Experimental branch: Known Traps


Although the central place for design discussion is ##crawl-dev on freenode, some may find it helpful to discuss requests and suggestions here first.

Swamp Slogger

Posts: 182

Joined: Monday, 2nd July 2018, 16:47

Location: United States

Post Wednesday, 12th December 2018, 16:12

Re: Experimental branch: Known Traps

Magipi wrote:Autoexplore takes you in stupid ways far away into unexplored territory. This is bad in the old system too, but much-much worse in the new system. The reason: alarm traps are more dangerous the farther you are from the stairs - and in the new system alarm traps are a million times more common than in the old system.

Good point. Autoexplore has a poor path choice, diving deep into the unknown, rather than concentric circles outward. I guess it's true that alarm traps are a major punishment for this exploration pattern, and thus any increase in alarm trap frequency will punish autoexplore relative to manual explore. This should be easily fixable by tweaking the frequency. (I'd suggest removing alarm traps, but, maybe one day autoexplore will be smarter and this won't be a problem).

VeryAngryFelid wrote:Consider this: in old system it moves me 5 tiles forward and I meet a monster, in new system I move 5 tiles forward and am teleported/shafted. Why? Because there is no goal to keep the same number of traps triggered per move as before, the goal is to make the game more "fun". With manual exploration I might move to that corridor and reveal just a few tiles per move.
Basically manual exploration becomes more attractive than ever because suddenly direction of your single step means different probabilities of being shafted/teleported which is a big deal for winning.

If I understand you correctly, you mean that with trap-on-sight, optimal play is to have each step reveal as few tiles as possible, thus incentivising manual explore (since autoexplore does not do this). That is a good point; maybe this change is just swapping one manual exploration incentive for another.

Blades Runner

Posts: 593

Joined: Tuesday, 11th December 2018, 19:14

Post Wednesday, 12th December 2018, 16:15

Re: Experimental branch: Known Traps

duvessa wrote:
TheMeInTeam wrote:As I improve as a player that number will grow
As you improve as a player, you will start to recognize that getting shafted or teleported is not a death sentence, and that the percentage of your deaths that were due to traps, rather than your own errors, is 0%.

Also, please read the last sentence of the OP again.

TheMeInTeam wrote:Zot trap absolutely can just kill you outright via paralysis, though your excluding that suggests you agree.
No. Paralysis, like shafting and teleporting, deals precisely 0 damage to you. There are Zot trap effects that deal direct damage, and those can kill you outright, but you need to be aggressively diving to actually reach a Zot trap with max HP that low.


Do you assert that all deaths to these have agency/viable player options without exception? It is logically inconsistent to make such a case unless you make that assertion, but it's a pretty strong assertion...

Even if traps that kill you w/o agency are very rare, these rare cases should not be considered good for the game. Also, they would be the only cases a "theoretical perfect player" would actually die/lose.

The "paralysis does 0 damage" reasoning is disingenuous. As you have no way of knowing the trap is there or mitigating it in advance, there will always be at least some scenarios where you spring such a trap on the same turn you discover multiple monsters.

Finally, you didn't answer my question, which is what meaningful decisions this kind of mechanic brings to crawl. Players play the same way prior to encountering the traps, and then they're presented with a situation where they either have the requisite escape resources or they don't. Shafts/dispersal/alarm will almost never kill someone with no agency, TP and Zot very rarely (with Zot more likely than TP but still rare). What's the point though? If you want to make the game harder, you could easily just reduce the number of generated escape resources similarly to how trunk has reduced digging wand count or how haste was removed.

Good point. Autoexplore has a poor path choice, diving deep into the unknown, rather than concentric circles outward. I guess it's true that alarm traps are a major punishment for this exploration pattern, and thus any increase in alarm trap frequency will punish autoexplore relative to manual explore. This should be easily fixable by tweaking the frequency. (I'd suggest removing alarm traps, but, maybe one day autoexplore will be smarter and this won't be a problem).


I'd rather see the auto-explore algorithm improved somewhat, though I'm not sure how feasible it would be. Maybe use closest known up-stair (without a star indicating unexplored) as an anchor or something?

Swamp Slogger

Posts: 182

Joined: Monday, 2nd July 2018, 16:47

Location: United States

Post Wednesday, 12th December 2018, 16:33

Re: Experimental branch: Known Traps

TheMeInTeam wrote:Even if traps that kill you w/o agency are very rare, these rare cases should not be considered good for the game. Also, they would be the only cases a "theoretical perfect player" would actually die/lose.

Actually, it is entirely possible to be a "theoretical perfect player", have agency, and die. Suppose you have a difficult choice of different actions, each with a different, nonzero probability of death - you have agency. The "theoretical perfect player" would be able to correctly identify and choose the action with the lowest probability of death. And death may occur.

Player agency is present until death is certain (or nearly certain, in which case player agency is so tiny it could be considered not present).
Last edited by stormdragon on Wednesday, 12th December 2018, 16:41, edited 1 time in total.

For this message the author stormdragon has received thanks:
duvessa

Dungeon Master

Posts: 250

Joined: Thursday, 27th November 2014, 19:12

Post Wednesday, 12th December 2018, 16:40

Re: Experimental branch: Known Traps

stormdragon wrote:If I understand you correctly, you mean that with trap-on-sight, optimal play is to have each step reveal as few tiles as possible, thus incentivising manual explore (since autoexplore does not do this). That is a good point; maybe this change is just swapping one manual exploration incentive for another.


That's not how probability works. Every tile gets rolled for a trap effect exactly once*, the rolls are independent and identically distributed; the total number of trap effects experienced to reveal a given set of tiles is independent of the method by which they were revealed. (* Except for those tiles revealed because the player was moved by an exploration-based-trap-effect.)

It is true that if you make a move that reveals more tiles, then on that particular move your chance to get trapped is higher than one that reveals fewer tiles. It's also true that if you make a move that reveals more tiles then on that particular move your chance to reveal a severely out of depth monster that can one-shot you is higher than one that reveals fewer tiles. Unknown tiles are already dangerous, a move that reveals more of them is already more dangerous than one that reveals less of them. This is not an implication of the new trap system (that is only a change to the danger level of unknown tiles). Auto explore is already disadvantaged compared to manual explore because of this.

This thread is getting derailed, I'm interested in playtesting feedback. I'm aware that I've also increased the trap rate and that that would be a player nerf independent of mechanical changes. Please go argue about whether or not traps are good or not elsewhere, the topic has been hashed out endless times on this forum.

For this message the author ebering has received thanks:
duvessa

Ziggurat Zagger

Posts: 8786

Joined: Sunday, 5th May 2013, 08:25

Post Wednesday, 12th December 2018, 20:38

Re: Experimental branch: Known Traps

TheMeInTeam wrote:Do you assert that all deaths to these have agency/viable player options without exception?
Yes
User avatar

Zot Zealot

Posts: 982

Joined: Monday, 29th September 2014, 09:04

Post Wednesday, 12th December 2018, 21:58

Re: Experimental branch: Known Traps

Perhaps traps could be inverted? They only trigger when you reveal previously explored terrain. I'm not sure "trap" would be the right flavouring though.

Dungeon Master

Posts: 625

Joined: Thursday, 23rd October 2014, 03:08

Post Wednesday, 12th December 2018, 22:19

Re: Experimental branch: Known Traps

that sounds terrible for backtracking

For this message the author CanOfWorms has received thanks: 2
chequers, duvessa

Ziggurat Zagger

Posts: 6454

Joined: Tuesday, 30th October 2012, 19:06

Post Wednesday, 12th December 2018, 23:33

Re: Experimental branch: Known Traps

ebering wrote:
stormdragon wrote:If I understand you correctly, you mean that with trap-on-sight, optimal play is to have each step reveal as few tiles as possible, thus incentivising manual explore (since autoexplore does not do this). That is a good point; maybe this change is just swapping one manual exploration incentive for another.


That's not how probability works. Every tile gets rolled for a trap effect exactly once*, the rolls are independent and identically distributed; the total number of trap effects experienced to reveal a given set of tiles is independent of the method by which they were revealed. (* Except for those tiles revealed because the player was moved by an exploration-based-trap-effect.)

This does however further incentivize "clearing" a level (killing as many of the creatures on it as is reasonable before descending) while revealing fewer than "all the tiles" (as every revealed tile has associated risk)

Furthermore it incentivizes killing monsters *in advance of* revealing as many tiles as possible, as many traps are only dangerous if monsters are around (paralysis, teleport, alarm)

I don't think that this is *necessarily* bad, but *contention* is that the amount of manual micromanagement required to intentionally reveal as few squares as possible and shout to bring creatures into known areas to kill them all prior to revealing all the tiles (or not revealing them and simply ignoring many of the tiles in a level) is at least as bad as the original problem that this solves.

I'm not sure that's true, but since this is all theorycrafting on expected "optimal play" (Which we've seen at least one example of under the old system) I'd like to see someone try to execute an "optimal" trap avoiding game in the new system to compare the two for tediousness (I suspect the new system, while imperfect, is an improvement over the old one).

It does seem to me that this change has traps as reinforcing what was already optimal play (luring monsters into cleared areas) which may or may not be a good idea.
Spoiler: show
This high quality signature has been hidden for your protection. To unlock it's secret, send 3 easy payments of $9.99 to me, by way of your nearest theta band or ley line. Complete your transmission by midnight tonight for a special free gift!

Ziggurat Zagger

Posts: 8786

Joined: Sunday, 5th May 2013, 08:25

Post Wednesday, 12th December 2018, 23:51

Re: Experimental branch: Known Traps

Siegurt wrote:This does however further incentivize "clearing" a level (killing as many of the creatures on it as is reasonable before descending) while revealing fewer than "all the tiles" (as every revealed tile has associated risk)
That's where floor items come in! Each unexplored tile could have an item on it, and items are good, good enough that it's not optimal to leave squares unexplored.

Item detection does break this by telling you which squares have no items in them, but item detection can just be removed; would anyone miss it?

Cocytus Succeeder

Posts: 2173

Joined: Saturday, 2nd February 2013, 09:52

Post Thursday, 13th December 2018, 01:26

Re: Experimental branch: Known Traps

duvessa wrote:
Siegurt wrote:This does however further incentivize "clearing" a level (killing as many of the creatures on it as is reasonable before descending) while revealing fewer than "all the tiles" (as every revealed tile has associated risk)
That's where floor items come in! Each unexplored tile could have an item on it, and items are good, good enough that it's not optimal to leave squares unexplored.

The right strategy seems to be to reveal as few tiles as possible until most of the floor is more-or-less cleared, then autoexplore to reveal the rest. When there are very few monsters left, alarm and tele traps don't matter any more.

If the goal of the change was to make UV4's tile-marking shenanigans obsolete, that was not achieved at all.

Ziggurat Zagger

Posts: 8786

Joined: Sunday, 5th May 2013, 08:25

Post Thursday, 13th December 2018, 01:39

Re: Experimental branch: Known Traps

It was achieved. Marking stepped-on squares, be they stepped on by players or monsters, no longer confers any benefit whatsoever. Some squares are still different from others, nobody's denying that. But as far as traps go, the game already keeps track of which tiles are unexplored.

For this message the author duvessa has received thanks:
tealizard
User avatar

Vaults Vanquisher

Posts: 454

Joined: Thursday, 1st November 2018, 02:33

Post Thursday, 13th December 2018, 01:59

Re: Experimental branch: Known Traps

Yes, clearly the goal was achieved. If there is a problem it is the narrowness of the goal.

Yet the work done and feedback on it suggests the way forward. In the new system, the exploration trap rate is essentially a numerical parameter (as I understand it). The balance of feedback suggests the number is too high. Perhaps in the fullness of time, its value will reach zero.
This is where mechanical excellence and one-thousand four-hundred horsepower pays off.

Cocytus Succeeder

Posts: 2173

Joined: Saturday, 2nd February 2013, 09:52

Post Thursday, 13th December 2018, 10:55

Re: Experimental branch: Known Traps

duvessa wrote:It was achieved. Marking stepped-on squares, be they stepped on by players or monsters, no longer confers any benefit whatsoever. Some squares are still different from others, nobody's denying that. But as far as traps go, the game already keeps track of which tiles are unexplored.

The game keeps track of which tiles are unexplored, but the question is not that, it is which tiles will reveal new tiles and which ones won't. Yes, you can calculate that, although it is probably way over the head of 99% of players given that LOS* is a terribly complicated and non-intuitive stuff.

Edit: *Sorry, not LOS, but line of fire. I mixed it up. So probably revealing new tiles is obvious at first sight for a lot of players. Not me, though.

Ziggurat Zagger

Posts: 4432

Joined: Friday, 8th May 2015, 17:51

Post Thursday, 13th December 2018, 13:27

Re: Experimental branch: Known Traps

Siegurt wrote:I don't think that this is *necessarily* bad, but *contention* is that the amount of manual micromanagement required to intentionally reveal as few squares as possible and shout to bring creatures into known areas to kill them all prior to revealing all the tiles (or not revealing them and simply ignoring many of the tiles in a level) is at least as bad as the original problem that this solves.

I'm not sure that's true, but since this is all theorycrafting on expected "optimal play" (Which we've seen at least one example of under the old system) I'd like to see someone try to execute an "optimal" trap avoiding game in the new system to compare the two for tediousness (I suspect the new system, while imperfect, is an improvement over the old one).

It does seem to me that this change has traps as reinforcing what was already optimal play (luring monsters into cleared areas) which may or may not be a good idea.


I am judging any game changes from the point of view how they will affect me or average player. Basically I don't care that some freak will stop writing down all stepped tiles (probably, nothing stops them from keeping doing it despite it does not make any sense in new system), I do care about autoexplore becoming more dangerous than before and higher number of shafts in general.

Edit. My point is that while optimal play considerations are important they should not be more important than considerations how average or new players play the game.
Underestimated: cleaving, Deep Elf, Formicid, Vehumet, EV
Overestimated: AC, GDS
Twin account of Sandman25

Blades Runner

Posts: 593

Joined: Tuesday, 11th December 2018, 19:14

Post Thursday, 13th December 2018, 16:19

Re: Experimental branch: Known Traps

I'm happy to be wrong regarding Zot paralysis.

On playing with this a bit more:

- I have to echo the sentiment about alarm traps and autoexplore. Autoexplore has been something of a false choice for a long time compared to careful play, but this exacerbates it. Perhaps the answer is the algorithm for it though, rather than making trap effects contingent on it. Still a significant tedium question though.
- Shafts are super common, I've been shafted at least 3 times in most games on this trunk since patch update. This assumes you live long enough to get through early areas or average will be less, but 2 shafts in dungeon alone is common in these runs. IMO a bit too frequent, but these less lethal than other traps too and can be used to isolate monsters.
- I do not like the possibility of downstairs --> teleported into monsters before moving. It does break the 1 move grace you get on first entering a floor in practice and is disproportionately dangerous compared to most everything else the new system introduces.

Actually, it is entirely possible to be a "theoretical perfect player", have agency, and die. Suppose you have a difficult choice of different actions, each with a different, nonzero probability of death - you have agency. The "theoretical perfect player" would be able to correctly identify and choose the action with the lowest probability of death.


While I disagree with some rationale in this thread and think the "perfect player will never die to trap RNG outright" is exaggerated, there are very few situations where "perfect play" enters into a scenario you describe, traps or not. A lot of agency in crawl is such that mistakes take a number of turns to manifest, sometimes a great number, creating some bias towards concluding a death was to RNG rather than a mistake (common in roguelikes).

It does seem to me that this change has traps as reinforcing what was already optimal play (luring monsters into cleared areas) which may or may not be a good idea.


Agreed, to the point where the current scenario buffs controlled noise to pull monsters from further away when ready.

Edit. My point is that while optimal play considerations are important they should not be more important than considerations how average or new players play the game.


I'd rather hedge against tedium in general. Optimal or not, manually exploring + generating noise to pull monsters is kind of iffy on the enjoyment...but it's strictly safer than going into the unknown with more monsters on the floor rather than less. While true for both old and new trap systems, it's still a tedium + questionable incentive partially created by both. As pointed out, revealing tiles is already dangerous due to monster presence, does the game need to further incentivize as much luring as possible with traps *in addition*, considering it's already an alpha strategy even when traps don't exist?

That new players won't do this and will thus be worse off is also not ideal, but IMO it's something of a degenerate incentive regardless and I'm not sure I like how it balances out stealth/stabbing based builds with other variants.

Dungeon Dilettante

Posts: 1

Joined: Thursday, 15th November 2018, 15:21

Post Thursday, 13th December 2018, 16:44

Re: Experimental branch: Known Traps

After trying the changes on trunk, I really feel like the alarm traps are disproportionately dangerous on the first floor of Lair and its branches. I can't say I ever worried about getting marked in these places before, and now it's affecting the way I explore and fight. Am I correct in understanding that every step on a wide open floor is multiple times more likely to generate a trap effect compared to walking down a corridor, since it reveals more tiles?

If I could make one change, I guess I would look for a way to restrict some of the trap rolls within Lair to corridors, vaults, and naturally-occurring chokepoints. This would make more sense flavor-wise, and would make getting marked feel more like a tactically interesting situation than a random kick in the balls. The current implementation feels too Xom-like.

For this message the author bees has received thanks:
sdynet

Slime Squisher

Posts: 392

Joined: Sunday, 11th September 2016, 17:21

Post Wednesday, 19th December 2018, 02:31

Re: Experimental branch: Known Traps

I liked the dispersal traps until I stabbed a sleeping Sonja who happened to be generated on one, meaning I couldn't collect the loot.

I approve of the revealed traps idea in general, but at this point Ashenzari should really be either reworked or removed. A big part of the god's identity is providing information, and both equipment ID and trap ID are now largely given away for free.

Dungeon Master

Posts: 250

Joined: Thursday, 27th November 2014, 19:12

Post Wednesday, 19th December 2018, 16:48

Re: Experimental branch: Known Traps

Do you have the version string for when that happened? Commit 0.23-a0-580-gf63cd86 was made 4 days ago to prevent that sort of thing.

Slime Squisher

Posts: 392

Joined: Sunday, 11th September 2016, 17:21

Post Wednesday, 19th December 2018, 17:12

Re: Experimental branch: Known Traps

My folder name contains 0.23-a0-577, so I guess I should update. Thanks.
User avatar

Mines Malingerer

Posts: 42

Joined: Wednesday, 18th July 2018, 23:51

Post Tuesday, 1st January 2019, 19:03

Re: Experimental branch: Known Traps

Since monsters don't trigger traps anymore if they're out of LOS, it's still possible to kill them on traps (like dispersal) with allies who can attack out of LOS (Yred servants, enslaved monsters, etc.)

Image
This is on trunk on CKO

(FWIW, I really think that monsters not triggering traps out of LOS is a bad idea, for many more reasons than Zot:5)
EthnicCake on CKO. It's an anagram.

For this message the author kitchen_ace has received thanks:
duvessa

Ziggurat Zagger

Posts: 8786

Joined: Sunday, 5th May 2013, 08:25

Post Tuesday, 1st January 2019, 19:08

Re: Experimental branch: Known Traps

Since there's code for pushing items out of doors now, would it be hard to push items out of permanent Zot/tele/dispersal traps?

Lair Larrikin

Posts: 16

Joined: Tuesday, 7th November 2017, 11:29

Post Monday, 4th February 2019, 08:12

Re: Experimental branch: Known Traps

Magipi wrote:
duvessa wrote:
Siegurt wrote:This does however further incentivize "clearing" a level (killing as many of the creatures on it as is reasonable before descending) while revealing fewer than "all the tiles" (as every revealed tile has associated risk)
That's where floor items come in! Each unexplored tile could have an item on it, and items are good, good enough that it's not optimal to leave squares unexplored.

The right strategy seems to be to reveal as few tiles as possible until most of the floor is more-or-less cleared, then autoexplore to reveal the rest. When there are very few monsters left, alarm and tele traps don't matter any more.

If the goal of the change was to make UV4's tile-marking shenanigans obsolete, that was not achieved at all.



I just want to quickly comment that before the new trap system, revealing as few tiles as possible per move is also optimal, to reduce the probability of multiple monsters coming into LOS at once. This isn't hypothetical: You can see zxc's videos of streaking games where he tries to avoid moving diagonally in D:1 for exactly this reason.
Yes, I play offline, tiles and stable. I am not l33t enough :_(

For this message the author Nasst has received thanks: 2
duvessa, stormdragon

Temple Termagant

Posts: 13

Joined: Wednesday, 11th July 2012, 04:29

Post Monday, 11th February 2019, 05:47

Re: Experimental branch: Known Traps

I know nothing I say seems to matter but I feel compelled to say my last bit on this since I'm so fed up with the game.

I was ambivalent towards this change at first, but the more I played with it the less I liked it. It just feels crummy. I'm not opposed to stuff that puts players in a dangerous situation but "traps" that trigger purely by LOS exploration doesn't work. It creates the weird scenarios where you can be standing still and somehow trigger a trap because something blocking your vision was altered in some manner. Say a Deep Troll Earth Mage digs open a wall with a bunch of enemies behind it! Boy that sucks, but also this reveals a bunch of LOS so now you triggered an alarm trap just by standing still. Feels really awkward. Dunno how you would explain this in a way that makes sense to a new player. It still doesn't make sense to me. Furthermore, it just feels cheap to be teleported, marked, or dispersal blinked on stairs before you've had a chance to even gather your bearings. Its really disorienting to go down to a new level and somehow end up on the other side of it before you even got a glimpse of where you entered from.

Overall, I just wasn't enjoying how this worked as I played. But the tipping point for me was playing Tomb. Tomb is just sadistic with these traps and not in a fun way. The sheer glut of enemies that flood your screen constantly means that "enemies triggering traps" feels less like a tactical concern and become more inevitable. If I'm on like D:5 and see an alarm trap I can think about its placement relative to the enemies I'm fighting. In Tomb the screen is choked with summons and death scarabs with erratic movement. Enemies trigger traps constantly, and the new dispersal trap is lethal here. Being blinked in such a positioning dependent area is awful. I got blinked into the center chamber, marked, and netted in just a few moves. This area was "balanced" around being able to stair dance originally. The changes that weakened stairdancing made it harder but less tedious, so I was okay with them. Now it just feels excessively cruel. Even as someone looking for a challenge, I do not enjoy this. The density of traps in Tomb is simply too much now.

I realize the old system wasn't perfect but it was far more tolerable. I don't know exactly how you'd fix this one. Some things that would make it more acceptable to me;

-remove the chance for traps to trigger on stairs. Inconsistent? Yeah, but this whole system is inconsistent in its logic. It just feels bad.
-mark is an MR check. Makes it feel like something you can actually do something to prevent rather than a guaranteed application.
-Do SOMETHING about Tomb. I don't know what, but the most low-effort-effective option would be to reduce trap density and make dispersal traps there one-use. I think dispersal traps are interesting as a concept and them being permanent is fine in most of the game, but you have to do something to accommodate for the sheer spam that Tomb throws at you

Probably done with this version of the game if this doesn't change. I've put up with a lot of changes I don't like but this is going too far.

Zot Zealot

Posts: 1004

Joined: Thursday, 16th August 2018, 21:19

Post Monday, 11th February 2019, 16:41

Re: Experimental branch: Known Traps

I agree the exploration based traps feel crummy.
Maybe it would feel more flavourful if the game picked a tile which "triggered" the trap by being seen, and put a little robot on it. Kind-of like having traps which are triggered by being seen. It could have a description like "This ancient contraption just marked you. Its magic spend, it is now just an empty husk."

Ziggurat Zagger

Posts: 6454

Joined: Tuesday, 30th October 2012, 19:06

Post Monday, 11th February 2019, 16:51

Re: Experimental branch: Known Traps

I don't mind the exploration-basedness of it (prior traps were exploration based, just on tiles stepped on rather than tiles revealed), what feels bad to *me* is when I am standing still in a chokepoint fighting things, and all of a sudden a monster appears from offscreen and drops a trap on my head, all of them are pretty annoying when this happens.

What makes this really awful is that it is nearly always *possible* to not fight in LOS of traps, however it means even more tedium while you lure things into good fighting squares where a trap isn't in LOS. As a result the correct choice is a tedious, optimal, and almost-never dangerous. There are situations where you are forced to fight in LOS of traps (say you haven't found a decent place on the level that's trap free yet) however the rare situation where those cause an interesting result isn't worth the tedium added. To make this sort of thing even more annoying, it's very infrequent for something to pop onto the screen from an unexpected place, so *nearly always* you can safely fight with a trap onscreen, knowing the monsters you'll fight won't come anywhere close to it, it's only very very rarely where fighting with a trap in LOS, but not anywhere near the monsters you're fighting will be a problem, so not only is "never fight with traps in LOS" tedious, it's an optimization that is very rarely impactful.

At least Zot traps were limited to very few places in the game, and you couldn't optimize about keeping them out of your LOS if you hadn't seen them yet.
Spoiler: show
This high quality signature has been hidden for your protection. To unlock it's secret, send 3 easy payments of $9.99 to me, by way of your nearest theta band or ley line. Complete your transmission by midnight tonight for a special free gift!

For this message the author Siegurt has received thanks: 4
chequers, duvessa, nago, svendre

Temple Termagant

Posts: 13

Joined: Wednesday, 11th July 2012, 04:29

Post Wednesday, 13th February 2019, 03:55

Re: Experimental branch: Known Traps

It was pointed out that under the new system, the optimal thing to do is lug around a piece of -tele equipment. Put that -tele gear on everytime you go down a staircase. Repeat for every staircase on the level. The cost for doing this is little if you don't care about turncount and potentially saves you consumables/death since you minimize the risk of being instantly teleported into a potentially dangerous situation the moment you enter the level.

For this message the author Internet Kraken has received thanks: 2
duvessa, VeryAngryFelid

Ziggurat Zagger

Posts: 8786

Joined: Sunday, 5th May 2013, 08:25

Post Wednesday, 13th February 2019, 04:36

Re: Experimental branch: Known Traps

For whatever it's worth, wearing -Tele while exploring was generally best in the old trap system as well.
User avatar

Zot Zealot

Posts: 982

Joined: Monday, 29th September 2014, 09:04

Post Wednesday, 13th February 2019, 05:45

Re: Experimental branch: Known Traps

The system is great in Zot:5 where having things go spectacularly wrong is part of the design. But I can't say I'm a huge fan of normal traps being zot-ified. I'd rather have more shafts, which I think lead to exciting situations.

Ziggurat Zagger

Posts: 4432

Joined: Friday, 8th May 2015, 17:51

Post Wednesday, 13th February 2019, 07:41

Re: Experimental branch: Known Traps

chequers wrote:The system is great in Zot:5 where having things go spectacularly wrong is part of the design. But I can't say I'm a huge fan of normal traps being zot-ified. I'd rather have more shafts, which I think lead to exciting situations.


I know devs don't want to add easier mode to the game but can they add a hard mode which creates more shafts, for instance?
Underestimated: cleaving, Deep Elf, Formicid, Vehumet, EV
Overestimated: AC, GDS
Twin account of Sandman25

Halls Hopper

Posts: 81

Joined: Monday, 18th March 2019, 22:11

Post Wednesday, 20th March 2019, 00:54

Re: Experimental branch: Known Traps

I figured I would share my thoughts on the trap changes for far. Overall I do like that shaft and teleport traps are more frequent through the game, but am generally not a fan of alarm traps do to its interactions with auto explore and it's tendency to take you away from the stairs. The main things I don't like about the change otherwise are triggering traps when descending (if that still occurs) or by revealing tiles when digging/by deep troll earth mages.

The one place where I don't like the quantity of traps is in Zot 5. With the latest changes, it seems like the safest way to clear it now is to make lots of noise to clear out the lungs. In addition, monsters triggering certain traps like net traps are unintuitive and it is likely a player hasn't encountered this before Zot 5. I would support reducing trap density in these lungs and mechanics to prevent noise cheese to ensure that a variety of approaches to Zot 5 can be taken (I know that I wouldn't ninja the orb with the latest changes). If difficulty is of concern in Zot 5, and I think there would be more interesting ways of making Zot harder like removing stairs and having only hatches (3 up and 3 down per level).

As for dispersal traps, I find them less interesting that shafts, tele and alarm traps in general and it can be quite annoying having them continuously trigger.

Edit: Oh and I generally am not a fan of the -Tele attribute on artifacts. I think teleports should either occur regardless of the artifact or a stasis amulet/ring should be introduced back into the game.

For this message the author acoolguy has received thanks: 3
duvessa, mgcat, nago

Ziggurat Zagger

Posts: 8786

Joined: Sunday, 5th May 2013, 08:25

Post Wednesday, 20th March 2019, 01:52

Re: Experimental branch: Known Traps

Net traps really ought to just be put out of their misery

Blades Runner

Posts: 593

Joined: Tuesday, 11th December 2018, 19:14

Post Wednesday, 20th March 2019, 14:55

Re: Experimental branch: Known Traps

The main things I don't like about the change otherwise are triggering traps when descending (if that still occurs) or by revealing tiles when digging/by deep troll earth mages.


Both of these still happen. It's not much different than getting shafted, unless you trigger a tele trap when first entering a branch. I don't like it either, it's unintuitive to unspoiled players and the present setup for traps has too many janky interactions/inconsistencies.

Dungeon Master

Posts: 250

Joined: Thursday, 27th November 2014, 19:12

Post Wednesday, 20th March 2019, 16:09

Re: Experimental branch: Known Traps

TheMeInTeam wrote:it's unintuitive to unspoiled players


This claim has been levied by many critics of new traps, and I really don't get it.

What is unintuitive about a clever minion of Zot putting a trap at a critical chokepoint (of which none is more critical than the stairs)? What is unintuitive about a less-than-clever minion of Zot setting up a trap badly, so that it doesn't trigger the moment you step on it, only after you've made a ruckus in that square?

Sure, you can "look behind the curtain" and see that the lore explanation I just gave isn't the implementation detail, but that's a spoiler. An unspoiled player is going to run on intuition and imagination. And it's not a far stretch of the imagination that someone might booby trap stairs in a dungeon meant to kill the player and keep them out!

For this message the author ebering has received thanks: 2
acoolguy, duvessa

Blades Runner

Posts: 593

Joined: Tuesday, 11th December 2018, 19:14

Post Wednesday, 20th March 2019, 17:51

Re: Experimental branch: Known Traps

ebering wrote:
TheMeInTeam wrote:it's unintuitive to unspoiled players


This claim has been levied by many critics of new traps, and I really don't get it.

What is unintuitive about a clever minion of Zot putting a trap at a critical chokepoint (of which none is more critical than the stairs)? What is unintuitive about a less-than-clever minion of Zot setting up a trap badly, so that it doesn't trigger the moment you step on it, only after you've made a ruckus in that square?

Sure, you can "look behind the curtain" and see that the lore explanation I just gave isn't the implementation detail, but that's a spoiler. An unspoiled player is going to run on intuition and imagination. And it's not a far stretch of the imagination that someone might booby trap stairs in a dungeon meant to kill the player and keep them out!


An unspoiled player will not:

- Anticipate that revealing a new tile will spring an already-revealed/passed trap (someone who has never played was watching me out of curiosity and was extremely confused as to what happened).
- Reasonably expect "trapped stairs" when they never observe traps on stairs, considering stairs are absolutely mandatory to progress in a way other tiles are not
- Immediately grasp the inconsistencies of what happens when something physically steps on a revealed trap (true for both player and monster stepping on the trap), or why these inconsistencies exist
- Guess in advance that traps on other tiles can be sprung on them while stationary without monsters stepping on the actual trap
- Know that traps can be sprung by stepping on tiles that were previously walls
- Realize that autoexploring is a trap itself, prior to experiencing or at least seeing alarm traps
- See a trap, and immediately avoid revealing further tiles while the trap is in LoS to the point of making noise or exploring elsewhere entirely. Besides, this is tedious.

Considering that traps largely lack agency *to avoid negative effects* for practical play until after they're triggered, a good comparison is hell effects. Every once in a while you just get RNG-smacked by something negative...because you're in hell so why not (hell effects did not confuse the same aforementioned spectator). The only things traps offer over hell effects is that *sometimes* you will know which effect is coming, and that they offer the opportunity to (for some traps) intentionally trigger them to player benefit. The latter existed under the previous trap model as well.

Finally, there's at least some disconnect when a player character is triggering a *known* trap unintentionally at the conceptual level. You can justify this away with some complicated explanations of what might be going on, but it's still an awkward interaction to get sucked into that hole 5 squares behind you when compared to the "you stepped on this poorly supported area and it crumbles so you fall". Even you are admitting the implementation details do not match intuitive lore explanation, and that does have gameplay consequences.

For this message the author TheMeInTeam has received thanks: 2
acoolguy, duvessa

Halls Hopper

Posts: 81

Joined: Monday, 18th March 2019, 22:11

Post Wednesday, 20th March 2019, 22:49

Re: Experimental branch: Known Traps

ebering wrote:
TheMeInTeam wrote:it's unintuitive to unspoiled players


This claim has been levied by many critics of new traps, and I really don't get it.

What is unintuitive about a clever minion of Zot putting a trap at a critical chokepoint (of which none is more critical than the stairs)? What is unintuitive about a less-than-clever minion of Zot setting up a trap badly, so that it doesn't trigger the moment you step on it, only after you've made a ruckus in that square?

Sure, you can "look behind the curtain" and see that the lore explanation I just gave isn't the implementation detail, but that's a spoiler. An unspoiled player is going to run on intuition and imagination. And it's not a far stretch of the imagination that someone might booby trap stairs in a dungeon meant to kill the player and keep them out!


One of the main things I like about dungeon crawl is that it does a good job introducing mechanics early in the game that maintain relevance through out the game (orc priests displaying smiting, orc wizards showing hexes like confusion, fast enemies like adders, hard hitter like ogres, etc). The main thing with these trap changes is that it doesn't effectively do this. In an average game, a trap won't trigger when an npc digs or goes down a stair, but it can. A player may not see an instance of a monster triggering a trap negatively effecting them (if they do, it may honestly be Zot 5). Even if they do, they may suspect it doesn't work on them for a particular type of trap (alarm vs net as an example). A player is at least able to see exploration triggering shaft, alarm and tele traps even early in the game.

Zot Zealot

Posts: 1004

Joined: Thursday, 16th August 2018, 21:19

Post Sunday, 24th March 2019, 21:28

Re: Experimental branch: Known Traps

I just had the experience of "you open a door" -> "you fall through a trap". It felt ridicules :lol:

I don't think it's spoilery, and I don't see any way to fix it.
User avatar

Vaults Vanquisher

Posts: 454

Joined: Thursday, 1st November 2018, 02:33

Post Sunday, 24th March 2019, 22:29

Re: Experimental branch: Known Traps

Looking at this conversation about traps triggered by exploration (or going down stairs, or opening a door, or having walls destroyed near you), the defense that it makes sense for somebody to "booby trap stairs" kinda makes you think... what if traps were actually attached to stairs?

As for exploration-traps, it has a lot of the same problems as tile-traps. It provides further incentive not to move into unexplored territory in combat, it encourages manual exploration, and so on. It only solves a very narrow problem, tile bookkeeping, while throwing to the wind any conceptual coherence traps might have. As with many things in crawl, the only reason to have traps in the first place was thematic and you've destroyed the thematic element with these mechanics. It would have been better to remove tile-traps and replace them with nothing. Easier too.

Now stair-traps are a different story. A staircase that collapses and strands the player on the lower floor or makes them fall to an even lower floor makes perfectly good thematic sense. A staircase that closes behind the player and marks them makes plenty of sense. A staircase that surprisingly only goes one way makes plenty of sense. These traps have the additional benefit that they partially address an actual problem in the game: Stairs.
This is where mechanical excellence and one-thousand four-hundred horsepower pays off.

For this message the author tealizard has received thanks: 4
duvessa, Kismet, sdynet, svendre

Halls Hopper

Posts: 81

Joined: Monday, 18th March 2019, 22:11

Post Monday, 25th March 2019, 01:50

Re: Experimental branch: Known Traps

I think removal of stairs in certain places is probably one of the most interesting changes one of the changes I like the most about hellcrawl (along with removing food). I am not sure if I like it quite as much to come from a trap (rather than a set of hatches and now stairs on two levels) though. I think both the trap and the hatch are fine, but would prefer it being known to be of a higher frequency than being unknown.

Ziggurat Zagger

Posts: 4432

Joined: Friday, 8th May 2015, 17:51

Post Monday, 25th March 2019, 04:48

Re: Experimental branch: Known Traps

Traps are stupid. Just yesterday I killed a monster on O:2 and decided to butch its corpse while waiting for regen (was at 90% HP, 70 HP as TeCj), teleportation trap landed me in the middle of main vault with 4 orc sorcerers, 3 orc high priests, 4 orc priests etc. Of course I had to use my only scroll of blinking immediately. What was interesting here? Waiting for 100% hp/mp before movement? Decision about chances to get paralyzed/bolted/smited instantly? Reading scroll of blinking???
Underestimated: cleaving, Deep Elf, Formicid, Vehumet, EV
Overestimated: AC, GDS
Twin account of Sandman25
User avatar

Zot Zealot

Posts: 982

Joined: Monday, 29th September 2014, 09:04

Post Monday, 25th March 2019, 05:26

Re: Experimental branch: Known Traps

It sounds like waiting until you were at 100% hp wouldn't have changed anything?

Ziggurat Zagger

Posts: 4432

Joined: Friday, 8th May 2015, 17:51

Post Monday, 25th March 2019, 07:39

Re: Experimental branch: Known Traps

chequers wrote:It sounds like waiting until you were at 100% hp wouldn't have changed anything?


Right, in this particular case I would still have to burn the blinking scroll. Yet in other cases (when there are less monsters after shafting/teleportation) resting first might help.
Underestimated: cleaving, Deep Elf, Formicid, Vehumet, EV
Overestimated: AC, GDS
Twin account of Sandman25

Blades Runner

Posts: 616

Joined: Thursday, 25th October 2012, 03:19

Post Monday, 25th March 2019, 17:24

Re: Experimental branch: Known Traps

If this whole thing is just to fix a problem with a few edge cases of players marking tiles, why not just limit the total number of tiles a player can mark to like 5 or something. That's what, 3 lines of code and maybe 2 minutes of work? If that isn't a good enough fix because someone could write an external program to accomplish it (or something equally ridiculous), really who cares?

What will be said if someone writes a program to calculate the odds of a trap triggering with each move, and then uses it as a tool to play more "optimally"?

In order to stamp out all traces of "optimal play", which have been decreed as "objectively not fun" when fun isn't an objective term, you'd have to remove fsim, wizmode, probably you'd have to remove access to the source code and on and on. So really what's the point? The net changes seem negative.

I use a combination of manual exploration and autoexplore which is guided by human intuition, but probably follows some kind of logic if I tried to analyze it. This change looks like it would punish the portion of using autoexplore, and just be generally annoying. Winning/losing isn't guaranteed in a roguelike for sure, but I think the best a dungeon crawler can be is be random with a high influence of skill. This change seems to support more chaotic results with skill playing less of a role instead of more.

If individuals don't feel they are getting enough fun from pure random factors, they can already worship Xom, which seems almost like what the new traps system is pushing onto every game.

For this message the author svendre has received thanks:
TheMeInTeam

Tartarus Sorceror

Posts: 1739

Joined: Tuesday, 13th March 2012, 02:48

Post Monday, 25th March 2019, 20:59

Re: Experimental branch: Known Traps

petercordia wrote:I just had the experience of "you open a door" -> "you fall through a trap". It felt ridicules :lol:


Is this possible with current trap system:

1. Read fog scroll.
2. Walk a couple steps forward into the fog.
3. Wait for fog to dissipate
4. Trap triggers!

Dungeon Master

Posts: 625

Joined: Thursday, 23rd October 2014, 03:08

Post Monday, 25th March 2019, 21:04

Re: Experimental branch: Known Traps

Rast wrote:
petercordia wrote:I just had the experience of "you open a door" -> "you fall through a trap". It felt ridicules :lol:


Is this possible with current trap system:

1. Read fog scroll.
2. Walk a couple steps forward into the fog.
3. Wait for fog to dissipate
4. Trap triggers!

Yes.

For this message the author CanOfWorms has received thanks:
Rast

Ziggurat Zagger

Posts: 6454

Joined: Tuesday, 30th October 2012, 19:06

Post Tuesday, 2nd April 2019, 17:22

Re: Experimental branch: Known Traps

So I'm not sure if this is intentional, or a bug:
  Code:
Done exploring.
Done exploring.
Done exploring.
Okay, then.
You eat one of the 22 rations. That ration really hit the spot! You set off the
alarm!
A sentinel's mark forms upon you.

No monsters in LOS, no tiles revealed, I ate a ration (didn't move) and was marked by an alarm trap, very confusing as my understanding was that traps should only be triggered by revealing tiles or a monster stepping on an already known one.

I have never seen that happen before, so it's pretty weird to me.
Spoiler: show
This high quality signature has been hidden for your protection. To unlock it's secret, send 3 easy payments of $9.99 to me, by way of your nearest theta band or ley line. Complete your transmission by midnight tonight for a special free gift!
User avatar

Shoals Surfer

Posts: 287

Joined: Friday, 19th August 2016, 21:21

Post Tuesday, 2nd April 2019, 18:25

Re: Experimental branch: Known Traps

Was there a fog area in los, like a dith altar?
make food great again

Ziggurat Zagger

Posts: 6454

Joined: Tuesday, 30th October 2012, 19:06

Post Tuesday, 2nd April 2019, 19:17

Re: Experimental branch: Known Traps

pedritolo wrote:Was there a fog area in los, like a dith altar?

Nope, in orc, after shopping, (also I did autoexplore right before I ate the ration and it said 'Done exploring.' So I sort of assumed there were no more squares to reveal)

It's *possible* that I monkeyfingered and hit some other key after the eating, but I don't know what it could possibly be that could trigger a trap on a completely explored level with no monsters which would generate no message in the log (such that triggering the trap would show up on the same line in the message history as eating the ration)

FWIW the ration's inventory letter was 'f' so that would've been where my finger was at the time.
Spoiler: show
This high quality signature has been hidden for your protection. To unlock it's secret, send 3 easy payments of $9.99 to me, by way of your nearest theta band or ley line. Complete your transmission by midnight tonight for a special free gift!
User avatar

Zot Zealot

Posts: 982

Joined: Monday, 29th September 2014, 09:04

Post Tuesday, 2nd April 2019, 23:37

Re: Experimental branch: Known Traps

Is the reply online?
Previous

Return to Game Design Discussion

Who is online

Users browsing this forum: Google [Bot] and 26 guests

cron
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group.
Designed by ST Software for PTF.