Dungeons & Dragons includes lots of sneaky creatures proficient in Stealth, from Abominable Yeti to Yeti. That alphabetical range somehow fails to make my point, but more than 500 creatures fit in between. The game includes creeping monsters who work as ambush predators like spiders (Stealth +7), crafty ambushers like goblins (Stealth +6), as well as potential jump scares from wights, and jumping oh-shit surprises from bulettes.

When these creatures get some chances to lie in wait or to sneak up for surprise, they show the traits that make them more interesting than a claw/claw/bite sequence. Combat scenes feel more varied. Plus the barbarian gets to show off Feral Instinct. However, I have almost never run sneaky monsters for surprise, and based on countless convention games virtually every other DM plays sneaky wrong too.

Part of the blame rests on the fifth edition rule for hiding before an attack. “Compare the Dexterity (Stealth) checks of anyone hiding with the passive Wisdom (Perception) score of each creature on the opposing side. Any character or monster that doesn’t notice a threat is surprised at the start of the encounter.” In the highly technical jargon of game design, this rule is rubbish.

The worst check among the hiders determines surprise, so with multiple foes and checks, the chance of success veers quickly toward impossible. Based on real life, you might suppose that ambushes often work. Based on the D&D rules, the word “ambush” describes a imaginary event that can never happen, at least for groups of sneaky monsters. (For a look at the reverse of this dynamic, see In D&D, Letting Everyone Roll Certain Checks Guarantees Success, So Why Bother Rolling?.)

To fix the rule so ambushers can surprise, skip the roll to hide. As a dungeon master, you can set the difficulty class to spot an ambush just as you set DCs for other checks. Then, compare each character’s passive Wisdom (Perception) score against the DC to determine if the character is surprised. This matches the procedure you might use for a trap.

As a technique for setting the DC to detect an ambush, just add 10 to the ambushers’ Stealth bonus, so hiding goblins with Stealth +6 require a DC 16 to spot. If the sneaky creature has advantage, add 5 to the DC. For disadvantage, subtract 5 from the DC. A group setting an ambush has the edge of planning, preparation, and familiarity with the location, so a DM focused on realism might allow advantage and enable many surprises, but not the kind players favor. More entertaining ambushes probably lead to a mix of surprised characters and characters who act immediately. Such situations reward characters with high perception.

This approach boosts sneaky monsters and perceptive characters, but it eliminates die rolls and the fun uncertainty that dice add. By choosing a DC, the DM essentially decides who becomes surprised, and it’s always the same, oblivious characters. The D&D designers favor passive perception because it avoids slowing the game for rolling Wisdom (Perception) checks. In dungeons that encourage players to look frequently for traps, passive checks can speed past a lot of rolls. But lurking monsters deserve a different procedure: Just have players roll a Wisdom (Perception) check before the likely roll for initiative. The characters who fail their check begin surprised.

In an ambush, initiative can start at one of two times:

• When the monsters attack.
• When one or more characters notice the potential ambush.

This second case leads to judgement calls from the DM, but can also lead to interesting scenes. How the realization plays depends on how the perceptive characters react to spotting the attackers.

If the ranger draws a sword and yells, “To Arms!” then initiative starts and the less perceptive players begin surprised. Meanwhile, the monsters are already aware of their foes, so they start without surprise.

If the ranger plays it cool and tries to alert the party, then the scene turns on whether the ambushers happen to attack immediately. Does the ambushers’ insight reveal that the archer spotted them? If so, they strike and may surprise some characters. Can the ranger deceive the ambushers into waiting just a moment longer while covertly signaling the party? Ask what the alert characters do, and then call for checks to play the scene.

Distance factors into the situation too. In daylight, if the eagle-eyed ranger spots a group of yuan-ti hiding among desert rocks a half-mile ahead, then any surprise hardly matters. During that first round, no one without meteor swarm is in range to attack. The players can select the best angle of approach or can just avoid the monsters. Since yuan-ti have darkvision to 120 feet, they surely wait until night when most adventurers with darkvision see 60 feet and still suffer disadvantage to Wisdom (Perception) checks unless they carry a light source. Such scenes allow characters to show their special traits, like the drow character’s enhanced night vision that lets the party turn the tables.

For that ambush in darkness, you must judge how close the party will approach the yuan-ti before any characters can detect the snakes. Perhaps at 60 feet, allow the ranger leading the party one check at disadvantage due to dim light. At 30 feet, in the moment the monsters spring their attack, allow a second check for the entire party.

Ultimately, many ambush attempts lead into situations that demand DMs ready to balance distance, skills, and traits to create fun and exciting situations. An ’80s roleplaying game might have included a table that factored Stealth scores, Perception scores, terrain and lighting into a roll that revealed encounter distance. The authors of the Player’s Handbook skipped all that complexity in favor of simplicity. The approach relies on a DM’s judgement, but isn’t that part of what makes being a dungeon master fun?