If you searched “how long should a D&D session be,” here’s the answer up front: three hours, with a hard stop, and if your real life only has two, two is a full-sized session and you can stop apologizing for it. The rest of this post is why that’s true, why the internet keeps telling you otherwise, and what to do with the slot you actually have.

Where the four-hour number comes from

Ask around and you’ll hear four hours like it’s in the rulebook. It’s in nobody’s rulebook. It’s an average of convention slots, actual-play shows, and the schedules of people who played in college, and all three of those are terrible models for your Tuesday night.

Convention slots are four hours because that’s how you fit two games in a day on a trade-show floor. Actual-play shows run long because they’re staffed productions with a cast that’s paid to be there and an editor who cuts the slow parts. And the college game went five hours because everyone lived four minutes away and had a 11am class at worst. None of these people had to relieve a babysitter.

Your table is five adults with jobs. The question worth asking about session length has a different shape: what’s the longest slot this specific group can hit every single week without anyone quietly starting to dread it? That number is the right length for your campaign. For most working-adult tables it lands between two and three hours, and the campaigns that die are very often the ones that refused to believe that.

What actually happens in hour four

I’ll describe it and you can tell me if you’ve been there. It’s 10:40pm. The fight has been going for fifty minutes because everyone takes ninety seconds to remember their bonus action. Two players have their phones out under the table, politely. The DM is narrating faster, skipping the flavor, because there’s a door the party was supposed to reach tonight and it’s still two rooms away. Somebody says “I should head out soon” in the apologetic voice, and the whole table does the math on whether to wrap mid-fight or push through, and either answer feels bad.

Nothing in hour four is as good as it would have been in hour two. Decisions get worse. Jokes get fewer. The DM’s improv, the thing the whole game runs on, degrades the same way anyone’s judgment does at the end of a long day on top of a long day. That hour buys you a worse version of the game you already had, and it charges the cost to how everyone feels about next week.

Long sessions feel generous while you’re scheduling them and expensive while you’re in them. The schedule that says “epic five-hour Saturday session, monthly” is a campaign that meets twelve times a year and loses the thread between every meeting. We wrote about that memory gap in when nobody remembers last session, and session length is one of the quiet causes: the longer and rarer the session, the more each one has to carry, and the more gets dropped.

Short and weekly beats long and sometimes

Here’s the arguable position, stated plainly. A two-hour session every week is a better campaign than a five-hour session every month. It’s better paced, better remembered, easier to prep, and far more likely to still exist in a year.

The math is lopsided even before you count quality. Weekly two-hour sessions are over a hundred hours of play a year. Monthly five-hour sessions are sixty, and that’s if none of them get cancelled, which they will, because a five-hour block is the first thing a busy month deletes. A short slot survives scheduling pressure the way a long one never does. It fits after dinner. It fits on a weeknight. It doesn’t ask anyone to burn a Saturday.

And the week matters as much as the hours. Momentum in a campaign comes from frequency, from the story being a regular visitor instead of an occasional event. Seven days is short enough that the table still remembers the cliffhanger. Thirty days is a cold start every time, and you spend the first forty minutes of your precious five hours reheating what the group forgot.

The shape of a good three hours

Whatever length you pick, the shape matters more than the duration. A good session has a warm open, a middle that moves, and an ending that lands on purpose. Here’s how that fits in three hours:

The first ten minutes are the recap. A real one, in the fiction, not “so, uh, where were we.” This is the highest-leverage ten minutes of the night and we’ve written a whole guide to it. On a short slot it matters even more, because you can’t afford the half-hour of collective remembering that passes for an opening at most tables.

The middle is two scenes, maybe three. A social scene and a fight. A fight and a discovery. On a two-hour slot, two scenes is a full meal, and knowing that transforms your prep. You stop building four-encounter nights that never finish and start building evenings that fit. Our 30-minute prep method assumes exactly this: prep scales with session length, and a short session is cheap to prep well.

The ending is chosen, not collided with. Pick your stopping beat about twenty minutes before the hard stop and steer to it. End on a reveal, a door, a sword raised and not yet swung. Ending on a comma is what makes a short session feel complete instead of truncated, because the table leaves mid-story on purpose, hungry, instead of trailing off when the clock runs out.

Do those three things and a two-hour session feels dense. Skip them and a six-hour session feels thin. Length was never the variable doing the work.

The hard stop is a kindness

Announce the end time at session zero and keep it like a train schedule. This sounds like it limits the game, and every table that adopts it reports the same two surprises.

First, the game gets better, because a known endpoint forces pacing. Fights tighten. Shopping episodes stop sprawling. The DM steers, because steering is now part of the job.

Second, attendance goes up. People flake on open-ended commitments and show up for bounded ones. “Tuesday 7 to 10, we always end at 10” is a promise a person with a job and kids can make for a whole season. “We usually go until whenever” is a promise nobody can make twice. If your campaign has a chronic attendance problem, check the end time before you blame the players; an unbounded session punishes exactly the responsible adults you most want at your table.

The hard stop belongs in your session zero scheduling talk, next to the day and the cadence. Our session zero guide treats scheduling as the top killer of campaigns, and session length is scheduling wearing a cloak.

So, concretely

Weeknight table of working adults: 2.5 to 3 hours, weekly or biweekly, hard stop, and guard that slot like it’s load-bearing, because it is. Weekend table: 3 to 4 if you genuinely have it, and watch hour four honestly for a month before you decide you do. Online table: knock 30 minutes off whatever you’d play in person, because screen fatigue is real and Discord hour three feels like in-person hour five. Play-by-post and two-player duets live by different clocks entirely and can ignore all of this.

And if the honest answer for your group is ninety minutes on a Thursday: run it. A ninety-minute session that happens is worth infinitely more than the four-hour session you keep failing to schedule. The campaign that’s still alive next November will be the one that fit inside your life.