From Storyteller to the Oracle of Delphi: On Running RPGs Without Controlling the Story

 Watching Quinns’ review of Stonetop got me thinking about how my way of running tabletop RPGs has changed over time.

In more mainstream games, the Game Master is often imagined as a director or narrator. The organizer of a story where the characters are actors and the campaign moves toward major revelations planned in advance. Over time, my own style drifted away from that.

I’ve come to see the GM less as a writer and more as an oracle. A weaver barely capable of glimpsing the players’ fate a few steps ahead.

The GM is a strange figure behind a screen, consulting dice and random tables like a shaman reading omens in scattered bones.

The players walk through a forest toward the next town, what happens on the road?

The GM rolls on an encounter table and consults fate. Maybe nothing happens. Maybe something appears. Maybe a storm tears through the camp during the night. But the GM doesn’t know what will happen either until the dice hit the table.The roll says there are wolves. So perhaps another table is consulted, or another roll is made to ask fate a second question:

Are they hungry?
Are they following someone?
Are they fleeing from something worse?

Random tables are a fundamental part of this way of running games. They are the foundation of divination.

Without them, the shaman would be completely blind. Tables give coherence to the rolls and allow randomness to make sense within the tone of the world. A fairy-haunted forest encounter table is not the same as one from a war-ravaged wasteland. The possibilities themselves already say something about the place before any result even appears.

The tables are the GM’s ritual bones.

They establish boundaries, suggest directions, and force the GM to interpret signs within a framework coherent with the setting. Years ago, I prepared campaigns differently. I planned large narrative arcs, final revelations, and important scenes. Now I’m much more interested in discovering what the dice want for that specific moment.

That does not mean a lack of preparation.

In fact, I prepare extensively before a campaign even begins. I define locations, factions, biomes, rumors, conflicts, backstory, and creatures. I build solid foundations upon which chance can later operate. Right now I’m running a campaign in Dolmenwood, a setting built around an enchanted forest filled with detailed hexes, strange villages, factions, trade routes, herbs, and ancient ruins. The world exists before the players arrive and will continue existing after them. But within that framework I leave enormous space for chance. I rarely prepare full sessions anymore. I mostly review what exists near the adventurers and let the dice do the rest.

A few sessions ago, one of the characters was cursed by a fairy and grew donkey ears. The players asked where they might cure the curse. I looked at the nearby locations and, two villages away, there happened to be a church capable of helping them.

That church already existed beforehand.

It did not appear conveniently to solve the plot. The characters simply happened to be relatively close to a useful place. That matters to me. In another moment during the campaign, the group found venomous creatures inside a tomb deep within the forest. They bit the party’s warrior, and there was no antidote and nobody nearby capable of helping them.

And that was the end of his fate.

No NPC appeared at the last second to save him. No convenient divine intervention arrived. The world simply continued on its course. That may sound harsh, but it became one of the most emotional moments in the campaign. The warrior slowly faded away while saying goodbye to his companions around the fire. And I think moments like that exist precisely because the world does not protect the characters. The players know that if they survive, it will be because of their own decisions, their intelligence, or sheer luck. Nobody is going to save them because “the story needs it.”

And because of that, victories feel far more real.

In many modern games, the world constantly bends itself around the protagonists. NPCs exist to be used. Solutions appear exactly when the narrative requires them. I prefer to think of the world as something indifferent and persistent. Sometimes the forest simply contains things that were never placed there for the players. An abandoned tower does not need to hide an important clue. A merchant may have personal problems completely unrelated to any central plot.

A dragon does not exist to be defeated.

Things exist there for their own sake, not to solve the adventurers’ lives.

The GM, like an oracle or village shaman, knows more about the setting than the players do. He knows where the rivers are, which faction controls the roads, which forgotten god sleeps beneath the ruins, and how close the mercenary band pursuing them has become.

But he does not know the adventurers’ final fate. That is built at the table. The diviner knows there is a nearby river where the players might lose the trail of their pursuers. But he does not know how they will attempt to cross it.

Will they swim?
Cut down a tree?
Use magic?
Build an improvised raft?

And then the dice appear once again.

Perhaps the rolls will favor them. Perhaps a natural 1 will drag them downstream forever. This style of running games also makes sessions more interesting for me. It forces me to stay alert. To constantly interpret what the rolls mean and what the logical consequences of events should be.

Because this is not pure chaos either.

Improvisation works best when the world has coherence. Knowing what NPCs want, how creatures behave, or how far a mercenary’s cowardice extends helps improvised decisions feel meaningful. Skeletons fight until they collapse because they do not fear death. Wolves flee when outnumbered. Mercenaries may abandon their companions before dying for a handful of coins.

That kind of logic prevents divination from feeling completely random. The GM is not inventing arbitrary events. He is interpreting signs within a world that already possesses rules, tensions, and a will of its own.

The GM’s job is not to protect a story or guarantee dramatic scenes.

It is to honestly interpret what happens when the players’ decisions collide with the omens revealed by the dice.