People have a problem with traps: They’re boring.
Not only are they boring when they’re triggered — with the DM arbitrarily telling you to make a saving throw at the penalty of suffering some minor amount of damage — they engender boring play by encouraging players to turtle up and methodically, laboriously, and excruciatingly examine every square inch of the dungeon in torrid bouts of pace-murdering paranoia.
And if you feel this way, you’re in illustrious company. Here’s Gary Gygax giving some of the worst GMing advice you’ll hopefully ever read (Dungeon Master’s Guide, 1979):
Assume your players are continually wasting time (thus making the so-called adventure drag out into a boring session of dice rolling and delay) if they are checking endlessly for traps and listening at every door. If this persists, despite the obvious displeasure you express, the requirement that helmets be doffed and mail coifs removed to listen at a door, and then be carefully replaced, the warnings about ear seekers, and frequent checks for wandering monsters (q.v.), then you will have to take more direct part in things. Mocking their over-cautious behavior as near cowardice, rolling huge handfuls of dice and then telling them the results are negative, and statements to the effect that: “You detect nothing, and nothing has detected YOU so far—” might suffice. If the problem should continue, then rooms full of silent monsters will turn the tide, but that is the stuff of later adventures.
Uh… yeah. Do literally none of that. But you can feel Gygax’s palpable frustration with the style of play his own killer dungeons had created boiling off the page.
Despite this, traps are a staple of Dungeons & Dragons. They date back to the earliest days of the hobby and they remain a prominent part of the game’s culture and its adventures. In fact, if you go back to the ‘70s and ‘80s you’ll find that traps weren’t just tolerated, they were gleefully celebrated.
Is that because people were clueless back then? They were just fooling themselves into thinking they liked traps?
No, in fact. It turns out that traps used to be different.
We’ll start by looking at how they were different, and then we’ll talk about why that’s important.
QUICK HISTORICAL SURVEY
If you look all the way back to the original edition of D&D in 1974, there are three things to note:
- Thieves didn’t exist yet, and there were no skills (or other checks) that could be used to find or disable traps.
- Traps did not automatically trigger. Instead, they triggered on a roll of 1 or 2 on a d6. (In other words, any time someone walked down a hallway with a trap in it, there was only a 1 in 3 chance the trap would actually go off.)
- Carefully searching an area for a trap took 1 turn. This was a substantial systemic cost, because the DM made a wandering monster check (with a 1 in 6 chance) every single turn.
In Supplement 1: Greyhawk (1975), the thief class was added. There was now a skill check that could be made to find and disable traps.
AD&D (1977-79) dropped the 1 in 3 chance of a trap triggering. This mechanic was still commonly found in published modules of the era, however, and, therefore, remained part of the meme-sphere for a time. However, as play moved away from open table megadungeons and DMs increasingly ran disposable dungeons designed for a single traverse, the 1 in 3 chance meant that some traps would never be encountered. The idea of PCs not seeing every single scrap of material in a scenario became a sort of heresy, and this mechanic phased out.
The use of wandering monster checks also became deprecated. First by significantly reducing the frequency of checks and, later, often eliminating the wandering monster check entirely. This eliminated the system costs associated with searching anywhere and everywhere.
Over the course of 2nd Edition, modules slowly standardized trap stat blocks. 3rd Edition then incorporated these into the DMG (actually presenting the most extensive resource of pre-built traps seen in a core rulebook up to that point). Whereas previously the presentation of traps had been organic and narrative, it was now largely formalized into a check-or-damage mechanical format.
3rd Edition also substantially reduced the amount of time required to search an area for traps from 1 turn (10 minutes) to, generally, 1 round.
Jumping to 5th Edition, we discover both the worst advice and some good advice for running traps jammed together on the same page.
The worst advice is the mechanical structure: Passive Wisdom (Perception) checks determine whether anyone notices the trap. If they do, an Intelligence (Investigation) allows the character to figure out how to disable it. And then a Dexterity (Thieves’ Tools) check determines whether they can actually disable it.
In other words, by 5th Edition the mechanical resolution of a trap has devolved into an entirely automatic sequence of mechanical interactions which the players neither initiate nor make meaningful choices during.
No wonder people think traps are boring! You could do this with ANY element of the game and it would be boring! Imagine if every social interaction was resolved with a passive Charisma check to initiate the conversation, a Wisdom (Insight) check to determine what you should say to them, and a Charisma (Persuasion) check to see if you say it successfully.
Both the fiction and the mechanics have atrophied, and the fiction-mechanics cycle has broken down.
The good advice is this bit:
Foiling traps can be a little more complicated. Consider a trapped treasure chest. If the chest is opened without first pulling on the two handles set in it sides, a mechanism inside fires a hail of poison needles toward anyone in front of it. After inspecting the chest and making a few checks, the characters are still unsure if it’s trapped. Rather than simply open the chest, they prop a shield in front of it and push the chest open at a distance with an iron rod. In this case, the trap still triggers, but the hail of needles fires harmlessly into the shield.
Why is this good advice? And what does it mean to actually put this advice into practice?
HOW TRAPS WORK
Let’s briefly sum up how traps used to work:
- There was a cost associated with initiating a search, so players had to make deliberate and specific choices about when and where to look for traps.
- The 1 in 3 mechanic made the outcome of even identical traps less predictable: It wasn’t always the guy in front who triggered the trap. Sometimes it would be the last person in line. Or maybe the trap would go off in the middle of the group. Or you might walk past it safely on your way into the dungeon only to trigger it as you were desperately trying to run back out again. Completely different dynamics (and experiences) in each case.
- There were no mechanics, so players had to creatively interact with a trap in order to both find and deal with it. And, on the flip-side, this also forced DMs to creatively define the nature of the trap beyond skill check DCs.
Let’s start with the cost. If you want to avoid every expedition being slowed to a snail’s crawl by paranoia (or players simply feeling resentful that they have to choose between having fun and avoiding an intermittent damage tax), then there needs to be a cost associated with searching so that the players have to strategically decide when it’s worthwhile to pay that cost. In other words, the cost forces the players to make meaningful (and interesting) choices.
This cost will usually take the form of time: Time wasted searching for traps makes you vulnerable to other threats. Wandering monster checks are one way of modeling an environment filled with active threats that can find the PCs. Adversary rosters are another. Any form of time limit can be effective, however, as long as the searching chews up meaningful chunks of that time.
Alternatively, recognize that there is no cost in the current situation and, therefore, no reason for the PCs to not laboriously search every inch and be as safe as possible. This usually means that no meaningful choices are being made during these searches, which is what The Art of Pacing describes as empty time. You want to skip past that empty time and get to the next meaningful choice. I recommend using Let It Ride techniques here.
Note: This may not always be the right call. If the players are having fun making those meticulous decisions, then they ARE meaningful choices and it’s OK to live in that moment. Similarly, these choices can also be used to effect. I’ve run horror scenarios, for example, where the fact that the PCs have been reduced to terrified paranoia is 100% the desired emotional space, and cutting past those moments of paranoia wouldn’t be the right call. The thing you’re trying to avoid here is boredom.
Next let’s talk about the trigger uncertainty. I don’t think it’s universally true that traps should have unreliable triggers, but it’s a concept that’s worth playing around with if you haven’t tried it. There’s a lot of fun stuff to be discovered in play here. To a large extent, you can just graft the old 1 in 3 mechanic back in. (Or use slightly different odds, like a coin-flip.) Alternatively, you might have a trap trigger 100% of the time, but randomly determine which party member or rank in the marching order it afflicts.
Finally, there’s creative engagement with the players. This is vital. If all you can do with a trap is make a skill check to Search for it, make a skill check to Disable it, and/or make a saving throw to avoid taking damage from it, then the trap will be boring. The players have to be able to creatively engage with traps the same way they can creatively engage other aspects of the game world.
However, achieving this does NOT require you to simply throw out the mechanics.
PLAYER EXPERTISE
In The Art of Rulings, I actually use a trapped chest to demonstrate the fundamental principles of making a ruling in an RPG because properly adjudicating a trap is an almost perfect example of how a GM can use the mechanics of an RPG effectively. To briefly review:
- Passive observation is automatically triggered.
- Player expertise activates character expertise.
- Player expertise can trump character expertise.
If we look at 5th Edition’s mechanical method for traps, it exists entirely in the first two categories: Traps are detected through passive Wisdom (Perception) checks that do not require a declaration from the players (i.e., passive observation is automatically triggered). Analyzing the trap and then disabling it presumably require player declarations, but the rote formulation is the most basic example of player expertise activating character expertise. It requires no meaningful decision-making on their part: You detect a trap, you say you’re analyzing it, and then you say you’re disabling it.
To make traps more interesting, what we want to do is push that entire interaction up the hierarchy: Instead of starting with passive observation and ending with shallow declarations of player expertise, we want to start with the players making meaningful choices and end by opening the door to players creatively figuring out how to trump the basic skill check.
Start by requiring player expertise to search for traps. You can use 5th Edition’s rules for passive checks if you want (I’m not a fan), but it should still require the players to say, “I’m going to check for traps.” As we’ve discussed, of course, there has to be a cost to this declaration for it to be meaningful. Otherwise it’s just a rote catechism of dungeoncrawling (make sure you say it or the DM will getcha!). What you want is for the characters to be making broad strategic choices about when and where and why they’re choosing to search (and, conversely, when and why and where they choose NOT to search).
In order for this to be effective, the placement of traps has to make sense. As the 3rd Edition Dungeon Master’s Guide says:
The solution is to place traps only when appropriate. Characters and creatures put traps on tombs and vaults to keep out intruders, but traps can be annoying and inappropriate in well-traveled areas. An intelligent creature is never going to build a trap that it might fall victim to itself.
If the placement of traps is random or capricious, there’s nothing for the players to base their decisions on. The result will be either frustration or resignation.
As a rule of thumb, you’ll know you’ve gotten the balance right if the players start actively trying to collect intel on traps. (They might question prisoners, check blueprints, cast auguries, etc.) If they’re doing that, then they both value those strategic decisions AND have faith in the logic and consistency of the game world.
Design Note: You can also explore – possibly at the prompting of your players – resolution options somewhere between not searching and detailed searching. An old school example is tapping a ten foot pole in front of you while walking down a dungeon corridor. This standard operating procedure probably doesn’t reduce the party’s speed, but still has a chance of prematurely triggering a trap before someone walks into it. This synced well with the old 1 in 3 chance of triggering a trap: The GM could simply add such a check for the pole-tapping (or perhaps a 1 in 6 chance to reflect that the pole was less likely than a full-grown person to effect certain triggers).
The method thus had a significantly reduced cost (in gold and encumbrance costs for the pole itself, plus a penalty to stealth tests from the tap-tap-tapping), but a similarly reduced efficiency in terms of actually detecting the trap.
If the players are expressing a desire for some sort of “extra caution, but not so extra that we have to pay the normal cost for a detailed search,” ask them what that looks like. Maybe they’ll come up with pole-tapping. Maybe they’ll come up with something completely different! Then see if there’s a way you can model that with a minor cost and/or minor benefit.
Another option is Matryoshka search techniques coupled to passive observation. Rather than saying “you found a trap,” you can instead use 3rd Edition-style Spot checks or 5th Edition-style passive Perception checks to incorporate details into your description of the dungeon which, if investigated in more detail, would reveal the trap. (For example, you might mention the line of decorative holes running down the length of the hall… which turn out to be the firing tubes for an arrow trap.)
PLAYER CREATIVITY
When it comes to the trap itself, the description of the trap should not be limited to a mechanical effect. Understand how the trap works and communicate that to the players (either in response to their search efforts or when the trap is triggered). It is these details which allow the players to engage the trap creatively – to “get their Indiana Jones on.” This is what begins to move a trap away from being a rote mechanical interaction and turns it into an interesting and interactive experience.
There’s no hard-and-fast rule for this, but if the PCs start doing stuff like scavenging the tension ropes that reset a spike trap in order to tie up a kobold prisoner or draining the alchemist’s fire through the nozzles of a flame trap to pour down the arrow holes of another, then you’ve nailed it.
You’ll also start seeing the PCs thinking about ways to bypass the trap, often in ways that also bypass the mechanical resolution of disabling the trap. (This is where player expertise trumps character expertise!) For example, they might use chalk to outline a pit trap so that everyone can walk safely around it. Or put a board in front of the arrow holes in the wall. Alternatively, some of these solutions might simply shift the mechanical resolution: Placing a board across a pit, for example, might require Dexterity (Acrobatics) checks for everyone to walk across instead of Dexterity (Thieves’ Tools) to disable.
And if the PCs do disable the trap, I recommend asking them how they actually do it. (Or, at the very least, describe it specifically when narrating resolution.) When they disable the pit trap do they wedge it open? Do they nail a board over the top of it? Do they wedge it with spikes so that it can support their weight one at a time? The difference will matter if they end up getting chased back down that hall by ogres!
Getting this type of specificity can sometimes be challenging with magical traps. Check out Random GM Tips: Disarming Magical Traps for some thoughts on how you can make these more interesting than just saying, “It’s magic!”
I always did think it was a mistake that 3e put the traps in the DMG and not with the monsters (especially in the case of certain hazards like molds). I think it A: would have prompted more use of traps, and B: would have encouraged folks to think of traps as something to be interacted with over the course of several rounds, instead of an instantaneous event.
The ceiling in Temple of Doom takes its time, giving Indy a chance to try several things to stop it. The Breath of God has a buildup before the blades come out. The boulder does not drop on Indy, it chases him..
Nice!
In the third edition I use three modes of progress: normal (base speed for 10 per minute), exploration (half speed) and obsessive (1.5m per minute). in exploration mode the PCs roll perception if there is a trap but I just tell them that they see something strange (holes in the wall, the floor sounds hollow etc). In obsessive mode they take 20 per perception roll. random encounters 1 per minute (the speed of advancement of the third edition is a tenth of that of the old editions).
in OD&D the PCs automatically see holes in the wall etc.
Are you familiar with some good sources of traps with description how the can function in the fantasy world? What is the internal mechanism, how they can be disabled or farmed for resources?
One thing about trigger uncertainty is that it abstracts the dungeon, making it easier to run theater-of-the-mind or at least not dragging things out with excessive detail; you don’t need to know exactly where everyone is stepping to handle a trap. For example, Caverns of Thracia area 6 (p. 23) has a spear trap that shoots out on a 1-4 instead of 1-2. To my mind, I envision a big floor tile, likely taking up the middle third of the hallway (so like three-feet wide), that’s hard to miss. The pit trap before room 11 (p. 26) triggers on a 1-3; it’s likely not as big, or maybe off to one side. The spear trap past the bars in room 17 (p. 28) is likely a floor tile as big as a grown-up’s foot, maybe a little smaller given that the trigger is likely focusing on the phantom dryad, or a tripwire.
I know this is prep-intensive, and not every DM will want to do this all the time (including me), but I like to frequently show players photoshopped pictures of the dungeon and its contents. Players might be lulled into just enjoying the scenery, but at best some of them say to the DM: “Hey, can I see that again? That section of the rough wall looks like the outline of a door…” and player skill determines if they find the secret door.
Similarly, smart players will notice the maroon dust and piled up dead insects and rats in front of a door which is leaking poisonous gases, and incautious ones will pry the door open immediately.
I liked your point about player characters using the trap components creatively. They are more likely to do so if the specific elements of the trap have been described in words or shown in the DM-prepared photoshopped picture.
Fantastic post Justin! I have never placed a trap at a door. I don’t recall even ever seeing a published adventure with a door trap, yet just about every group I ever DM’d want to constantly stop and check doors for traps. Regarding cost, in OD&D just about every d6 roll has positive results on 1 or 2, except for the wandering monster roll which is positive on 6. I tell my players straight up they can search or listen etc. all they want, but if they roll a 6 there will be a monster.
You might also be interested in the advice on traps from Snider and Arneson’s Adventures in Fantasy game (1978) “… since the beasts traveling in and out of the lair all the time would probably avoid the bother of setting are resetting such things,…” They suggest traps be placed “at the end of a passage that dead ends (and thus one not used to often by the inhabitants)”. p22
My practice is quite similar. My general rule is that a trap that is found but difficult to bypass is more interesting than a trap that is not found and springs; and that a trap that is suspected but not necessarily found (or found and not necessarily bypassed) is more interesting than a trap that is not expected and springs, or is suspected and found.
So for hallway traps, which are not expected, I make them hard to detect visually (ie by passive perception when walking or running – I set the DC higher than it is reasonably likely that any PC of an appropriate level will have), but easy to detect using standard procedures like a 10′ pole. I do make the 10′ pole procedure slow down the party significantly (although no more than mapping, if different party members are responsible for each), so that there may be a cost; but I am more interested in how they get by the discovered trap than in actually springing the trap.
Where traps are expected, like doors and chests, there is more flexibility. If the trap isn’t found, or if it is found and discovered to be difficult to disarm, or if it is not clear whether a disarm attempt is unsuccessful, there is uncertainty that requires the party to consider what precautions to take. Chests containing breakable treasure like potions kick this up a notch, since many precautions can spoil the loot.
Player skill trumps skill checks, or at the very least improves your chances. I always have a general idea of how the trap functions, so that I can adjudicate attempts to use player skill. Courtney Campbell’s PDF on “Tricks, Empty Rooms and Basic Trap Design” is great for this (his website was better, but he deleted most of it).
Relying on a skill check is a risk. Assessing how a trap works and correctly guessing how to bypass it is a much lower risk (although it can take more time and lead to further wandering monster checks). Even figuring out only part of the mechanism can be a lower risk, if I think it is worth a significant bonus to the roll.
I have always been a fan of the occasional “partly triggered” trap. E.g. On a 1 or 2 its triggered, on a 3 its partially triggered.
Thog the Heavy gets unlucky and rolls the 3 moving over a deep pit trap.
“Thog, you feel the ground beneath you give way for a moment and there is a splintering noise. You see you are in the middle of a flimsy cover over some pit, and it is precariously close to giving way. If you move suddenly it will probably drop you like a rock into the pit below. What do you do?”
To be honest I have come to the conclusion that traps and some other things are better when the players know about the information and it’s more of a puzzle than a surprise. And basically your thoughts on traps kind of says why.
When i basically gave strong hints about a trap in a dungeon before (and there was never you can just roll disable traps to disable it) it usually created interesting scenes where they tried to counteract the trap, exactly like you said with putting boards on holes or chalk around a pit trap.
Basically traps really should just be another obstacle that players have to deal with instead of a surprise you’ve gotten less hp now.
@Xercies: Very much so.
Something I talk about in the Art of Rulings but didn’t specifically call out here is that you can literally just raise the threshold required for players to activate character expertise.
For example, if they say, “I disable it.” You can simply decide that this is not enough specificity and say, “How do you do that?” Even if it’s still ultimately resolved with a Disable check, you’ve pushed the specificity into the game world.
You can also require that extra specificity to encourage/teach a particular style or emphasis of play, and then back off it once the players know that option exists (and are comfortable exercising it when they
(See Part 2 of the Art of Rulings. I actually use traps as a specific example there, too, when talking about this.)
@DanDare2050: I like the partial/full trigger. I’ll have to play around with that! Also something to be said for keeping something similar in mind with partial failures on Disable checks, too.
Traps seem to me to be something that could use some clean procedures. (Sneaking around, too.)
A couple of questions come to mind. First, how would you deal with tripwire traps? Most tripwires are trivially easy to avoid once you’ve spotted them. If you use passive Perception (which I gather is *not* how you’d handle it), you risk making the trap a non-event. On the other hand, it seems kind of unfair for the players to have *no* chance of spotting a tripwire unless they’re explicitly searching for traps.
My second question concerns trigger uncertainty. With a party of four, rolling a die four times every time they walk down a trapped hallway could become rather tedious, and also risks tipping them off that the hallway is trapped. (Rolling for *every* hallway eliminates the second problem, at the cost of greatly exacerbating the first. Having the players roll reduces the first problem but exacerbates the second.) How would you prevent these problems?
@Percio: Out of curiosity, is 1.5 m your own conversion, or is there a regional translation of the 3e rules that actually uses metric?
@Wyvern:
1. To use 3E terms for clarity, make a Spot check. Actually roll the dice.
2. When using this procedure, I roll # of dice = # of party members (or # or ranks in marching order) and then read them left-to-right in marching order.
There are other ways of doing trigger uncertainty: Checking for the whole group instead of individual members (and then randomly determining rank if the trap does go off). Alternatively, skip the group check (so traps trigger 100% of the time) and just randomly determine rank (uncertainty is which party member triggers it, not whether it’s triggered).
If you’re doing a full group check, of course, you’d probably want higher odds than if you’re checking everyone individually. Instead of 1 in 4 or 1 in 3, might make more sense to do 3 in 4 or 4 in 6.
3. Star Wars D20 used 1.5m for 5-foot squares/steps, IIRC.
I wondered about rolling all the dice at once, but I wasn’t sure how you’d determine who tripped the trap that way. Reading left to right is a clever solution.
When would you use a full group check? Is that what you’d do if a trap affected the whole room at once, so it didn’t matter who set it off?
I just checked, and the Star Wars rules use 2-meter steps. Percio specifically mentioned “third edition”, however, and the examples given sounded like they came from a D&D context.
An alternate to the normal passive perception rule is as follows:
1 if the group is specifically checking for traps do the normal stuff.
2 otherwise secretly roll a group “perception base” every so many turns. During that time their passive perception works with that dice roll instead of the fixed number used in the normal rules. If they are hurried reduce the result by some reasonable amount. Then passive perception works as normal but the players have no idea how high or low their perception is currently.
An alternative is to keep a secret passive perception result for every player so sometimes the less perceptive spot things the more perceptive missed.
Another option is to add a bonus when the group is fresh and subtract a bit from the passive perception when the group is getting fatigued.
Wyvern: I’ve been doing that very thing for years.
Big gang of monsters attacking one PC? Roll five d20 and go left to right.
Fireball hits all the monsters? Roll all the saving throws at once and just go left to right.
Initiative for several different creatures? Roll all the dice and go top to bottom.
You get pretty quick after a while, especially if all the modifiers are the same. That’s kinda the simplest use case, and pretty good for situations like this where you want to know which position(s) in a line of characters trigger the trap.
So, i’m curious; if you are running something with a skill that addresses the “fix” stage of the trap (such as disable device), do you just… not have that skill? Or is the disable device DC effectively the level of expertise that it takes for the rogue to actually achieve whatever solution the party has come up with (aka, once they find a solution, only THEN are they allowed to make the disable device check?)
@Arturo:
“You’ll also start seeing the PCs thinking about ways to bypass the trap, often in ways that also bypass the mechanical resolution of disabling the trap. (This is where player expertise trumps character expertise!)”
Check out The Art of Rulings for an in-depth look at how character expertise and player expertise interact, but the short answer is: No. The skill check (i.e., the character’s expertise) exists. But if you allow player expertise to trump the character’s expertise (either replacing or providing a bonus to that check), you’ll encourage the players to creatively interact with the game world rather than just defaulting to a vanilla skill check.
Alternatively, you can also raise the threshold required for activating character expertise (i.e., requiring players to describe exactly how they’re disabling the trap, instead of just saying “I make a Disable Device” check).
I was thinking something similar: where if a Perception check is made, than an in-detail description of the trap is made as to how it works, what it looks like, maybe hints as to what it needs in order for it to be working.
Then, once the characters come up with a solution, that solution’s success depends on the Disable Device check (modified by just how intelligent or ridiculous the solution attempted actually is) not unlike how a player can *say* something meaningful but still has to roll diplomacy, except in this case, it actually feels *less* disjointed because a GM can say “that solution was so simple it doesn’t require a check”
To give an example based on Pathfinder’s Electricity Arc Trap
ELECTRICITY ARC TRAP
Hint: The air in this area smells like ozone
Type: Mechanical;
Perception DC 25 – You’ve noticed a one by one foot square of tile; the grout here is uneven and missing compared to the rest of the room, and closer inspection reveals that it’s clearly designed to bob up and down to a certain amount of weight. Testing with your tools quickly reveals that there’s some kind of magnetic field beneath it; it pulls your tools towards it ever so slightly.
DC 28 – You follow notice the slight magnetic pull seems to follow along the floor, then climbs up the left wall and continues up onto the roof some 20 feet back to what appears to be copper piece jutting from the stone.
Trigger: Touch (1-3); 1 by 1 foot floor tile, 5lbs
Reset: none
EFFECT electricity Arc (4d6 electricity damage, DC 20 Reflex Save for half damage); multiple targets (all targets in a 30 ft. line)
Disable Skill: Disable Device 20
Bypass: The lightning-rod from Lobby 3A was used to allow guests to pass through this hallway unhindered hundreds of years ago. if inserted at the beginning of the pathway, the arc harmlessly strikes it instead.
Grimtooths Traps (multiple volumes) is the source of utterly fiendish traps. Very very deadly but huge fun.
@FILIP GRUSZCZYNSKI
Traps are very context dependent.
Just read about real-world hunting traps (from trees-and-ropes to tripwire shotguns, as well as non-lethal ones) in old hunting books, with illustrations, and whatever you can find on booby traps.
Then play Dwarf Fortress for a while (then read about “dwarfputing” on DF wiki, and play some more). You’ll start to “think with portals” where traps are concerned.