If a player is rolling a perception check, there is a possibility that something has already gone wrong.
In AD&D, the base chance for a first level Thief to find a trap is 20%. The base chance to remove said trap is also 20%. Which really used to bug me as a young player, because I couldn’t justify the incompetence of the low level Thief, particularly since it took so very long to have a halfway decent chance of doing anything.
But I now think we were doing it wrong, or at least doing it differently than had been intended. I believe primary method of finding traps, and dealing with the potential hazards of them, was intended to be through careful investigation and deduction from the clues provided by the DM. The find traps roll was only a safety net, a chance to find something if the players were stumped.
Pits, for example, were supposed to be found by running a 10′ pole along the floor, or prodding the floor ahead. Other traps were to be found by paying attention to the clues and making appropriate inquiries of the DM.
Looking for an example, I grabbed my copy of Tomb of Horrors and looked at the very first entry, the first “False Entrance Tunnel”. The entry reads:
The corridor is of plain stone, roughly worked, and it is dark and full of cobwebs. The roof 20′ overhead is obscured by these hanging strands, so casual observation will not reveal that it is composed of badly fitting stones. Daylight will be sufficient to reveal that there is a pair of oaken doors at the end of the passageway. If the roof is prodded with any force, or if the doors are opened, the roof of the tunnel will collapse and inflicts 5-50 (5d10) hit points of damage upon each character inside of it, with no saving throw. The doors open outwards by great iron ring pulls. The cobwebs must be burned away to be able to inspect the tunnel ceiling.
If I was to encounter this as a player now, I would be immediately suspicious of the cobwebs, and the fact that I can’t see the ceiling. Hopefully I would be smart enough to clear the cobwebs and look at the ceiling. The DM should then tell me, without a perception check, that the ceiling is composed of badly fitting stones. Which I would hopefully interpret as presenting a risk of cave-in, and start worrying about what might cause it.
Even if I didn’t do that, I might be careful with the door. Noting that it opened outwards and had rings, I might attach a rope to the rings and pull the door open from outside of the tunnel. Rocks fall, nobody dies. No perception check or thievery check required.
Probably the best series out there on adjudicating actions is the Angry GM’s series on GM Basics (which you can now find since he finally organized his site). The first article in the series, Adjudicate Actions Like a Motherf$&%ing Boss!, sets out the steps for properly adjudicating actions, item 3 of which are as follows:
3a. The DM determines whether or not the action is even possible
3b. The DM determines whether the outcome needs to be randomly determined
3b1. The DM determines how to randomly determine the outcome
3b2. The DM makes a die roll or instructs the player to make a die roll
3b3. The DM determines the outcome of the die roll
3c. The DM decides an outcome
3d. The DM describes the results of the action of the players
What we are doing here, is giving the players an opportunity to control the outcome by jumping from 3b to 3c; in other words, they can avoid random checks by careful investigation. If the players properly investigates the scene and properly interprets the clues, the outcome does not need to be randomly determined. If they don’t, or don’t want to, then the outcome does need to be randomly determined.
Note this does not prevent the player from relying entirely on skill checks. That player can still say, “I search for traps,” and risk death or injury to his character with the roll of a die. It just rewards the player who wants to figure it out for himself by giving him control of the situation.
One thing I would like to point out is the lack of extraneous detail. Pretty much all of the details in the Tomb of Horrors entry are related to the trap. While there can be red herrings, it is probably the best practice for the red herrings to be related to the trap, and for the red herrings to be discoverable as such.
For example, look at this pit trap from Hack & Slash. There is an obvious trip wire, which is a red herring. Jump over the trip wire to avoid it, and you end up in a pit. The pit is detectable by tapping. The red herring is therefore part of the trap, and is detectable.
This is different from a red herring which invites an investigation that has nothing to do with the trap. That just wastes time and encourages pixel-bitching.
It would have been very helpful to have an explanation of this style of play in the “Find/Remove Traps” entry in either the PHB or the DMG. But in fact, if it is explained anywhere in either book (other than some hints in the sample play description), I can’t find it.
Which is no doubt part of the reason for modern playstyles having developed the way they did.