Who's having failures not trigger meaningful consequences? The chandelier did fall, the character did believe the lie. Is anyone arguing for a lack of meaningful consequences for failure?
Chaosmancer seems very opposed to failure *always* having meaningful consequences, and IIRC, a few people have voiced agreement with that. I very much doubt anyone is opposed to failure *ever* having meaningful consenquences. Certainly no one here has given that impression.
The problem, I think, that folks are having here is how you can both telegraph a trap AND surprise a player with that trap.
By using telegraphs that are subtle enough that a player has to be paying attention to notice it. Think of it like Dark Souls. Part of the notorious difficulty of dark souls is ambushes and traps that get you when you're least expecting it. But these traps and ambushes are always telegraphed. If you are paying careful attention to your environment, you can pick up on the clues and avoid places you know are likely to be trapped. If you're rushing, you're likely to get nailed by them. This is why it is often said that Dark Souls is difficult, but fair. This is something I strive to emulate in my D&D games.
One effective way to do this is to have an observable pattern to the traps within a particular dungeon. Maybe in an old dwarven ruin, there are many statues of ancient dwarf heroes that flank various doorways and passages, some of which (but not all of which) mark traps; anyone who walks in-between the statues triggers spears that shoot up from the floor and skewer them. Early in the dungeon, you put a skeleton between a set of statues, with puncture marks in its armor. Now, it's very unlikely that the players will be surprised by
this specific trap, if they have an ounce of genre savvy. That's not the point of this trap. This trap is a "teach." Now the players have enough information to know that in this dungeon, they should be cautious of statues flanking doorways,
if they're paying attention. The next time they come to a doorway flanked by statues, they might remember the previous time, and they might investigate to look for holes the spears might come out from, or a pressure plate that might trigger the spears, or a secret passageway around the trap. Or, they might not have picked up on the pattern yet, and they might just get nailed by some spears. Either way, they're definitely going to be cautious the next time they see such statues. So, maybe next time, the statues have been largely destroyed. There's still a bit of the stone pedestal the statues sat on, but no statues. Again, players who have picked up on the pattern might realize that this is part of it if they realize these pedestals are the bases of destroyed statues, but then again they might be surprised by the trap if they miss that detail. And, you'll want some sets of statues that don't trigger such traps. They're just ordinary statues, flanking doors and passageways. If the players are paying attention, they're sure to be cautious, but when they can't find anything and nothing happens when they walk between them, this is going to get them thinking about what marks the difference between the trapped statues and the safe ones. And there should be a difference they can pick up on, if they're specifically looking for it. Maybe the trapped statues all depict dwarves from a particular clan of dwarves, and you need Proficiency in History (or Stonecunning) to recognize it. Then the broken statues are going to be real head scratchers.
Or, and this is my bigger issue, if you tell a possible consequence of failure to the player and then go with a totally different consequence, don't your players get annoyed?
I mean, if I'm swinging on that chandelier (why is this always an example? In 30+ years of gaming I've NEVER seen a player try this) and you tell me that I'll miss the jump if I fail my check and then have the chandelier break, don't your players react pretty negatively?
I imagine they would, which is why I wouldn't do that.
I've been repeatedly told that allowing my players to roll first will result in the players being angry for things like "Well, I didn't SAY I was doing that". I avoid that by being pretty clear up front that actions are resolved AFTER the roll, which means that the DM gets a smidgeon of control over the character from time to time. Making that roll indicates that you are okay with that.
Which is perfectly valid and fun way to do things. No judgment here, if that's what you and your players prefer, by all means, enjoy!
But, if you tell me X is going to happen if I fail and then Y happens, how is that not a bait and switch? "I wouldn't have done that if I thought that THAT could happen."
I'm not sure where you're getting the idea that anyone is doing this. You don't tell the players one thing will happen and then have another thing happen. You tell the players what it is reasonable for them to know will happen. Again with the illusory pit example, "if you fail, you'll fall in" doesn't give away that it's an illusion, doesn't mislead the players into thinking one thing will happen when really another will, and gives the players enough information to make a good decision about how to proceed. You don't have to tell them that they'll take 10d6 damage when they fall, just telling them they'll fall is fine. You don't have to tell them the magic sigil will explode if they mess up deactivating it, just telling them the volatile magical energies will be released unsafely is fine. You don't have to tell the players there's an unaware ogre on the other side of the door that will hear them and have enough time to prepare for them if they fail to break down the door in one go, just telling them that the noise will alert nearby creatures to their presence is fine. The idea that the only options are "tell the players information their characters couldn't know" or "don't tell the players the consequences of a failed roll" is a false dichotomy.