"Railroading" is just a pejorative term for...

[The Shaman]

This always trips me up in these discussions. It seems to me that most people associate sandbox play with pre-determined (or randomly determined) places, NPCs, etc. The "ultimate sandbox DM" would then have the entire world mapped out before hand in every detail, large and small. In reality, however, no one can do this. At some point a DM has to make up details that were not pre-determined or there isn't a random table for. And many of these times, there are a multitude of choices that would not, in fact, be at odds with ingame causal logic. It's just a choice.

Forgive me for quoting myself, but I addressed this in another thread back in April.

In the games I run, the adventurers can't "waltz out of the sandbox," because it's all sandbox. The adventurers can't go 'off the reservation' because it's reservation in every direction.

I think of my prepration time as 'prepping to improvise.' I can't detail an entire game-world, or game-universe for some games, so I'll detail a few obvious locations then focus my preparation on what I need to know to differentiate the cultural and natural landscape the adventurers may discover in their travels. From this I can draw things like npc characterisations on the fly, and from there I'm simply reacting to whatever the adventurers do.

In my experience, successful improvisation comes from knowing the setting well, not in terms of where this city or that river is located, but how the inhabitants of this area differ from the inhabitants of another area, in their outlooks, lifestyles, and subsistence, then bringing that out in response to the actions of the adventurers.

To use the example of MerricB's Alliance base, how does an Alliance base on a core world differ from one on the frontier or the edge of the Black? This gives me some guidance on how base personnel perceive themselves and their jobs, what resources they can bring to bear, and so on, which makes reacting to the adventurers much easier.

The same is true in the game I'm prepping to run: how does the outlook of a noble with a small estate in Languedoc differ from one in Aunis? I don't need to know every valley of the Cévennes or beach of the île de Ré to create a (hopefully interesting and distinctive) characterisation of each.

I also prep random encounters in advance of actual play. For me random encounters are the 'living setting' - I spend time thinking about the origins of the encounter, identifying the motivations and methods of the antagonists, and so on.

For example, a randomly generated 'bandit' encounter becomes rebellious Huguenots in the Midi foraging for supplies for the duc de Rohan, or ragged, half-starved mercenaries returning from the Holy Roman Empire and resorting to brigandage in Picardy, or chauffeurs roaming the pays of Normandy looking for victims to capture and ransom. In this way there are no 'generic' random encounters; each is a reflection of the game-world where the adventurers are standing at the moment.

As part of preparing to run a game, I'm building the tools I'll need behind the screen once we're actually playing.

Frex, I'm fortunate enough to have an exceptionally detailed map of the setting for my game, but even with this as a resource, there is no way I can reasonably attempt to key every settlement, every church, every abbey, every fortress. So far I've focused on detailing certain cities which I believe are most likely to come up in play: Paris (the central city of the setting), Rouen (the gateway to New France), La Rochelle (the Protestant stronghold and another important port), Toulouse (an important cultural and administrative location, the 'Paris' of southern France), and Marseille (the gateway to the Mediterranean). The importance of thse cities is such that I reasonably expect the adventurers to visit at least two or three of them in the course of playing the game.

But suppose the adventurers visit Bordeaux, or Clermont, or Lyon instead? Part of my prep focuses on being able to improvise so that a visit to each is distinctive in some way. Bourdeaux is home to the only significant population of Jews in France outside of Paris or Avignon and trades extensively with Portugal. Clermont is situated on the edge of one of the most geographically inhospitable regions of France, giving it the feel of a frontier town in the heart of the kingdom. Lyon is a historic banking center with strong cultural and business ties to the Swiss and the Italians. This allows me to take a generic random encounter and reskin it so that it's geographically appropriate to the setting and reinforces the cape-and-sword genre feel of the game.

And that's, for me, one of the approaches which characterizes running a 'sandbox'-y setting. I'm not improvising encounters to 'steer' the adventurers on the 'right track.' I'm not improvising encounters, or features of the game-world, in response to something on a player's character sheet. I'm not improvising 'level-appropriate' encounters.

Instead, I'm improvising encounters largely without regard to who adventurers are. If the adventurers head off cross-country in Provence, they may encounter a Roman ruin, and one of the more likely forms of wildlife to be found hiding among the fallen columns and crumbling foundations is a viper. That's about as close as I get to 'aiming' an encounter at an adventurer.

The exception to this is encounters which occur as a consequence of the adventurers' actions. If the adventurers rescue Princess Pinkflower from the château de Bauchery, a random encounter with troops or guards gets assigned a chance of being the baron's henchmen trying to steal her back.

 

[Umbran]

But nor do the rules of the game model that logic. They are to a significant extent external to it.

I find this is one of the language ambiguities that causes problems.

In the sense that every detail of the game rules directly resembles an element of the game world logic, you are correct, the game rules don't need to model the logic.

In the sense that the results of the game rules are the same as the results of the game world logic, then the rules darned well ought to model the logic! If the results of using the rules are not the same as the results of the logic, then the characters will have experiences that aren't really supposed to happen in the game world.

Say you're running a four-color supers game. In four-color supers, characters don't die frequently. If your game rules have one character death per weekly session, the game rules are not modeling the game world, and you've got a problem.

 

[Hussar]

If your game is openly billed as a scenario -- "The Musgrave Ritual" or "Battle of the Five Armies" or what have you -- then anyone who pejoratively calls it on that account "a railroad" simply has no business signing up for it and then complaining about its being just what was advertised.

If you claim, "You are free to move about not only the cabin but all of Madagascar," but it turns out there's really nowhere to go except down the trap door under the rug in the sitting room ... then people may reasonably resent being misled.

That's a rather odd definition of scenario though. I mean, you're basically saying that something like Shackled City - a series of twelve or thirteen distinct but related adventures - is a single scenario and not a campaign.

A scenario that takes you from 1st to Epic level in 3rd edition? Most people would call that a campaign.

Likewise, GDQ, nine very large modules which can easily take a year of play time is a fair bit more than just a scenario. It's not that hard to call it a campaign.

That's like saying T1-4 is just a scenario. If a scenario can cover an entire play life of a group of characters, then what is a campaign?

 

[Ariosto]

what is a campaign?

It's whatever you say it is, Hussar.

And that thing we used call a D&D campaign is whatever y'all decide we're allowed to call it, whenever you make up your minds.

 

[Hussar]

Well, Ariosto, I don't know about you, but, I generally say that an adventure and a scenario are pretty much synonymous. To me, a campaign is a series of adventures linked in some fashion, whether thematically, or simply by the fact that the same players played through all or most of them.

Well, I tried, but I see that playing pin the tail on the definition is still the game du jour so, I bid you adieu.

 

[pemerton]

In the sense that every detail of the game rules directly resembles an element of the game world logic, you are correct, the game rules don't need to model the logic.

In the sense that the results of the game rules are the same as the results of the game world logic, then the rules darned well ought to model the logic! If the results of using the rules are not the same as the results of the logic, then the characters will have experiences that aren't really supposed to happen in the game world.

Say you're running a four-color supers game. In four-color supers, characters don't die frequently. If your game rules have one character death per weekly session, the game rules are not modeling the game world, and you've got a problem.

Umbran, I had in mind a slighlty different distinction - though I don't dispute the importance of the point that you make.

The difference that I had in mind is that between games like (on the one hand) Rolemaster, Runequest and Classic Traveller, and (on the other hand) HeroQuest or (as I interpret it) 4e. In the former games, the rules very closely model the ingame causal logic of the gameworld (in Forge terminology, they are purist-for-system simulationist). In the latter games, the rules are closer to "meta-rules", distributing permissions to narrate ingame events, and establishing the parameters that constrain such permissions.

Any number of examples could be given. Just to pick one: both RM and RQ have fumble rules. If a fumble result is rolled on the dice, then not only does the action that the player has initiated fail, but we know that, in the gameworld, the PC has made a hash of the attempt. In 4e, on the other hand, a "1" is an automatic miss on an attack, but doesn't dictate anything about what has happened in the gameworld. It might be that the PC attempted an attack that went wild, or was fumbled, but equally it could be that the PC made a brilliant attack, thwarted only by the equal deftness of the opponent. The only constraint on the gameworld imposed by the rules is that whatever is said to happen, it must be consistent with the PC failing to reduce the opponent's ability to fight (because no hp have been delivered, given the miss - and putting to one side powers that do damage even on a miss).

To tie this back to your point - the world of RM and RQ is a world in which even the greatest heroes occasionally fumble their attacks. The world of 4e, on the other hand, and depending upon the story preferences of the players/GM (in different groups, a different person might actually enjoy the relevant narrative authority) can be such a world, but equally can be a world in which (at least some) heroes never fumble, and fail on occasion only because they face opponents almost as powerful and skillful as they are.

 

[Hussar]

Forgive me for quoting myself, but I addressed this in another thread back in April.As part of preparing to run a game, I'm building the tools I'll need behind the screen once we're actually playing.

Frex, I'm fortunate enough to have an exceptionally detailed map of the setting for my game, but even with this as a resource, there is no way I can reasonably attempt to key every settlement, every church, every abbey, every fortress. So far I've focused on detailing certain cities which I believe are most likely to come up in play: Paris (the central city of the setting), Rouen (the gateway to New France), La Rochelle (the Protestant stronghold and another important port), Toulouse (an important cultural and administrative location, the 'Paris' of southern France), and Marseille (the gateway to the Mediterranean). The importance of thse cities is such that I reasonably expect the adventurers to visit at least two or three of them in the course of playing the game.

But suppose the adventurers visit Bordeaux, or Clermont, or Lyon instead? Part of my prep focuses on being able to improvise so that a visit to each is distinctive in some way. Bourdeaux is home to the only significant population of Jews in France outside of Paris or Avignon and trades extensively with Portugal. Clermont is situated on the edge of one of the most geographically inhospitable regions of France, giving it the feel of a frontier town in the heart of the kingdom. Lyon is a historic banking center with strong cultural and business ties to the Swiss and the Italians. This allows me to take a generic random encounter and reskin it so that it's geographically appropriate to the setting and reinforces the cape-and-sword genre feel of the game.

And that's, for me, one of the approaches which characterizes running a 'sandbox'-y setting. I'm not improvising encounters to 'steer' the adventurers on the 'right track.' I'm not improvising encounters, or features of the game-world, in response to something on a player's character sheet. I'm not improvising 'level-appropriate' encounters.

Instead, I'm improvising encounters largely without regard to who adventurers are. If the adventurers head off cross-country in Provence, they may encounter a Roman ruin, and one of the more likely forms of wildlife to be found hiding among the fallen columns and crumbling foundations is a viper. That's about as close as I get to 'aiming' an encounter at an adventurer.

The exception to this is encounters which occur as a consequence of the adventurers' actions. If the adventurers rescue Princess Pinkflower from the château de Bauchery, a random encounter with troops or guards gets assigned a chance of being the baron's henchmen trying to steal her back.

I've been thinking about this and it finally struck me why I was having a problem.

How is this not describing Celebrim's Rowboat Campaign?

I mean, you're saying that the players can go anywhere and do anything, but, they are not guaranteed to find adventure everywhere they go. Instead, they proactively decide to head cross country and come across a Roman Ruin. And their reward for being proactive is to meet ... a snake.

How is this not telling the players, "Sure you can go anywhere you want to go, but, if you head to places I didn't really anticipate, you are now in a "no-adventure zone" and be prepared for boredom?"

I'd much prefer that adventure seems to lurk under every rock and for some bizarre reason, trouble just seems to find the group, regardless of what we do.

 

[Celebrim]

I've been thinking about this and it finally struck me why I was having a problem.

How is this not describing Celebrim's Rowboat Campaign?

I can't tell if it is or isn't, but as best as I can tell if it does avoid it, it's because he's got only two players and he's running a successful Small Drama campaign heavy on theatrics and concern for the daily hardships and interpersonal conflict. Basically, it's a soap opera set in 17th century France, possibly run with only one player at a time allowing the DM to lavish time on the players. The fewer players you have, the more you can personalize the game and the more you can make the small stuff like making camp, shopping, small talk with NPCs and the like interesting. Plus, he's a history buff and if he's got writerly skills its probably like getting immersed in a good historical novel.

It very much makes a difference whether you are running a game for one person, or two, or six, or forty, just like it very much makes a difference if you are running a game for people you've never played with before or you've been DMing for 18 years with the same group of people.

If he's successfully running a Small Drama game for more than two players, then its because he has a group of amateur thespians that spend most of their time talking with each other IC. The group matters.

Of course, I can't really know what's going on in his game without sitting in on it. These are all generous guesses as to what makes a game like that work. I can't say however that the game he runs appeals to me as a player. I find the whole thing terribly boring as described, and I always try to run the game that I as a player would enjoy. But, also, it takes all kinds. There is no gaurantee that everyone would enjoy my game (though I'm smug enough to assert that so far, everyone has).

 

[Nagol]

I've been thinking about this and it finally struck me why I was having a problem.

How is this not describing Celebrim's Rowboat Campaign?

I mean, you're saying that the players can go anywhere and do anything, but, they are not guaranteed to find adventure everywhere they go. Instead, they proactively decide to head cross country and come across a Roman Ruin. And their reward for being proactive is to meet ... a snake.

How is this not telling the players, "Sure you can go anywhere you want to go, but, if you head to places I didn't really anticipate, you are now in a "no-adventure zone" and be prepared for boredom?"

I'd much prefer that adventure seems to lurk under every rock and for some bizarre reason, trouble just seems to find the group, regardless of what we do.

The rowboat is a degenerate form of a sandbox that allows the players infinite choice but strips that choice of meaning by requiring those choices to be made without context of a situation or in response to the environment.

A sandbox allows the player to make meaningful choices. Meaning implies decisions made with knowledge/context of the environment and the environment reacting in plausible and understandable ways.

Deciding to strike out into the wilds chasing nothing in particular is a choice. It is the players gambling that the choice will lead to something unexpected or that the expected ressult is preferable to their other current choices (for example, they are being hunted in town and don't want to confront the hunters). Sometimes such gambles pay off othertimes they don't. Always having the gamble pay off strips the choice of a lot of meaning for me.

Part of the draw for a sandbox for me is those range of choices include choices that are better for some strategies/outcomes and some are worse. Players should weigh the expected result of their choices against their goals and preferred outcomes.

 

[Janx]

Well, Ariosto, I don't know about you, but, I generally say that an adventure and a scenario are pretty much synonymous. To me, a campaign is a series of adventures linked in some fashion, whether thematically, or simply by the fact that the same players played through all or most of them.

Well, I tried, but I see that playing pin the tail on the definition is still the game du jour so, I bid you adieu.

And thats why he's the only person in my Ignore list. It does not appear possible to have a discussion of gaming elements with him.

Whereas, Shaman's tables of coincidentallly linked random encounters seems like a fascinating way to generate random game content (and thus the GM has little clue what's coming next). Shaman could sell me on that idea within the framework of how I do things now, and that's a useful dialog.

I'm not sold on the idea that a sandbox has to explicitly be non-level appropriate. Just as GMs in non-sandboxes don't literally make every encounter and every entity level appropriate. A GM in the adventure path style writes content that is appropriate to levels needed to the adventure, and stuff that's not involved (and thus not assumed to be directly attacked) could be wildly varying in level. Contrast that to the Elder Scrolls IV: Oblivion sandbox. Wherein everything levels up with the PC. So much so, its a valid strategy to play the game and never sleep (so you never level up), thus making the end encounters easier.

Basically, nobody has a campaign world where every encounter, every NPC, every place is literally level appropriate. Its a style choice of whether the GM gives a hoot if the PCs go into too dangerous parts or parts he hasnt planned on.

Likewise, there are no rules to break when determining the reactions of NPCs, the level of the encounters, or any other event or entity in the game. Rocks fall, everybody dies.

Any GM who deliberately counters players choices to force them through his content when they deliberately chose to avoid it is a crap DM. When the players say, "we ally with the giants", or "we go south to avoid the giants" then the GM has to react to enable that, not actively try to make it fail so he can force them back into fighting with the giants like his adventure says they should. That is such obviously bad GM behavior, they invented a term for it, "Railroading" that it shouldn't bear discussing or debating with any experienced GM such as would appear in this thread.

The questions to be pondered should not be the obvious stuff, as if that proves anything else is also bad. The question should be what ideas from other gaming styles could be adapted to your own to avoid unfun situations and to enhance your game, not replace your game style with somebody else's.