If you believe countless Dungeon & Dragons adventure reviews, nothing ruins an scenario as quickly as a linear design. In a linear adventure every group follows the same plot thread, through the same scenes, to the same conclusion. At best, critics accuse linear adventures of eliminating the players’ choices between scenes. At worst, critics say linear adventures require dungeon masters to abuse their power to shunt players along a railroad. Instead of steering the adventure, players follow a fixed story.

In my last post, I explained that players don’t hate linear adventures as much as reviewers and dungeon masters think. We tend to judge harshly because we see the lack of options. In a successful adventure, players never see the walls.

Many gamers conflate linear adventures with railroading, but that mistake tars decent adventures. Players seldom mind linear adventures, but few players tolerate railroading. DMs who railroad deserve player complaints.

“Railroading is not linear prep,” Phil Vecchione from Gnome Stew explains. “Railroading is the game master’s reaction to a player’s action, in an effort to drive the game in a specific direction. That reaction is to typically negate, reverse, or shut down a player’s action, in order to get the game moving in the GM’s desired direction.”

A successful linear adventure invites certain choices and makes assumptions about outcomes, but it never forces a result. Some of the success of an adventure depends on the designer’s ability to predict choices and outcomes. (See Actions Players Always Take and Choices Players Never Make.) When the predictions fail, adventures tempt DMs to railroad.

In some game situations, when the players veer from the plan, the temptation to railroad becomes nearly irresistible. These situations appear in nearly every DM’s career. Instead of succumbing to temptation, what should we do?

1. When an action leads in a direction you never anticipated, improvise.

Every DM eventually faces a player decision that nullifies all the planning that prepared for a game. “If you can’t improvise, you’ll eventually hit a wall you can’t climb over, or find yourself trapped in a corner and unable to talk your way out,” D&D senior producer Chris Perkins writes.

“Improvisation demands equal measures of intuition and confidence. DMs who lack sufficient intuition or confidence tend to have trouble improvising at the game table. The good news is that DMs, being creative souls, rarely fall short in the intuition department. They know a good story from a bad one, a well-developed character from a cardboard cutout, and so forth. However, confidence is a far more rare commodity, and DMs who lack the confidence to trust their intuition often have trouble improvising behind the DM screen. I know because I’ve been there.”

Entire books aim to help game masters improvise, but often the trick comes down to making a leap into the unknown–or unprepared. If you find yourself stuck, call for a break, spend a few minutes finding inspiration, and then go with the idea that seems most fun.

2. When an action may deliver an easy win that cuts an adventure short, reward the ingenuity and then add complications.

Mike “Sly Flourish” Shea likes to note that we DMs match our one brain against the players’ six brains—a serious mismatch. Sometimes players invent a plan that threatens to skip the middle of an adventure and deliver an easy victory.

As one possible response, DMs can grant players an easy win. Players relish a chance to thwart the villain (and the DM) with an ingenious plan. Perhaps the characters built around deception or infiltration get to shine.

But an easy win can’t cut a 4-hour convention adventure to a half-hour assassination. If you want to spare more of your preparation, reward ingenuity with some success, and then add reasonable complications that make the plan unravel. Perhaps the players reach the throne room, but discover that the lich-king has left to sow terror and whatnot. Now the players must decide how long they dare to wait in the heart of an enemy stronghold.

Many ingenious plans start with the players impersonating the villains’ cohorts and seeking a free pass to the boss’s war room. In those scenes, I seek ways to help the heroes while also stirring trouble. Often some of the party must pass as prisoners or foreign allies. What if guards demand to take prisoners to the dungeons and allies to the rest of their delegation? Sometimes I add tests of loyalty. “We’ll take you to the Prince of Murder, but first help us execute your prisoners.” Even when a deception succeeds, such tests suggest that smart foes hold some natural suspicion.

3. When players try to start fights during an interaction scenes, pause the action.

Sooner or later, every dungeon master sees players stop a role-playing scene to start a fight. As the party talks with a scheming queen, a player blurts out, “I hit her with my axe!” Picking a fight during such an interaction typically causes problems because the adventure expects the queen to bridge the way to the rest of the adventure. The attack burns the bridge and leaves players running from her heir and her army.

These scenes tempt me to add a hidden pit trap between the charging barbarian and the queen. Actually, that makes some sense. If I were boss of some D&D land, my throne room would feature a trap door.

If the archer or warlock launches a ranged attack, every DM feels tempted to turn the queen’s guards into invincible minor minions who crush the party. Then the queen threatens to hang the characters unless they continue the scheduled adventure. Steal from the classics.

A better, non-railroad response includes 3 steps: (1) pause the attack, (2) learn the root cause of the attack, and (3) reroute the adventure.

Instead of letting the instigator roll damage, pause the action.

If only one player wants the fight, Teos “Alphastream” Abadia explains how he lets the party intervene in-game. “I’ll freeze time. ‘Everyone can see that your character is about to kill this person. Everybody has a chance to stop this. What do you all want to do?’” Teos makes it clear that the single player stands alone against everybody else in the party. See What To Do When A Player Interrupts A Role-Playing Scene To Start A Battle.

If the whole party seems eager for battle, look for the root cause of the attack.

Perhaps the players see the queen as a bad ally, so when the adventure leads to an alliance of convenience, the players rebel. Murder in Baldur’s Gate assumed characters would support one of three patrons who vied for power. The patrons start unsavory and, as they gain power, become worse. My players wanted no part of it. I needed to find a more agreeable patron.

Often, the players see the queen as a villain they will fight eventually. Why not solve the problem now? As DM, tell the players how their characters’ lifetime of experience in the game world reveals flaws in the players’ plan. “The reputation of the queens’ knights leads you to believe that they can easily defeat you.” If the players attack anyway, finish the fight, and then find another patron.

4. When plot features recurring villains, but the party blocks their escape, plan for escape, but prepare for a new villain.

Every DM loves a recurring villains. But to establish one, you need to introduce the villains and then somehow invalidate the players’ decision to murder them.

Typically, DMs underestimate the players, and so potential recurring villains die during their first scene. Our odds stand at 6 brains to 1. As a slim advantage, we have time to plan. Don’t make a potential recurring villain the most threatening target in an encounter. Don’t leave the villain exposed between their turns. Plant a potential barrier along the escape route. (I’m not above an escape via spinning bookcase.) Start the escape while minions remain to block pursuit and while the villain still has enough hit points to survive the players’ biggest attacks. Accept the (probably) inevitable premature death. Prepare to call another foe from the bench.

5. When the story includes the players’ capture, but the players win instead, wait for another chance.

The most egregious crime of railroading comes when a DM wants players taken captive. In adventure fiction, heroes get captured regularly. So DMs dream up similar stories, and then try to force a capture despite the players’ determination to never get taken alive.

To engineer a capture, DMs needs to hide an encounter’s threat to the players, block the characters’ attempts to flee, beat any signs of an unexpected rally, and so on. During all this, if the players see signs of their DM bending the odds to thwart their escape, they will feel railroaded. You can’t plan for a capture.

None of this means your players’ characters will never be captives.

Use capture as an alternative to a total party kill. Save your escape-from-the-dungeon scenario for a time when players ignore warning signs, make bad choices, suffer setbacks, and ignore any chance to run. Those times happen—trust me. Then, instead of rolling new characters, have the old characters wake in chains. The players will feel grateful for a second chance.