The Monsters Know What They Re Doing

What the title actually means for your table

When I first started running games, I treated monsters like stat blocks waiting to be attacked. Kill the thing, get the XP, move on. That approach works fine until you run a pack of wolves and the players realize they have no idea what to do because the wolves aren't even trying to herd them into a corner. The concept behind The Monsters Know What They Re Doing is basically that monsters in your game should behave like intelligent predators or strategic combatants rather than stationary damage sponges. This doesn't mean every goblin is a military tactician. It means the dungeon's inhabitants make decisions based on self-preservation, territorial instinct, and whatever resources they actually have access to. The rules literally say that a creature would flee if it took heavy losses and had a reasonable chance to escape. Most DMs ignore this because it's easier to just run the stat block straight. But when you start giving monsters the instinct to break off combat under the right conditions, something weird happens. Players stop treating every encounter like a spreadsheet exercise. They start negotiating, setting traps, running away themselves, or using the environment in ways they never considered before. The game gets messier and more interesting almost immediately. The tricky part is figuring out how much intelligence each monster actually has. An owlbear isn't going to coordinate a flanking maneuver, but a bandit captain absolutely should. A gelatinous cube doesn't know anything, obviously. The challenge is applying the right level of decision-making without overthinking every single encounter. I usually draw the line at any creature with an Intelligence score above 4 or any humanoid monster. Everything else gets animalistic instincts and basic survival drives.

Here's what I learned the hard way. I ran a session where I had three gelatinous cubes and twelve skeletons in a corridor. The players were higher level than I expected and were chipping through the cubes fast. I had the skeletons fall back behind the remaining cube as cover, which is what The Monsters Know What They Re Doing would technically suggest. Instead of running away, the players just waited out the round and let the skeletons retreat on their own. I had prepared an entire tactical plan that went completely nowhere because the monsters weren't smart enough to execute it. After that, I stopped over-planning monster tactics for low-Intelligence creatures. I let them do what they naturally would and only design elaborate behaviors for things that could actually follow through on them. There are some situations where this approach breaks down completely. If your players are optimization machines who come prepared with area denial spells and crowd control, giving monsters smart behavior might just lead to longer, more exhausting combats where nothing gets killed efficiently. I've seen this happen with well-coordinated parties that can isolate and pick apart individual enemies before the monsters can react meaningfully. In those cases, the monsters fleeing just delays the inevitable and makes the session drag. You're better off leaning into numbers and environmental hazards instead of tactical retreats when facing that kind of group. Another problem shows up when you're running a low-level campaign with weak monsters. A bunch of level 1 goblins with The Monsters Know What They Re Doing isn't scary. They'll scout the party, decide they're outmatched, and scatter. That's logically sound but narratively disappointing for players who want a fight. I switched to giving them a incentive to stay rather than flee. Maybe the treasure is behind them and they'll defend it, or maybe they're ordered to hold position by a smarter leader. The behavior is the same. The motivation is different.

For practical implementation, you don't need to redesign your encounters. You just need to ask two questions before every battle. First, would this monster reasonably know the party is a threat? A guard patrolling a hallway knows. A random skeleton in an unexplored crypt does not. Second, does this monster have a viable escape route or reason to stay? Answer those and adjust aggression accordingly. It usually adds maybe thirty seconds of prep per encounter and changes how the fight plays out significantly. I also found that tracking monster morale separately from HP makes this easier to run. Instead of deciding on the fly whether a creature flees, I assign a small morale score at the top of each encounter note. Something like 8 for intelligent humanoids, 5 for beasts, and 0 for undead or constructs. When average HP drops below half and casualties accumulate, I check if morale triggers. It keeps the game moving without turning every fight into a management simulation. This method typically reducesDM prep time by fifteen to twenty minutes per session compared to designing full tactical sheets for every encounter. The biggest mistake I see people make is applying this philosophy too rigidly. Not every monster needs to outsmart the players. Sometimes a CR 1/4 rat swarm should just be a rat swarm. Overthinking this concept turns encounters into chess matches when your players just wanted to swing swords and roll dice. Use it as a guideline, not a rulebook. Let the monsters know what they're doing sometimes, not always.