traps means that they have less of those types of situations to deal with, but have more of other types since no character should be good at everything when starting the game.

I can't see how walking around a trap in the middle of a room can ever be interesting.

[molybdenum]

This would probably be a bit further down the line, but what about traps being specific cases of dungeon triggers? A quick example of a trigger might be a jail-like special room: a bunch of cells containing some nasties that are behind a locked, un-pickable door. Somewhere in the room is an unlock trigger. There could also be alarm triggers scattered about (or right on the unlock trigger itself), which would wake up the monsters.

[Derakon]

Anybody else got any ideas for some sort of "learning curve" system that would strike a balance between always auto-detecting all traps from the start vs. unacceptable chance of insta-death in the late game?

I would additionally amend this with "and encourages the player to make interesting decisions beyond choosing whether or not to wear searching gear or pick different race/class combinations".

How about this, as a start: the character can tell that there is a trap somewhere in an area, but not precisely where. We can represent this by declaring a radius R based on the character's trap detection skill (indeed, their trap skill might actually just be R), picking a random tile within R of the trap, and then marking every tile within R of that tile as being potentially trapped. Characters that are better at searching would have a decreased radius; with a radius of 0 the character would always precisely identify the trap location.

Then the player would know that somewhere in that circle is a trap, but they would not know precisely where. Moving through a tile would clearly mark it as trapped/not trapped, as would manually searching, but this becomes rather burdensome with larger values of R (which I expect would cap somewhere around 6 or so), so characters that are bad at searching might well rather just take their chances.

My big concern with this approach is that if the player is not in under particular stress, they may want to turn searching on so they can pinpoint the trap, which would be somewhat tedious. The UI also could get complicated when there's multiple traps about, plus of course the problem of indicating a potentially-trapped tile that also has an item, door, etc. on it.

[Nomad]

DCSS used to have a trainable traps skill which was removed a few versions ago. We should probably figure out why they removed that. My guess is that training a trap skill is one of those things that sounds good in theory but isn't all that interesting in practice.

I'm not sure how well the DCSS implementation would necessarily translate to Angband, though - I don't think I played that particular version, but it's a pretty different game what with the skill system, auto-explore, etc.

Anybody else got any ideas for some sort of "learning curve" system that would strike a balance between always auto-detecting all traps from the start vs. unacceptable chance of insta-death in the late game?

[fizzix]

Maybe you could have a searching skill that's 'trained' by choosing to disarm traps? So at first you're stumbling about near-blind, but every time you successfully disarm a trap it adds x-percent to your search success until you eventually get 100% effective trap-spotting? (And perhaps instead of having a pval rings and amulets of Searching should just convey 100% search skill so weaker class/race combos can keep one equipped until they've trained their natural skill level up enough to do without it.)

DCSS used to have a trainable traps skill which was removed a few versions ago. We should probably figure out why they removed that. My guess is that training a trap skill is one of those things that sounds good in theory but isn't all that interesting in practice.

[Nick]

There are new builds up for Windows and OS X, with some bugfixes by takkaria, a fix to Nomad's assert bug and change so that created traps are immediately apparent.

[Nomad]

An additional idea, which I'm uncertain if I like, is to have a kind of ID by use for traps - so the first time you encounter a trap type you hit it unnoticed, and from then on you can always recognise that type.

Maybe you could have a searching skill that's 'trained' by choosing to disarm traps? So at first you're stumbling about near-blind, but every time you successfully disarm a trap it adds x-percent to your search success until you eventually get 100% effective trap-spotting? (And perhaps instead of having a pval rings and amulets of Searching should just convey 100% search skill so weaker class/race combos can keep one equipped until they've trained their natural skill level up enough to do without it.)

Individual trap types could have some sort of difficulty rating that affects both how hard they are to spot/disarm and how much of a skill bonus you get for disarming them. That way it becomes a tactical choice whether to wear equipment with Searching and slowly build your natural skill by only tackling safe and simple traps, or leap in there and risk tackling every Summoning trap you see to get to 100% search skill early on.

[spara]

About the detection. How about this:

* When there are no monsters in LOS, @ automatically searches and finds traps and secret doors.
* When there are monsters in LOS, @ does not automatically search. If wanted, searching can be turned on, but it slows @ down.

The idea being that @ can only concentrate on one thing at a time. Two things, keeping an eye on a monster and searching takes more time.

Pluses for searching could give percentages for finding traps when there are monsters in LOS.