The Dark Side of OSR Traps

One of the worst sessions I've ever run was a modified one-page dungeon a year or two back where I repeatedly shot down player ideas to escape a trap because they didn't fit with my internal notions of how a trap-mechanism worked. In trying to get the players to problem solve rather than just roll to disarm traps, I fell into a common pitfall of puzzles where there's no easy answer- the dreaded 'Read the GM's mind to proceed' scene.

THE SCENE
A stone-walled windmill on an island in a shallow lake. The interior was a central stone pillar, a sarcophagus full of treasure that could only be be interacted with via mirrors, and two exits- a heavily barred ceiling window, and the main doorway. Oh, and some corpses of past adventurers.

Interacting with the little riddle-clue hinting of the secret coffin and sniffing about with mirrors was fun. That wasn't the issue.

The issue was when the sarcophagus was looted and made visible, the trap activated- a waterfall of flaming oil pouring from the entrance and running outside. The players were trapped behind iron bars and fiery doom.

My idea was that the windmill pumped up a reservoir of oil as it span, and had enough oil to last for about 16 hours. It could be turned off by weighing the coffin down with weight PERFECTLY equivalent to the taken treasure and closing the lid ( That level of detail and insistence on exactitude should be a warning sign that you're more invested in your own story than the quality of the game, whether you're talking traps, your character, a plot, etc). A horde of Shadows would come from the forest, slip across the lake, up the side of the windmill, and through the barred grate, converting any thieves to shadows and  returning the stolen treasure to its resting place.

That little idea of how the trap operated seems innocuous, but here's where things started to go wrong- I never hinted at the weight mechanism, figuring it would be obvious to the players that if emptying the treasure and taking off the lid from the coffin triggered the trap, it could be un-triggered the same way. An easy way to communicate this would be to have the coffin rise up as the lid and treasures were removed until it was above ground level, elevated by a pressure plate.

Here's what the players tried, and why I shot them down, and why I shouldn't have. The reasoning might seem to make sense, but here's the common issue with the rulings I made- they were based on my own headcanon of how some bullshit fantasy trap mechanism works, not rulings on how to make the session engaging and interesting.
1. Replace the Coffin Lid 
 Almost, but not quite-they needed more weight. If there had been a clue, like the pressure plate beneath the coffin sinking slightly, they would have figured it out. But they tried a half-measure that they didn't know was a half-measure and so figured the coffin wasn't the key to it all. A player even climbed inside the coffin- that weight probably should have turned off the trap, and then there would have been the problem of a player being left behind in an invisible coffin. A good scene. Or maybe it could have sunk the coffin too far down the pressure plate, indicating to the players 'aha, we need exact weights.' Ideally they then could've used the old corpses to make up the difference, but one player had rolled them all into the flames already, so they'd have to sacrifice inventory items... good stuff.
But 'nothing happens' because it didn't  perfectly fit my notion of the trigger mechanism communicated to the players 'it's not the coffin, try something  else.' And so they did.
2. Apologizing to the Goddess the place was dedicated to 
 After all, there was clearly magic going on with the invisible coffin. But I decided the goddess didn't really care if these tomb-robbers lived or died, and plus, it was a secretive goddess, not like the pushy, chatty other deities of the setting. But heck, a sepulchral voice moaning 'return the treasure' or a feeling of guilt growing when looking at the treasure would've moved things forward.
3. Filing the ceiling bars
I decided this would take too long and the Shadows would show up before they could be sawed through, because they're very thick bars designed to prevent that sort of thing from happening. To which I now say, REALLY? A player went to the trouble of having files in their pack and got told, 'no, the tools you brought to do something like this aren't good enough.' So much for 'use your inventory to problem solve,' cripes.
4. Checking for secret doors
There were none, so none were found. Pretty straightforward, right? Well, maybe not- after having all their attempts at escaping through the obvious entrances shut down, it makes a lot of sense to go 'oh, there must be something we're missing.' But the only thing they were missing was actual information on how this place worked, and I provided no way to access that information.
5. Plug the nozzles shooting oil
 The nozzles, like the coffin, were only visible when viewed in a mirror and only your reflection could interact with it. I figured the torrent of flaming oil was too powerful to simply stuff with someone's cloak.
But I could have had clogging a nozzle create an opening in the flame. Oil could back up and start bubbling from the stone pillar that contained the pump, threatening to ignite everything and everyone if they didn't flee quickly. The thief who climbed up to reach the nozzles would be in perilous danger from spraying oil and there woulda been someone holding a mirror for them as they used their hands to climb. It woulda been a fun scene.

Real time, an hour or two had passed. The initial sparks of creativity were being replaced by frustration and boredom. Player morale was breaking down fast by now. The problem player starting harassing other people for not describing their actions as completely as they did, whining and nitpicking ooc. One player declared their character was just gonna go to sleep until something happened. Another player, trapped outside, had just been throwing rocks at the windmill's sails for hours.

6. Hoist the coffin lid over their head and use that to shield them from the flames.
I mentioned that damage would still be taken from the oil on the ground, and that dissuaded this plan for the moment. What I did NOT communicate was how much damage this would be- some players probably thought it would be 'instant cremation.' Had I said '2d8 damage and you have to jump in the lake afterwards' they probably would've done that immediately, but I assumed they knew flaming oil on the ground damage and knew that's what I thought the damage would be. Don't assume or imply mechanical stuff- just say what the dice will be. That's how the mechanics of the world work, so players should be able to know those mechanics to simulate their characters making reasonable decisions. Not everything should be described via simulationist roleplay.
7. Use the rotation of the windmill to rip the bars off
 Firstly, the rock-thrower had ruined the windmill's wind-catching ability. Secondly, there wasn't wind (though there probably should've been a good updraft from all this burning oil).
This led into
8. Try to reverse the rotation of the windmill to shut off the oil flow
The player on the outside could've been rewarded for being cautious and not getting caught, climbed the windmill, been passed the rope, used his bodyweight to spin the arms, and rescue everyone.
But I decided that the windmill only rotated one way and couldn't be reversed, and anyway it had an oil resevoir that was fueling the trap so the rotation was completely pointless save to pump more oil for a later activation of the trap.
Yup. Just shutting down a player idea of how the mechanism works, without giving them any idea of how it actually works. That's bad. Don't do stuff like that, mmmkay?
9. Call me out for making bad rulings and disengage
Thuvrig Mountaincloak-"Let's just wait it out guys, we've been just getting completely denied for an hour now."
The truth hurts don't it. And it's a good idea- there's no reason you can't tell your GM 'bruh this puzzle is boring and also the suck.' Anyway, even 'wait it out' got denied, because I ruled it was night now and the army of shadows showed up.

Attacked by hordes of ethereal monsters with no magic weapons, they went back to plan#6 and fled through the oil waterfall with the coffin-shield, taking 9 damage each and escaping, the end. This was a good 3 hours of play I believe, all for 'you take 9 damage and get the treasure.' Yikes.

THE MORAL OF THE STORY
A bad trap is just like, you walk down a corridor, you rolls dice -fail to detect traps, Dave gets hit by an arrow for 1d8 damage. There's no player agency there, it's just random math.

A bad way for players to interact with traps is to simply intone 'I roll to disarm traps' upon encountering one, rather than trying to deal with it in 'the OSR way', like holding a shield between you and an arrow trap, or poking the trigger with a 10-foot pole, having an expendable goat take point, or using the trap against your enemies, or whatever.

Traps like the flaming oil windmill are good- just LOOK at all the stuff the players came up with to escape. In the beginning, it really had them thinking and asking questions and being clever and engaging with the fiction rather than the maths. The trap was scenery and context, not just a penalty applied to their HP, soon forgotten, as so many official D&D traps are.
But GMing like I GMed it, is what makes people think OSR trapfinding is stupid pixelbitching baloney. If you're ever GMing something and the only thing happening is the players getting told 'nothing happens' that's a good sign you're making some sort of mistake, with keeping them un or mis-informed, having no stakes or pressures to drive a scene, or perhaps being too concerned with your vision of 'how things are' as opposed to 'how fun things are.'

Hopefully this lesson in what NOT to do helps you see the nebulous, ever-changing form of what TO do as a GM.