I have believed in the value of so-called sandbox DMing for a long time. In September, 2008 I tried to codify my views on the subject in a blog post on the ENWorld gaming site. It went over pretty well. But now seven and a half years later, i wanted to revisit it. Sandbox gaming is more popular than ever — I attribute it to the ease with which one can run such a game using 5th Edition D&D — and my perspective has probably evolved.
What follows is the original post (with some minor editing for spelling, etc…) with my annotations in bold italic text. If you are interested in reading the original version, follow this link to ENWorld.
Seven Sandbox Essentials
The “sandbox” setting, in which players are not only allowed but encouraged to make their own fun by exploring and interacting with the setting, is, in my opinion, a fundamental requirement for satisfying, rewarding D&D play. Adventures and setting that force players along certain paths — or worse, away from certain paths — cannot hold a candle to sandbox play. No DM’s or designer’s story has a hope against the story that the players themselves create through their actions and the consequences of those actions (with a healthy dose of DM input and dice-based uncertainty,to boot).
Well,that was kind of absolute — and we know who speaks in absolutes. It is interesting to see how one’s perspective on internet communications changes over time. I now cringe at the preceding paragraph, all to aware of the intended authoritative voice despite not possessing much in the way of authority besides some years running games and a few RPG sourcebooks.
Here are seven essential elements for a good “sandbox” setting.
- Big, but not Too Big: A sandbox setting should cover a relatively large geographic area, with room for varied terrain and environments, as well as multiple political entities. However, one of the things that makes a sandbox game so enjoyable is the players’ ability to engage it as a whole, to see all four corners and to uncover its nuances and secrets. Too big a sandbox makes each part indistinct — the DM likely doesn’t have time to flesh out every aspect of a whole world; nor do players likely have time, or even interest, to visit it all. By limiting the scope of the setting and containing it geographically, the DM has the opportunity to delve into the whole setting and so do the players. Something on the order of the British Isles works well, size and scope wise.
Not surprisingly, I was running a D&D campaign set in a mythic post Roman Britain at the time. That said, I actually agree with myself here. Both of my very successful convention based ongoing hexcrawl sandboxes have hewed to this rule. The Valley of Tombs and the Isle of Dread are both contained geographical regions with varied terrain types and locations.
- Lots to Do, Lots to See: As stated above, a sandbox needs to be diverse in regards to where the players can go and what they can do when they get there. It isn’t enough to litter the place with 100 dungeons. Rather, there should be a handful of dungeons, a handful of towns, a handful of active fortresses, a handful of mystic locations, etc… This applies to political, religious and mercantile groups, as well. A monolithic nation that covers the whole setting won’t do. At the very least, there should by various states or provinces with different cultures and conflicting interests. Even better, numerous small nations or city states work well. In addition, even if the DM chooses to have a dominant religious entity, schisms and sects within the church, with their own temples and own interpretations of scripture are necessary. Players should want to explore the setting to see what is around the next bend or over the next hill.
And here I think I overstated the need for such a high degree of diversity. In retrospect this advice runs counter to the previous advice. Geographical containment or not, having multiple states, multiple institutions and multiple other kinds of factions is a heavy workload that can lead a DM to never actually finishing preparing his sandbox. In my experience with the Valley of Tombs and the Isle of Dread as sandbox settings, a couple settlements and a couple factions of important are plenty.
- A Life of its Own: The setting should “live” like a real place. The diverse locations and groups discussed above should interact and those interactions should be both internally consistent and produce movement within the setting. Groups that are opposed might be moving ever toward open conflict. The heir to a city state might try and hasten along his inheritance. A lowly peddler might slowly climb to the top of the mercantile heap. A dark cult might be waiting for a soon approaching celestial event to unleash their dark master on the setting. It isn’t that the setting ignores the players and their characters’ actions. Rather, there’s something of a timeline or assumed evolution to the setting that the players can interact with and disrupt. Knowing what would happen in a given situation without the involvement of the players allows the DM to better interpret what happens when they do get involved, as well as allowing the DM a plan if the PCs don’t bother with a particular setting element or subplot.
This advice still rings true. What’s more, I think it applies to all kinds of campaigns, not just sandboxes. In my experience, players respond to a living, breathing world and having some idea of what everyone might be doing if the PCs weren’t throwing wrenches into the works helps the DM maintain that kind of setting.
- No Scaling: The sandbox should not scale to the level of the PCs. If there’s a “12th level” monster in the Darkenwood when the game begins, that monster remains there — barring its involvement in the above — whether the players choose to go to the darkenwood at 2nd level or 20th (or both). A scaling setting breaks verisimilitude and suggests to the players that whatever growth they have is irrelevant. instead, the setting should include a wide array of “levels” of adventure locales, NPCs and monsters, distributed throughout the setting in a plausible and internally consistent manner. This is not to say that the setting can’t include “zones’ that are geared toward certain levels of play, but too much of this inhibits the open nature of the sandbox. Mix it up instead. But make sure that there aren’t too many “invisible” major threats. Powerful creatures and characters produce legends and rumors and even inaccurate information will give players a glimpse into what lies ahead, allowing them the opportunity to make a meaningful choice as to where to go and what to do.
This advice is built around the issues of Challenge Ratings and the steep power curve of D&D 3.x and Pathfinder. Since 5th Edition’s power curve by level is more shallow, reminiscent of 1st and 2nd edition where lower level characters could get lucky and take on higher level threats while higher level characters might get slaughtered by even common enemies likes orcs, I view it a little differently now. I usually apply an assumed CR to a geographical locale or “zone” and then include both “easy” and “deadly” as well as typical encounters for that locale on the chart. In this way, the uncertainty of random encounters is preserved at the same time as some internal consistency is maintained for the locale.
- Wandering Monster Tables and Random Encounters: One of the key conceits of the sandbox is that players are free to go where they will, do what they wish and engage the setting through their characters without being pulled or prodded into the DM’s “story”. This requires a lot of work on the part of the DM, creating many adventure sites, placed encounters, NPCs and organizations before play even begins. But even with all the work done, there’s still a good chance the players will go somewhere the DM hasn’t thought too much about or had a chance to flesh out. This is where the value of random encounter charts comes in. With such tables, built specifically for the setting and informed by the detail the DM has done, can provide fun for everyone even during those sessions where the players simply strike out down the road. These random encounters shouldn’t simply be a collection of monsters listed by terrain type. Instead, the charts should include little glimpses into the setting. What tribe are those orcs from? Who are the bandits? Where is the merchant caravan heading? By making specific encounter charts for the setting, the DM ensures that more of the work he has put into the setting sees use. By exposure, players are given hooks to choose for themselves to investigate and engage aspects of the setting.
Still solid advice. In fact, as time has gone by I have become more and more comfortable with “on the fly” DMing and relied more and more on random encounter tables for inspiration. The part about not everything being a fight is true, too. When the encounter comes up, a little creative application can turn what was likely a short fight with a couple specters into a tragic love story the PCs got to not only witness but resolve (as happened on the Isle of Dread at CarnageCon last November).
- New Blood: Characters die. They retire. Players get bored or want a break from the usual. Inevitably, a player is going to need a new character, or a new player will join the group. It may seem obvious, but it is often overlooked: the setting must allow for the introduction of new heroes (i.e. PCs) without breaking plausibility. As such settings that are mostly wilderness or wastelands with few settlements don’t work as well as those that provide a diverse selection of races and classes within the population. As a related aspect, this means the setting must be adventurer friendly and reasonably wide ranging in regards to which races and classes are available. While it is okay for the DM to establish some limitations to better suit the genre or setting he has in mind, too many restrictions hampers the introduction of new characters and should be avoided.
Both the Valley of Tombs and the Isle of Dread include gateway towns for this very reason. they are frontier locations where adventurer types come to test their mettle. The only requirement for a new PC is that they want to head off into the wilderness for fortune and glory. With the more shallow power curve as explained above, new PCs can be introduced at whatever the “starting level” of the sandbox is (be it 1st or 10th) and likely both survive and be able to contribute as they quickly close on the veteran PCs, level wise.
- Meaningful Choices and Meaningful Consequences: Most important of all, the actions of the players should have direct, noticeable impact on the setting, at least insofar as the PCs degree of influence. Information should be plentiful enough to allow the players to choose which actions they will take, and those actions must have consequences. Who the players ally with, and with whom they make enemies; what meta-setting secrets they uncover, and which they bury; those monsters they kill and those that they merely enrage: these all should change the setting to some degree or another. If the players feel their adventures and explorations within the sandbox have a real impact, they will be both more inclined to engage the setting, and more thoughtful of the consequences of their actions.
I think this is another bit of observation that rings true regardless of the sort of campaign you run.
Revisiting an older piece is an interesting experience. I may do it again sometime.