You could not call such bloodshed a victory. Dwarven pike had met draconic spear wall, neither buckling until lines were met. Hradakh had been in the thick of it, stabbing his way through the brutal melee, and when the fighting stopped, he and his dragonborn kin held the field.

What few of them had survived anyway. The field was carpeted in blood and bodies, and perhaps one in three of his comrades stood unaided. Another victory like this and there would be no war for lack of troops to fight it.

By default, combat in 5th Edition–derived systems is a deathmatch. Initiative is rolled, and combat ends when one side runs out of combatants. It is what most rules and encounters are written around, what most groups of players are familiar with, and the basis of monster and encounter design. It is also an invisible limiting factor on encounter and monster design.

Combat-as-deathmatch puts constraints on what outcomes in a fight are acceptable as well as what sorts of storytelling are mechanically supported. All enemies in a battle are likely to die, and even when taking prisoners is an option, rules for doing so don’t exist. Conversely, a player loss is defined as a party wipe, and players are mechanically encouraged to fight to the end instead of surrendering or retreating. This in turn shapes how combats, enemies, and adventures are designed.

If the players losing threatens to end the campaign and all enemies are expected to die in a fight, then campaigns need to be designed around the assumption that players will win most (if not all) of their fights and in the process kill the hostiles opposing them. Losses need to be rare and ideally have some sort of mitigating factor to keep players alive. A way to retreat or enemies who are uninterested in killing them. Similarly, if you want an enemy to show up more than once, they need to either avoid combat or have some way to escape it. The plot needs to at least function when the players stand a reasonable chance of killing any villain they are in stabbing-or-casting distance of. Hostiles and locations are designed around this paradigm, an unspoken axiom that pervades many elements of how the game is designed and played. While there are optional rules like morale and guidelines in the community to encourage the use of enemies who flee when they’re losing, these mitigate the problem—they don’t solve the core problem.

If those constraints are comfortable for your campaign, this isn’t an issue. However, when you try to work around them, you are likely to run into problems. It can feel like the game system is actively working against you, like you’re constantly balancing between absolute disaster and fights being inconsequential. Fortunately, this is a problem that can be solved without homebrewing.

The key is in breaking away from the unspoken default. Designing fights that aren’t meant to be to the death. Ones based around a discrete objective and that end when that objective is completed—whether or not the players win. Fights where the penalty for loss isn’t necessarily death but could be the loss of desired resources, a need to find a new route forward, or other plot-facing repercussions. One where enemies might show up again and again, facing off against the players over some discrete goal and retreating when they fail to accomplish it.

In short, you want to introduce skirmishes. A fight around a concrete objective where most of those involved will live to see another day. This is a tool in your arsenal as a GM and storyteller rather than a new default. Something to be mixed in to enable more storytelling options, introduce more dangerous (or merely recurring) adversaries, or promote a real chance of failure without risking a total party kill.

That said, most of the guidelines in the game revolve around deathmatches, and most players and groups are used to them. When applied to skirmish design, much existing guidance is going to result in fights that are too easy or where the optimal solution is still to kill every hostile on the field. Similarly, increasing the danger of opponents can result in lethality due to a systemic focus on physical danger as threat guidance.

As such, designing skirmishes is harder than designing deathmatches. You’ll want to be in constant communication with your group as you introduce them. See how people feel about the encounters and whether some players feel like they aren’t getting to do things due to how the encounters are designed. You’ll also need to consider how to initiate and end combats and how to set up combat goals and enemy roles around those goals to create satisfying and interesting encounters.

While this series will eventually cover the entire process in more depth, here’s an overview of skirmish-based combat design to tide you over until the next entry:

• What is the objective of the skirmish? Are you fighting to escape, protect a point, complete a ritual, or kill a specific target? How many turns might it take for the players to complete this objective, and are they on a time limit?
• What ends combat once the objective is completed? Will the enemies retreat or will the PCs? Is there a steady stream of hostile reinforcements on the way? Is the building on fire? Or is part of the objective ending combat?
• What does each enemy contribute to this fight? Are they threatening PCs, controlling an objective, or simply limiting movement options? Each enemy should be doing something even if that’s just “making it dangerous to complete the objective.”
• How does the terrain affect the fight? It doesn’t have to, but time limits, burning buildings, chokepoints, and weirder terrain effects are excellent ways to control encounters and escalate tension without simply adding more hostiles.