I've participated in games that have run for vastly different time periods. Some games I've been in or run have been epic in scale. Some have been so fast and intense that they've been a single memorable blip. I've been in games that ran too long, games that ended too soon, games that were unappreciated in their time, games that seemed great in the moment but were forgotten as soon as they wrapped.
Along the way, I have developed some opinions on how long a game should be.
Essentially, most of my answers are going to be somewhat vague, because different games require different things, but most of my answers will be phrased along the same lines. What should make a game longer and what should make a game shorter.
How long should a game be?
The basic answer is simple: long enough to tell the story. But that isn't a firm answer. There are stories that can unfold in three sessions and stories that take five years to tell. Overall, a game will suffer if it is compacted too much or stretched too much. A game that's too compacted will feel rushed and can be confusing too much detail in any one point, no time for processing. A game that's too stretched may end up losing momentum or having players not see how things fit together. How do you know how long the story takes? That's a hard question. I tend to break down stories into steps or dramatic beats. In order for the neighboring kingdom to invade, they need to manufacture a casus belli, amass troops, move those troops to the border, seize the first line of defense, the second line of defense, find a way to break through the final line of defense and then seize the capital, install a puppet government and demobilize their army except for an occupational force. So something as simple as an invasion has 9 steps to it. Some of those might be combined, some might require more than one game, but having the steps of the process laid out can tell you that the story of the invasion of a kingdom is going to take at least 8-10 months. Every story can be broken into parts, some big, some small.
Beyond just the story length answer are a number of other factors.
I think a LARP should be short enough to maintain a consistent central cast. There should be people there at the beginning who see the end. And while a little cast rotation can be a good thing in a healthy game(eliminating those who aren't interested, adding those who found out late), a game that can maintain the continuity of a central cast is generally better than one that can't.
A game should be short enough to avoid excessive power creep. I've spoken about the perils of experience systems. The longer games go on, the more power tends to become consolidated/imbalanced. This isn't just an XP problem, it can be a plot access problem. When the older players have lots of plot-centric influence they've built over chapters 1-4, those who start in act 5 may feel marginalized.
A game should be long enough for people to settle into their characters and factions. In games where players create their own characters, there is generally a period of settling into the character. In games with political factions, it takes time for groups to gel. A game should be long enough to allow that process. Some of this can be taken care of with pre-work, but pre-work is just that, work. It's much more interesting to play out a massive internal factional schism than to simply talk about it.
A game should be short enough to avoid staff burnout. It takes so much work to run a game that after a while, the staff can get very tired. Very long projects can feel stifling, as staff members(who tend to be creative project types)want to move on to the next project rather than lingering on the current one. This can be addressed by adding new staff, but that can dilute the central vision of the game.
A game should be short enough that every part serves a purpose. Not every game should be plot heavy episode after plot heavy episode, but there should be a minimum of filler content. Some sessions may be designed to be light on the eventfulness, a chance for people to explore facets of their characters or mix up the general patterns of the game. Things like bottle episodes or tabula rasa type games can be great for resets. IC event games like weddings or funerals or holiday celebrations can be fun in a game situation.
A game should be like an ideal gas, it should expand to fit its container. Running games in many contexts means tailoring the game to the time constraints. If you're running a game at a con, it should be a quick one shot, set up to develop and resolve within the time slot. If you're running a game at school, it should be able to be set up and completed in a semester. I've tailored timelines for games based on local weather patterns.
Overall, finding the right length for a game is a complicated process. I've already stated my opinions on bounded storytelling, but this is a great time to reiterate that. A game should have an end. Endless games are shapeless and memorable not as a whole, but as parts. If you're going to treat the game as a whole, you should think about the entire shape of it.
