The word "encounter" has taken on narrower meaning since the advent of Third Edition. An encounter has become a discrete unit that can be weighed and measured according to mathematical standards. I'm not knocking this approach; it's a great boon to DMs who want to stage finely balanced fights. It has, however, cloistered the term into a tighter cell than it had in the past. Sectioning PC/NPC/monster interactions into fights, skill challenges, and roleplaying scenes erects artificial walls where there should be only clear, smooth pathways.
3rd and 4th Edition D&D also placed far more emphasis on movement grids and miniatures than previous editions did. These have the wonderful effect of reducing confusion and arguments, but they also drive a wedge between the combat and noncombat portions of an encounter. The physical process of placing dungeon tiles or drawing the battlefield on a grid and setting up miniatures interrupts the natural flow of an encounter and reinforces the idea that what happens before and after that process are separate situations.
In fairness, this has as much to do with player perception as with what the rulebooks actually tell DMs to do. It's the toolshed phenomenon at work: when you get a brand new, cordless jigsaw, you make every cut with it, even when your simple, scuffed up old coping saw or keyhole saw could do the job quicker and cleaner. 3rd and 4th Edition provided tools for balancing combat that were great improvements over what D&D had in the past, but they created an impression that storing combat and noncombat activities in separate bins was a good idea. It's not.
It's unfortunate, then, that we use the term "random encounter" for this discussion. A new term—random incident or random event—would unshackle us from all that baggage. But random encounter has a long tradition behind it, so I'm going to try to rehabilitate those words.
The point of that long introduction is that it's easy and tempting to treat all random encounters as combat encounters, but it's also a mistake. I'd go so far as to state that if fewer than half of random encounters turn into fights, you're probably doing something really well.
To help you achieve that goal, here are a few keys to keep in mind.
Plan ahead. Just because they occur unexpectedly doesn't mean they're off-the-cuff. Whether you create your own random encounter tables or use someone else's, spend some time thinking about each possible encounter beforehand. If it involves NPCs, what do they want? Why are they on the move? What do they have to offer the characters, and what do the characters have that the NPCs might want? If monsters are involved, know how they typically behave. In a more formally built encounter, answering those questions is part of the design process. Random encounters don't need to be planned to that extent—improv is part of their appeal—but you should at least know the motivations of those involved, so you can act accordingly.
This is doubly true if you're using generic random encounter tables or those from a packaged campaign setting instead of tables that you designed. The person who wrote those tables knew what he meant, but you might not. Why would a big, flying predator like a manticore be in dense forest? Unless you have an answer handy when the encounter begins, your solution probably will be "roll initiative," and everyone loses out on what could be a very interesting exchange.
It's always wise to have easy access to whatever monster or NPC stats are required. That might mean copying stat blocks into a digital file, if you organize things that way, or just including a book and page reference on the random event list so you can turn directly to the stat block when it's needed without flipping through several other books first.
The amount of planning you need might be very different from the amount of planning someone else needs. If you're intimately familiar with the nuances of the setting, then you'll handle whatever the dice throw your way without breaking stride. If not—if you skipped over that portion of the guidebook that explains about the ruined manticore civilization in the forest, thinking you'd come back to it later—then the time to look it up is when you're reviewing the random encounter tables (or placing manticores on your own), not when "Manticores, 1-4" comes up on the table.
If you run combats on a battle grid, then thinking about the location is also an important part of planning. Many random encounters won't turn into fights, but many will. Some are guaranteed to. Be ready to sketch out the battlefield on a moment's notice. Prepare a half-dozen or so generic layouts that you can call on instantly and tweak as needed.
Spice it up. Another reason why random encounters have a bad reputation is because so many of them begin with "you meet some manticores." No one ever just meets some manticores. Conditions surround every encounter.
Older editions of D&D included tools for determining two of those conditions that are very important: range and awareness. That is, how far away are the monsters or NPCs when the characters first notice them, and who was spotted first?
Those two conditions affect everything that follows. If 50 goblins haven't noticed the heroes, then the goblins can be ambushed or avoided. If they are charging the characters but are a quarter-mile away, then the heroes can run for it, seek a strong position in the rocks, or gird themselves with protective spells.
A third condition is reaction. I lament the loss of NPC reaction tables from D&D. In their absence, DMs tend to default toward hostility, and players adopt arrogance as a standard response. A random reaction roll creates the potential for an encounter with 50 goblins that becomes a trademeet over shared ale instead of a battle with knives and claws. Obviously, a DM can steer any encounter in that direction if he chooses to, but we're back to the tools issue. When handed 50 heavily armed goblins by a random encounter table, most DMs' thoughts will not run toward peaceful intentions. A reaction table can generate odd combinations that lead to compelling roleplaying opportunities.
By combining range, awareness, and reactions, you can have any number of random encounters with 50 goblins, and each will be fresh and different with little effort on your part.
Harassing encounters: I don't advocate building random encounters with an eye toward harassing the party and depleting its resources on the way to the main adventure site. It's annoying, it seldom works, and even when it does work, it tends to be self-defeating. Players don't press on into danger with their arsenals half empty. If their spells or supplies are dwindling, characters just turn back to resupply or hole up to recover strength.
The only time harassing encounters are really dangerous is when regrouping is not an option or resources are already low. The ideal moment is after the climactic fight, when the big spells and daily powers are exhausted. That's when characters who shot their bolts can really be challenged by a wandering monster.
Random events. As noted above, it would be better for everyone if DMs and players thought in terms of random events instead of encounters. Random events can be anything:
- fights against monsters
- wary trading sessions with goblins
- chess games with manticores
- swapping news and buying supplies from passing merchants
- an abandoned mine entrance
- the bodies of NPC adventurers who died around their campfire from drinking spoiled wine
- screams in the night that no one can quite identify or zero in on
- a boy looking for his runaway pony
- a werewolf seeking to buy silver
- an amazing auroral display
- a flash flood
- mastodons in a herd that takes hours to pass
- someone's coin purse disappearing
- a roadside shrine
- a blind crone led through the alleys by an albino rat
- an earthquake or landslide
- a forest fire, or a fire that breaks out in the inn or stable overnight
- news of the local baron's death
- a dog that follows the characters everywhere and stares at them unnervingly
- a military press gang sweeping through town
- a wanted poster with the heroes' description
In other words, just about anything you think of can be put on a random "encounters" list. Good examples are often easier to find in board games than in RPGs. The D&D board games, for example, include excellent random events. They occur in board games far more often than you'd want in a roleplaying situation, but that doesn't diminish their value as examples.
Some of the items on the list above are fine adventure hooks, and indeed, they can grow into short or full-blown adventures. Why, then, would you put them on a random encounters list instead of treating them as scripted adventures in the first place? Because by letting them occur "naturally" as random encounters, the decision whether they are or aren't the stuff of an adventure is put in the players' hands instead of the DM's. This touches on a larger subject of randomness in scenario design, which I'll come back to in the future at much greater length (probably until you're sick of hearing about it).
The point I want to stress is the fundamental and huge difference between dropping events into the heroes' laps with a wink and a nod that says, "here's your next adventure," versus having small things happen around the characters from which the players choose their own paths to adventure. The ideal approach is a mix of both. You can have your ongoing story that stretches across months of real time with consequences that compel characters to action (or pull them along by the nose, in the worst case). But you should also sprinkle their path with minor events and encounters that can spice up the journey, distract the characters from their major troubles, let them solve the occasional small problem or interact with the world in ways that don't shake its foundations, give them a sense that they are steering their own course, and just maybe veer onto a path that everyone will find intensely satisfying while it lasts.
With that in mind, the final piece of advice is, don't lock yourself into thinking of random encounters as small and limited. Keep your mind open to all possibilities. If players determine to follow up on a random incident, let them. Be willing to let them lead for a while. It's their game as much as yours, and nothing sustains player interest as much as the feeling that they're steering their own destiny.