What makes a good random encounter?

Encounters are information as much as obstruction

The oldest justification for random encounters is attrition — they make travel and delay cost something. That still works, but it is the weaker half. The stronger half is that an encounter table is a description of a region. Three entries about starving wolves, refugees on the road, and a burnt-out waystation say more about the state of the frontier than a page of gazetteer prose, and they say it while the party is doing something.

This is also what makes a table reusable. A table written as a list of creatures is exhausted once you have rolled through it; a table written as a list of situations keeps giving, because the party's response is different each time.

Four things a strong entry has

Any entry missing the second one is a monster, not an encounter.

• A subject Who or what. This is the part most tables stop at.
• An activity in progress What it is already doing when the party arrives. Something interrupted is a scene; something waiting is a stat block.
• A reason to be here One clause connecting it to the region. This is what turns the table into worldbuilding.
• At least one non-combat handle Something to trade, follow, avoid, warn, or exploit. It does not have to be the likely choice — it has to exist.

When to roll, and when not to

Rolling is worthwhile when the outcome could genuinely change the party's plan — during travel they chose, in a dungeon where noise carries, on a watch they set. It is not worthwhile as a rhythm section. An encounter check that fires while the party is walking to a shop they have already decided to visit adds time and removes nothing.

The related discipline is being willing to skip a result. A rolled entry that does not fit where the party actually is has no authority over the fiction; either place it somewhere it makes sense, or roll again. Random tables are a prompt, not an oracle, and treating their output as binding is how sessions end up with an ice troll in a summer marsh.

There is no universal correct frequency here. Attrition-focused play wants regular checks and meaningful resource loss; a table that plays for scenes and intrigue wants far fewer, weighted heavily towards encounters with people. Decide which of those your campaign is, and build the table for it rather than importing someone else's assumptions.

Table-building habits that pay off

• Weight towards the mundane Most entries should be traces, travellers and weather. The rare entry lands harder when the common ones are quiet.
• Include entries with no creature at all A cairn recently disturbed. Smoke on the next ridge. These are the cheapest way to make a region feel inhabited.
• Let entries reference each other The refugees on entry 4 are fleeing the thing on entry 9. Rolling both across a journey builds a story neither entry contains.
• Age the table After a major event, change three entries. A table that never changes tells the players the world does not either.

Before you put an entry on the table

• It is doing something when the party arrives.
• There is a reason it is here rather than anywhere else.
• The party has at least one option that is not combat.
• It reveals something about the region.
• You would be happy to roll it twice in one campaign.