So, you don't want players to actually face save or die effects. You want them to fear save or die effects and when they don't react appropriately, then kill them.
Because otherwise, why would you change how poison works after they know poison is in play? Honestly, that is just so much worse to me. Player 1 gets hit, and takes damage, but is fine. Player 2 gets hit next round and dies. The Dm shrugs and says "sorry, player 1 got hit with a lower dose than you did"
No, my point is that there are ways for the PCs/players to know that poison is in play, and that they need to be prepared because poison is often save or die.
Notice that in my snake example (becuase everyone knows snakes are venomous) I pointed out that knowing that might not matter. If you are thrown into a pit of snakes, you don't have a lot of other options, whether you know snakes are poisonous or not.
Quite true. Although in any edition, being thrown into a pit of venomous snakes has a good chance of being fatal. On the other hand, it made for a dramatic scene in
Raiders of the Lost Ark where the "DM" gave the "PC" an opportunity to change their tactics.
But again, this seems to miss the point. Sure, the person who fell and missed all the spikes didn't die. So now they know traps are in the area. But they don't know that a single failed save is going to kill them. And, they likely went into this expecting traps anyways. The Tomb of Horrors, if memory serves, tells the players that the tomb is a death trap full of deadly traps. It doesn't need a trap in front to remind them that there are traps in the dungeon.
If they know that poison is frequently save or die, then they do know that.
When you feel the need to make it less likely the players will fail, you need to look at why that is.
...
If you find yourself granting advantage on saves, because a failure will kill them, and you don't want them to die. Then maybe instead of giving advantage, you should make the trap not deadly.
It's not really a situation of "making it less likely to fail." It's about building suspense and tension.
Consider this, comparing the mechanics):
DM: Make a Perception check.
Player: Crap, a 7.
The floor suddenly drops below you and you fall into a pit, landing on several spikes for 11 piercing damage. Make a Constitution save.
Player: 17
DM: You feel a slight burning, probably poison, but it doesn't cause any other effects.
or
DM: The floor suddenly drops, a pit trap? Roll percentile dice.
Player: Crap, I hope I make it. 37.
DM: You tried to leap free, but the floor dropped out too quickly. Roll a d6 to see if you land on any spikes.
Player: Spikes? Ugh. OK, A 4.
DM: You manage to rotate enough to land between two of the spikes, they glisten slightly with a thick black substance.
Player: Probably poison. Whew, that was close.
The first one is quick and doesn't build any suspense. Also, failure to detect the pit always results in damage. This is the 5e mechanic in the updated ToH.
The second one is the original mechanic. It builds more suspense, especially since players know that poison is save or die. Now they know there are pit traps with poison.
Yes, they expected traps, and are probably taking precautions. But this still increases the tension (in my opinion) because it highlights that the traps are likely to be deadly. This is also the opening of the adventure, letting you know that the adventure is likely to be deadly.
In 5e, how deadly is it? What does it tell you about the rest of the adventure? There are traps, and poison. Yeah, as long as we don't take too much damage in a single hit, we can expend Hit Dice, or stop and rest if we need to.
--
Another example, from ToH. It's often said that it has too many arbitrary traps that are too deadly. But once again, examining the math shows that the chance of death is actually very slim, and that it's a device to ramp up the suspense and tension. For example, the Knockout Corridor.
There's a sleep gas that has a good chance of putting most of the party to sleep. Although it doesn't specify whether elves are immune, they should be (in 5e, they are knocked unconscious, eliminating this ambiguity). So the rest of the party (which could be somewhat large with hirelings and henchmen), are knocked out. A stone juggermaught might come out, and start coming down the corridor, and crush anyone in its path. Mathematically, it's very rare that somebody will actually get crushed. Once could argue that it could be made so nobody gets crushed.
This plays out the same way in 1e and 5e, they didn't alter it other than the wording to unconscious. The events are often not experienced by the PCs, it's the players that are the audience while their PCs lie helpless. Since the rest of the dungeon has ramped up the difficulty, by this point their expectation is that it could be deadly. The fact that occasionally it is (and players talk to other players), added to the reputation of the dungeon. That is, the dungeon among players shared the same reputation as it did within the game.
I'm not saying this is a perfect design, and I know that not everybody will like this approach to playing. But the mechanics in both of these cases in the original are building the suspense in a way that is missing in many of the mechanics of 5e.
Again, to each their own. But I've built off this approach for decades, and my players over the years have loved it. And I think the reason why is that it's often very difficult to make the players fear for the lives of their PCs. And it's become harder to do that over the years as the power level of the PCs have increased.
Again, I'm not saying one approach is better than another. They are simply different approaches to adventure design and mechanics, and provide a different feel and play experience. Different groups will like different approaches.
