What makes a Sandbox?

[howandwhy99]

If there's some series of murders for a ritual (and a few more murders remain) or the bad guy has a time table of other tasks and places to be, there's some linear elements. Does that in turn makes writing the adventure less "sandboxy" and more AdventurePathy?

I suspect time tables and such encourage a certain style of content writing, namely adventure path.

A timetable for NPCs actions is just the scenario as written for any given session. It isn't a path to be followed by the PCs. Rather they tend to derail this timetable. I would say, any published adventure that presumes a set sequence of events will happen regardless of PC interference is not a sandbox adventure. A predetermined sequence of events over many sessions, an AP, makes even less sense to me as none of it occurs as a consequence of player decisions. IMG, any session's scenario timeline changes because of the involvement of the players. This means a new scenario is needed each session. Something I cannot know before the end of the previous session.


Also, I think it would be wise to also define "Plot Hook". In my view, they are portion of the game world where a particular web of NPC plots may be entered into by the PCs. This could be anything from walking into a thieves guild, hiring on for a service, or attempting a goal in conflict with NPCs' goals. The more complicated and larger in scope plots tend to be more interesting, but being small doesn't make something not a plot either.

I'm off to game...

 

[Janx]

If it is, then it is less far toward sandbox along the sliding scale between sandbox and linear than some other campaigns. At the far end, closest to full sandbox, the area is only defined by where the PCs choose to go. This is pretty rare. While something might look like a full sandbox, there still might be limitations like the inability to planar or time travel. A campaign might be suggested to be "anywhere on the map" but in reality only allow for certain areas of the map to be explored because of physical limitations (a mountain range) or inobvious social limitations ("Evertime we get three miles from town, one of my PC's relatives is kidnapped!"). Each restiction nudges a campaign along the scale farther from full sandbox mode and closer toward the linear mode.

Unless you're playing in a world where the PC is a god and can make anything happen, a PC is always operating under constraints.
if he has 10HP, he can't take 11 damage and walk away from it
if he has 10GP, he can't hand out 11GP to a shopkeeper
if he has no other means, he can't travel faster than by foot

Those would be "rational" limits within the context of the world.

"Evertime we get three miles from town, one of my PC's relatives is kidnapped!" is a better example of the DM over-riding rational capabilities to create an artificial limit. And its more of actively blocking, rather than a limit. Actively blocking is a railroading symptom.

Ignoring railroading symptoms as the extreme, I would think however, that "making stuff up" to entice the PCs to do the DM's thing, instead of giving them opportunity with no stick to choose their own path is anti-sandbox.

Thus:
giving the military PCs orders to travel to XYZ and do ABC
kidnapping a relative by an enemy of the PC
a dame with legs that don't quit drops by the office with a problem...

these are valid events in a game. They tend to be the tools of non-sandbox DMs to route the party into the planned adventure.

therefore, is it true that:
a sandbox has fewer events that route PCs to a path, especially "out of nowhere"

Basically, filter out "stuff you can't do because your PC doesn't have the means", as not relevant to the conversation.

I get the sense that a sandbox is "prepared" for the PC try to do whatever he has the means to do, though a PC might not actually attempt or consider it.

Thus, while the paladin will "never" attack the city guards or rob a bank, if he did, the sandbox GM would roll with it and would just calculate the reaction of the world to his actions.

Whereas an AdventurePath GM will have a what-the-heck moment.

 

[Janx]

A timetable for NPCs actions is just the scenario as written for any given session. It isn't a path to be followed by the PCs. Rather they tend to derail this timetable. I would say, any published adventure that presumes a set sequence of events will happen regardless of PC interference is not a sandbox adventure. A predetermined sequence of events over many sessions, an AP, makes even less sense to me as none of it occurs as a consequence of player decisions. IMG, any session's scenario timeline changes because of the involvement of the players. This means a new scenario is needed each session. Something I cannot know before the end of the previous session.

I get what your saying and that's where I was heading, I think.

My hunch is that if you have a timeline of what's going to happen, either within the session, or across sessions, it is probably not a sandbox trait.

Timelines are certainly a tool, but they have their own risk (railroad bait) and they can be harder to abjudicate as changes are needed. They tend to encourage a DM to try to minimize the damage to maintain the timeline (just as some sci-fi fiction insists that you "can't change the timeline"). This leads to blocking, which defeats player choices, which is the crux of railroading.

Timelines are almost something you have to write down, to run and manage. which definitely puts in the "published adventure" domain. I think the introduction of events and timelines in published materials is what shifts the game from sandbox to not-sandbox.

Another thing I'd consider out of the topic is "stuff that defines the world" or "stuff that's in the background".

Ravenloft says you can't leave it. It's the nature of the world, though that doesn't make it Not a sandbox (just not a very big one)...

Having a timeline that the Titanic will sink on a certain date is just background fluff if the PC aren't in a position to change it or interact with it. If the PCs are ON the Titanic, then the timeline of events is relevant. If they're on dry land, doing some other mission, and they get back and hear that Aunt Nelly was on board and was lost, that's just fluff.

 

[pawsplay]

I get what your saying and that's where I was heading, I think.

My hunch is that if you have a timeline of what's going to happen, either within the session, or across sessions, it is probably not a sandbox trait.

My feeling is that if you don't have a timeline of what's going to happen, nothing is happening, and the characters live in a big, blank world that bears little resemblance to reality or is very boring.

 

[Janx]

My feeling is that if you don't have a timeline of what's going to happen, nothing is happening, and the characters live in a big, blank world that bears little resemblance to reality or is very boring.

I agree.

If you run a dungeon with rooms of monsters, and the monsters never leave the rooms, the first time experiencing D&D, the players may not notice how "unrealistic" it is.

And when they get to town, if nothing's going on, but shop keeps buy unwanted gear and sell new gear, that's also pretty basic, but gets old.

As DMs and players get more sophisticated, they try to make it feel more alive.

That happens by moving the pieces around and "making stuff happen". Making stuff happens looks like events, but they may be events running on a timeline.

 

[Nagol]

I don't know where this line of questions will go, but I'm hoping it reveals ways that a timeline or murder mystery can fit in a sandbox.

How did you determine the murder was going to happen?

My sandbox is composed of the geographics areas in question, thier contents, and the intended actions of its inhabitants. As part of the intended action, I create a timeline of events that are likely consequential to the sandbox. In this case, one faction needed the baron gone. I had a timeline of actions thta culminated at a particular date with his death if the players did not interfere.

Was it planned before the session?

Yes.

Was it trigger based (you had notes that said "the day after the PCs show up, this NPC is murdered")?

No. The timeline is a set of actions, their date, level of success, public results, and expected consequences / new situations caused by the event. Player activity can modify the timeline or results depending on player action.

It is possible in some cases for a trigger to be set. For example, party A needs to accomplish some hazardous task, but doesn't have the ability. Party A will approach the first group to appear to ask for aid. Often, these triggers will have an review date or expiry date on them. I.e. if the PCs do not make themselves known within 3 weeks, Party A will set forth to the larger town to look for aid.

Was it something you made up on the fly, as it made context within the session?

I knew the basics -- the predetermined level of success, the level of ability of the assassin, and the expected strategy to prevent simple resurrection. Running the session was done on the fly including a couple of points where the party could have modified the result.

Does a timeline of planned events interfere with a sandbox?

The timeline is part of the sandbox. Environments evolve over time so should the sandbox.

What I mean is, between sessions, it's pretty easy to look at what the PCs did, and make adjustments to the world, move stuff around, make up new content as "reactions". If the PCs clear out a dungeon and 6 months go by, you can fill it with something new. If they whacked an overlord, somebody new fills in the vaccuum.

To the players, that looks like it was all planned and after the fact, there's a cause and effect and order of events.

Within a session, moving some orcs around, because they heard a noise in the dungeon, or the guards didn't report in also make sense, and the players might think there's a timeline, but there's not.

Running an actual timeline is a lot trickier, because a deviation early on can change everything, throwing off a chain of thought the DM had (which I believe can cause railroading).

Consider:
the butler (who will be the murderer) knows the household goes to sleep at 10PM.
at 10:30pm, he grabs the candlestick from the dining room
11:45 he listens at the master's door for snoring
11:55 he opens the door and sneaks in
12:00 he strikes with the murder weapon, killing the master
12:05 he wipes down the candlestick (missing a bit of blood in a groove)
12:10 he joins some stable hands in a game of poker, as an alibi, knowing they don't tell time too well
2:00AM he puts the candlestick back in the dining room
8:00AM the maid discovers the master is dead
9:00AM the inspector arrives to examine the scene

I'm not saying a murder mystery is run this way, only demonstrating a timeline for a murder, pre-PC interaction.

If the PCs are staying at the house that night as in Nagol's example, the original "murder timeline" might play out that way. It's even possible to abjudicate that the PCs slept through the murder like everybody else and awake in the morning to deal with the environment of the murder mystery itself.

However, its also possible (especially in a more complex case) that the PC are active, and thus can alter the timeline. That in turn presents a challenge in a complex case where the DM has built a mystery to be unraveled by controlling the variables to set it up.

Basically, a timeline before the PCs get involved is no big deal. That's fluff to help the DM make it make sense.

A timeline woven around the PCs is a bit more complex and might have some non-sandboxiness to it.

My timeline for the event would be much less detailed. It would look like this:

Date 99/99/99 Event -- murder by butler
Butler will attempt to coup de gras master overnight with candlestick from dining room. Establish alibi with stablehands by joining card game. Then restored cleaned weapon to dining room when safe to do so. Maid will discover body during normal rounds in the morning.
No intervention: success

The butler's personality will dictate how much interference he'll accept before adjusting/cancelling the event.

Think of the timeline as a guideline of how the starting situation will achieve equilibrium (or at least flip to the next state) in the absence of PC action.

Player interference forces the actors to adjust their actions, timings, and goals. In other words, the players choices have effect on the game world.

I might also note that a timeline is not the same as a time limit. The murderer having a ticket to board a train at noon tomorrow or the room filling up with water is a time limit. Time limits are fairly easy to abjudicate. WHereas each element in a time line assumes the outcome of the previous elements. Change an element and the whole thing is at risk.

This is unlike a dungeon, where by DM laziness if a PC clears out room 1, it doesn't have to affect room 2.

A timeline is one way to add a fourth dimension to the sandbox -- the PCs aren't the only forces dynamically affecting the environment. The universe continues to evolve in a plausible fashion in the absence of the PCs. The PCs could enter a dungeon and return a week later to discover the overthrow of their king. some events are known in advance and the PCs have the luxury of proactive action. Others come as a surprise and the PCs have to react.

 

[pawsplay]

When NPCs are acting on a timeline, I call that a trajectory. I know where they will land if their action is not interfered with.

 

[Umbran]

Hm. Take a look back at the OP for a moment. "What elements must be present in a game for it to be a sandbox?" Looking for definition by inclusion.

Much of the following discussion has been definition by exclusion - looking for things that make something "not a sandbox".

If it looks like a duck, walks like a duck, and quacks like a duck, it is a duck. If the thing also has lasers, that doesn't make it not a duck. It is just a duck with frikkin' lasers.

One can take the question one step further - What can a GM do to foster and use player-direction?

 

[Beginning of the End]

I'm arguing that if there are no dragons in a sandbox setting and you want the DM to include some then you should be able to talk to him out of character.

Then why are you arguing with me? What part of, "There's nothing wrong with openly communicating with your GM about what your goals are and what you'd like to see in the campaign..." didn't you understand?

It's simply not true that the only way to do a dragonslaying adventure is to:

(a) Tell the DM you want dragons; and then
(b) Sit on your ass and wait for him to send somebody around the local tavern to tell you where the dragon is.

It is if you've looked everywhere for dragons and not found any.

OTOH, what part of "...but that communication with the GM doesn't require that your character sits on his ass and waits for God to deliver something interesting for him to do as if it were a pan pizza." didn't you understand?

You're insisting that "OOC communication about what people want in the campaign" and "sitting on your ass waiting for the GM to hand you an adventure" are inseperable. They're not.

In fact, the pairing seems quite odd. If you're being proactive OOC (by communicating what you want), why wouldn't you also choose to be proactive IC (by going out and looking for what you want)?

Correct me if I'm wrong.

Allow me to join the legion of posters saying, "Been there. Done that. Not going to waste my money on the t-shirt."

But does it invalidate a campaign as 'sandbox'? Or does it just make the campaign less of one?

This is probably a good opportunity to take a step back and revisit a point that Snoweel has been ignoring:

(1) There are sandbox-compatible techniques.
(2) There are techniques that are inimical to sandbox values.

A sandbox campaign is defined by being predominantly made up of #1.

But this is not a purity test. It's simply not useful to treat "sandbox" as if it were the campaign equivalent of sexual virginity.

By the same token, it's ridiculous to suggest that railroading, for example, "makes a sandbox" because any given sandbox campaign can tolerate a little bit of railroading without ceasing to be a functional sandbox.

A discussion about what techniques are sandbox-compatible, sandbox-beneficial, and sandbox-inimical seems valuable. Trying to come up with some sort of meaningless purity test doesn't seem valuable at all.

Large scale collaborative world-building, ie giving the players authorial input to the setting, is inimical to the primary purpose of sandbox play, which is exploration (and possibly conquest) of the sandbox.

There are different degrees of input, however. There's a continuum between "I want to fight a dragon!" and "Here's the lair of a dragon that I mapped up for my PC to go explore". The latter end of that scale would, IMO, move us pretty solidly out of sandbox territory; but the other end of the scale wouldn't.

But if the "maze" is in front of a screen being drawn up by players together, than I do not believe the game is a sandbox game. That's just my honest opinion. I see the other as a game of making a maze together and not solving one, a necessary component to external discovery. Are the players discovering their own wants and desires in such a collaborative game where they jointly draw up the maze? Sure, but there are no unknowns about the fiction to explore. In fact, unknowns would be detrimental to play because one cannot explore their own feelings toward a concept never presented to be addressed.

To play a bit of devil's advocate: What if only some of the players are part of drawing up any given part of the maze? Can it still be a sandbox?

For a real world example: Unless I'm mistaken, Rob Kuntz continued to play in Gygax's Greyhawk campaign even after he became a co-DM and began developing his own portions of the campaign. I'd argue there's nothing in that arrangement that would be inimical to the sandbox.

I think this is a good thing and just represents that when the PCs are busy with a BIG priority quest that they may or may not have generated for themselves, that other matters tend to fall by the wayside for awhile.

I dunno if I'd qualify that as "ceasing to be a sandbox", though. I don't think a sandbox stops being a sandbox just because the PCs have decided to spend 10 sessions pursuing a single goal.

If there's some series of murders for a ritual (and a few more murders remain) or the bad guy has a time table of other tasks and places to be, there's some linear elements. Does that in turn makes writing the adventure less "sandboxy" and more AdventurePathy?

There's nothing about a "series of murders" or a "time table" that suggests linear adventure design to me. It certainly can be, but it doesn't have to be.

"Series of a ritual murders" is a toy in the sandbox that the PCs can either choose to pick up and play with or ignore completely. In this it doesn't differ from "dungeon infested with monsters". The PCs can go into the dungeon and explore; or they can go "into" the series of ritual murders and explore. Or they can ignore them both and head into the next hex.

I run a sandbox campaign. Early on, the players got involved in a murder mystery simply by arriving at the location where a murder was about to be committed.

The plarty was approaching the local baron about the possibility of being allowed to bear restricted arms and armour. He was killed their second night there as part of a pre-established timeline.

This is something that I think tends to be under-appreciated by people who haven't actually run sandbox campaigns: It takes a surprisingly light seeding of material before you'll discover the PCs getting involved in things you never expected them to get involved with.

For example, in my last campaign I had prepped several "backdrops". These are basically timelines of events that run in the background of a campaign -- they're the news headlines and large-scale events that the actions of the PCs are unlikely to interact with (in much the same way that you and I are unlikely to interact with the death of Michael Jackson or parents convincing the world that their kid is stuck in a hot air balloon). One of these involved a political leader declaring rebellion against his Overlord.

The PCs managed to trigger the leader's declaration of rebellion months ahead of time because they inadvertently handed him political ammunition that he could use to make it happen. And then they managed to accidentally stumble onto the site where this guy was going to betray and murder a bunch of the Overlord's sympathizers, leaving me scrambling to provide stats and specifics for an event that I never thought would be played out "onstage" (so to speak).

The fallout from both of those events completely shifted the shape of local politics in the campaign and left the PCs deeply entangled in a series of events that I had originally planned to be nothing more than background scenery of the "larger world".

And this kind of stuff happens all the time.

A timeline woven around the PCs is a bit more complex and might have some non-sandboxiness to it.

I'd recommend checking out Don't Prep Plots, particularly this example of using scenario timelines.

It all pretty much boils down to how you're using the timelines. If the timelines become straitjackets, then you're no longer in a sandbox. If the timelines are just a default plan of attack that is freely discarded or modified when the PCs interact with it, then you're still in the sandbox.

Part of the art of running a sandbox is learning what constitutes useful prep and what constitutes prep that isn't going to survive contact with the PCs.

In general, I find it useful to have a default that assumes the PCs don't interfere because it gives me a baseline to improvise off of. (And it also tells me what happens if the PCs don't get involved in this particular scenario.) Almost all contingency planning beyond that is useless, IME. It doesn't hurt to jot down a few notes on cool ideas you might have flitting about, but any time spent on trying to plan for "what happens if the PCs do X?" is generally a waste of time because the PCs are just as likely to do any of the other 25 letters in the alphabet.

Used properly, a timeline can greatly enhance a sandbox: Without a baseline timeline, it's very easy for elements of the game world to default to a static "nothing happens unless the PCs are looking at it" state.

OTOH, you're not a computer and you can't realistically keep an entire world running in the background. You have to pick your battles. As a general rule of thumb (which I'll violate whenever it seems appropriate), I try to design things in static holding patterns ("the orcish slavers continue operations as usual") until the PCs "touch" them and then I'll start tracking that element of the campaign world in more detail until the PCs "finish" it ("if they return within 2 days, the orcish slavers have sold d% of their stock; in 5 days, the orcish slavers have sold all their stock and started packing up their camp; in 10 days, the camp has been abandoned").

When NPCs are acting on a timeline, I call that a trajectory. I know where they will land if their action is not interfered with.

The use of the word "trajectory" is quite brilliant, I think.

 

[Janx]

Hm. Take a look back at the OP for a moment. "What elements must be present in a game for it to be a sandbox?" Looking for definition by inclusion.

Much of the following discussion has been definition by exclusion - looking for things that make something "not a sandbox".

If it looks like a duck, walks like a duck, and quacks like a duck, it is a duck. If the thing also has lasers, that doesn't make it not a duck. It is just a duck with frikkin' lasers.

One can take the question one step further - What can a GM do to foster and use player-direction?

good point. As I proposed some of those exlusions, what my main point was that some attributes of a sandbox were phrased to imply that ONLY a sandbox had those attributes, and thus making that attribute the definition of a sandbox.

So saying a sandbox has attribute X is OK by me.

Saying X defines a sandbox, is a fallacy, if it also applies to other styles.

I'm also easily convinced by arguments that "I have a sandbox and I do X" when I've postulated that having X isn't a sandbox. Personally, I'm working through the traits of a game, and seeing what is more like a sandbox than not.

For instance, timelines seem to me to lead away from a sandbox, but as several have noted, they have used them in a sandbox. Sounds plausible enough for me then that a sandbox could include timelines.

I think what a sandbox DM needs to be is more detached from the campaign world elements. As pawsplay mentions, the timeline is more of a trajectory. In this way, the DM is mentally prepared to react to the PCs then run through a sequence of events.