8 lessons learned from my experience running a successful sandbox campaign

Intro

An era has just ended. For two full years - since roughly September 2022 - I have been running a regular open table sandbox campaign in Swords and Wizardry. We aimed at meeting weekly, and probably averaged a game about every 2-and-a-bit weeks across the two years, taking into account small periods away and substantial periods of regular attendance.

This was, without qualification, the best campaign I have ever personally been involved with. Testimonials from players confirm this. The core group was made up of a mix of 5e veterans and completely new players, all of whom had an amazing time right up until the end of the last session - 7.5 hours of continuous play - and the final folding up of the GM's screen.

All things must come to an end (credit: Nicolas Delort)

The campaign represents a honing of my own techniques for running games, and the culmination of a process of refinement, of setting up a campaign to reflect what I personally want out of a game, and what I view as the best, most authentic form of OSR play. Of course, no two people have exactly the same view on this sort of thing, but given the success of the campaign, in terms of the enjoyment the players got out of it, I wanted to share some insights I felt I had from the experience - things I did that helped me out, that helped the players have fun, and that saved my ass from time to time. I think the reason I haven't posted to the blog in a while was that I didn't feel the need to write about my rpg theories, because I was getting to put them into practice and test them so frequently. Thus the advice here is tried and tested through experience, and is thus proven to work.

Why eight lessons? Because that was how any I could think of, without subdividing them into tiny pointless details. I've tried to structure these loosely in order of importance, but don't take that too seriously at all. Truth be told, each one of these could probably be a post in its own right, and I may return to them and write them up in more detail as and when. For now, though, I hope they prompt useful thoughts for other GMs.

1: Open your table

The Alexandrian article I linked above gives a good account of this. Basically, on the GM's end, what you want to do is set a player threshold and a schedule. For instance, I made the rule that the game would run every Wednesday afternoon, provided we could get together any three or more players. This has several good effects. First, it makes the game very low-commitment, so you can justifiably say to prospective players that they can give it a try, no expectations, and come along if and when they're free, if they want to. This means you can get a campaign mailing list (or messaging group) with dozens of people, any of whom can show up if and when they want.

The second advantage, and the more important, is that it makes your campaign much more robust. A lot (the majority?) of modern-style campaigns are afflicted with 'scheduling conflicts' that can make it almost impossible to get a game in, and can break down the campaign's momentum. If you run every Wednesday, come hell or high water (provided you can scrape together enough warm bodies), everyone - including the flaky player who can only make it once every four weeks - gets more gaming. The Tentpole dungeon helps with this, because it makes things smoother if you can think of each session as an  expedition, to explain changes in the cast makeup.

An interesting thing that will probably happen is that, after casting the net wide and recruiting lots of people to your open table group, you get a core of committed players who turn up every session, and a group of others who turn up once in a blue moon, or never. This is all fine, and to be expected; the open table format lowers the barrier to entry, so you can recruit new players easily, but the game won't stick for everyone. The fact that you're casting the net wide and letting everyone try the game before deciding what level of engagement they want means you'll find those committed players more easily (and wring more gaming out of them) than if you did the traditional thing of tapping four or five friends you think are reliable, who may or may not gel with the game as you run it.

2: Keep your secrets

I actually have written a post on this one (actually, I've written two or three), but I'll say it again here: When players feel like the game-world contains no facts other than the ones they get shown, they treat the world like a dream-world or an illusion, rather than treating it like a real place. When the players know that the world contains secrets that they may never discover, and that you are willing to allow to remain secret forever if the players don't stumble upon them, then they'll start to treat it like a real world with its own facts. This sometimes means pulling your punches, giving the party clues that something is there without revealing anything substantive about it, and allowing things to sit. Sometimes you'll have to sit on an amazing adventure or location you created and watch the players walk right around it. That's fine; if they don't get round to it, you can always use it for the next campaign. Avoid the temptation to tell them what they're missing out on, or to have adventures track them down (although it's fine for events to catch up with them if it makes sense in the situation), and definitely avoid the temptation to give them a peek behind the screen. There are still things my players don't know anything substantive about even now the campaign is over, and I intend to keep it that way; I want them to look back on it as a real place they were exploring, not as a convincingly woven set of stories.

This is my response to the Blorb principles. It's tricky to convince players there's more to the world than what you're showing them, when all you can do as a GM is show them things. If it sounds like a paradox... well, it is. Read the article to learn how to resolve it.

3: Give them a (bad) map

This is one that I've been doing for a while, but this campaign validated its utility for me beyond doubt. Again, this is one that I could write a whole post on (I'm beginning to think that the next year of the blog is just going to be expanding on items in this list). I see a lot of people cropping up on my usual loitering space of D&D Reddit asking questions that would easily be answered by this technique.

When you design an overworld or wilderness, make yourself a hexmap. Make it nice and accurate - I tend to do contours for elevation, especially in mountainous regions, rather than drawing little mountain pictures, just because you can get more detail. Draw on lots of things you don't expect the players to find yet, and some things you haven't yet planned out in detail. This represents the reality of your world - if it's on there it exists. When you're finished, it should look something like this:

My behind the scenes map, for a different campaign as it happens.

Do not give the players this map. Instead, go and get a plain piece of paper, preferably quite thick, and draw a picturesque, Tolkien-y fantasy map on that - the type with pictures of mountains and little broccoli forests. If you like (and I do recommend this as a mini craft project), rough up the edges a bit (I used a rock), put some folds in it, and stain it with some tea to give it an aged look. It should end up looking something like this:

Same region, different map. You may not be able to tell, but this one is yellowed and aged,

This is the map you give to the players. The point is to give the players a handle on the world, so that they can start planning their way around it, but not to give them perfect information - the player map is missing a lot of detail (most of it, actually), and isn't quite to scale, but is good enough if you're navigating by landmarks. This is to do with immersion, but it's not actually to do with giving them a touchy-feely prop - this isn't fog machine immersion. Rather, what you're doing here is bringing the play options afforded to the players into alignment with the in-world options afforded to the characters.

In simpler terms, you want the players to be making the same choices their characters are confronted with. You do this, first, by ensuring they have the same information. When players have the same information as their characters, and leverage it to achieve the same goals, they find themselves faced with the same choices. And it's making the same choices as their characters that makes players immersed, because in those moments, essentially, they don't notice that they and their characters are different people.

Some bonus points about the player map: First, I like to put different writing in different colours on my map. This creates the impression that it has passed through different hands, who may have different concerns, creating something like 'layers' to the map, and highlighting points of interest the players may also choose to investigate. This also encourages the players themselves to write on the map - I always tell my players to do so, although many of them are hesitant. But it really helps them to feel like they're exploring the space, and discovering new things, when they get to note down the new place they blundered into so they can try and get back there (just try to resist telling them exactly where they are - instead, get them to make a mark where they think they are, based on what they can see etc.). Indeed, having an imperfect map stops them thinking of the map as the space, and gets them to think of it as a map of a space that exists independently.

The same, by the by, goes for dungeons. You don't have to give the players a partial map of the dungeon (I didn't in this game), but if you do, don't give them an infallible one. A huge part of the character of old-school play comes from mapping out dungeons yourself. Again, if the players have a perfect map they'll start to think of it as being the space, or at least a sort of 'playing area'. Really, you want them to treat the dungeon as a real space they're feeling out and trying to grasp, rather than a thing on the tabletop that you reveal to them as they go along. It shouldn't be trivial, for instance, for the characters to leave the dungeon - no saying "We go back the way we came" or pointing to the route on a VTT. If you can, you should always try to have the players describe to you the escape route they take, because some of the most fun, most tense gaming experiences happen when the party thinks they should be coming to an exit, and actually find they don't know where they are.

4.1: Improvising: Know your limits

I could write a whole post on improv do's and don't's, and probably will at some point. People seem to be afraid of improvising, but most of what the GM does at the table is actually improv - no-one scripts every possible thing every NPC could say, for instance. For now, here are two things I've long thought and put into practice, that experiences in this campaign has confirmed to work.

The first thing is to know your own limits and blind spots: Get to know the things you as a GM are and aren't good at coming up with on the fly. (I was going to write "are comfortable with" there, but comfort improvising only comes with practice - being good doesn't imply being comfortable.) As a sandbox GM, you can't prep every detail of everything the players will encounter - you need to find a balance of breadth and depth, which means leaving some blank spots. By learning what sorts of details you can and can't come up with at the table, you learn which details you need to prep, and which you can leave out, or outline in broad, abstract terms.

For example, I know that I'm good at coming up with distinctive and memorable towns, NPCs, gods, nonsense peasant festivals, an so on. However, for whatever reason, I suck at creating quests, rumours and suchlike - NPCs with urgent business for the PCs to help with, half truths for the party to follow up on, and other things of that ilk are my blindspot, when it comes to spur-of-the-moment play. I can have a whole scenario prepped in the next valley, complete with mysteries, encounters, adventure locations etc., but at the table I'll always struggle to come up with a reason the local innkeeper tells you to go there. It's just something about how my brain works. So what I do is prep things like rumours and scenario hooks in some detail, and make sure I have a really good idea of where they are in the world, and when they're likely to come up.

So my notes for a whole district of a city might comprise the words 'Temple complex: Many gods worshipped, mostly lesser deities', and I might determine on the fly that the peasants of this village are celebrating New Milk Day (where they tip cows and chase the New Milk's Fool), but when it comes to explaining why Doug the blacksmith will pay 20gps for you to go into the local dungeon, it takes me a good afternoon to come up with all the details. Your own blindspots might be different - maybe you're really good as mission hooks, but you struggle to come up with NPC names (a very common one), or taverns or whatever. In that case, those are the things you need to prep, while leaning as heavily as possible on your ability to improvise the other stuff.

In short, then, prep you could improv at the table is prep you don't have to do. Everyone has different strengths and weaknesses with improv (if you're thinking "I only have weaknesses", you just haven't found the right setup and the right support framework - see Make your own tools). If you learn what you can create  to a high standard with little or no prep, you can give your players the impression of a seamless world filled with detail that goes all the way down.

4.2: Improvising: Make links

One reason I think a lot of GMs are afraid to improvise is because they think improvising involves pulling something that is coherent and fully formed straight out of thin air. I imagine this is because, when you see people who are good at improvising doing it, this is exactly what it looks like they're doing. But this is actually a misperception - it just seems like that person is creating something impossibly coherent out of whole cloth right in front of you because they're experienced and know the tricks.

A while ago, I had a conversation with a friend who was part of a long-form comedy improv troupe. (Aside: If you want a good time, try playing D&D with improv comedians.) I asked if their show used the David Gilmour style of improvisation - rehearse several cool set-pieces, then use improvisation to connect them up. He told me that their show, which was fully improvised and lasted about an hour, had no rehearsed scenes or gags, but did have a sort of formula. The first bit of the show would be all free play. After that, they would introduce elements related to the prompt. Finally, in about the last third of the show, they would make an effort to come to a resolution, specifically tying in elements from earlier.

This last part is important, because the recurrence of those elements - the fact that they suddenly became tied to the resolution in some way - created an impression of a coherent narrative that seemed, from the outside, to be impossible to achieve without some sort of preparation. There was no rehearsing plot structures, not hurried conversation offstage about planning ahead (there was only ever one person offstage at a given time); the group just practiced folding in the bits and pieces they spun up earlier to link them, in whatever way, to whatever was happening at the end. The result is a bit like a magic trick, where it seems like the story must have been planned out, without there being any time it could have been planned in.

Hopefully the relevance is obvious. When you're improvising as a GM, you're not creating something whole-cloth at the table. Instead, you're filling in gaps and joining dots, using what you know about your world's character and tone to bridge from one established detail to another. For example, when I needed to iprovise a reason for a Balrog to show up in a peacable hamlet (see Use (the right) random encounter tables, below), I folded in things the party had already encountered (a lowly demon they freed from the dungeon, the big Wizard War in the distant past) to situate and explain the new content, creating the impression of a world of interconnected facts. Again, this is much easier than creating things from scratch; you just need to think of a link - any link - between the thing you know must be there (a monster, the occupant of a house, an item, a room) and the details you already have prepared. As an aside, this is one of the ways Deep Lore can be useful to you, i.e.: by providing these links - if the players know about the Wizard War from something else, everything you link to it in improv will have the appearance of being part of a grand tapestry of information that is being gradually revealed.

5: Keep a commonplace book

You never know when inspiration will strike, and you need to have somewhere to pin it down when it does. Follow the advice of Dirk the Dice and keep a notebook or 'commonplace book' with you. Mine is filled with pages for rough ideas, dungeon and wilderness maps, and procedures and tables (see Make your own tools). I tend to write anything that definitely makes it into the campaign on index cards for filing and to make it easier to find, but I often end up using stuff directly out of the notebook - so much so that my players got me a new one for my next campaign as a parting gift. A simple recommendation, but an important one.

6: Set up a tentpole dungeon

I just wrote a post on this. That says basically everything you need to know. Having a tentpole (mega)dungeon gives your players a focus at the start of the campaign, and lets them learn the ropes in a controlled environment. It's also a really good way of showing them up-front that the world has its own facts (see Keep your secrets), and that their destiny very much depends on their own decisions and strategy, rather than on anything you have planned.

A sub-tip, since I feel it doesn't necessarily merit its own heading, but is definitely required to make the tentpole dungeon work in the way you want, is this: Use GP for XP. I was one of the sceptics about this for the longest time, but it's one of those things you have to try (in a sandbox campaign, with multiple options for adventure) to see how it channels player motivation, and encourages and rewards them. I'm with Dwiz on this - think of your players as collecting for their eventual lordship, when they have to pay for castles and retainers and such with their legendary ill-gotten gains.

7: Make your own tools

Related to Improvising: Know your limits, this one is about working out tools that work for you. A lot of this is recognising gaps in your skillset and creating tools - usually tables, sometimes mechanics - to plug them. As an example, when a combat against a foe composed of ordinary infantry is initiated, my mind tends to go a bit blank, and I tend to set the combat in a bit of a featureless void. (Although I love miniatures, I play without them for the vast majority of combats, but I've also seen this problem crop up with GMs who always resort to the battle map.) To plug that gap, I created the 'D20 what's interesting or distinctive about this particular combat' table, and associated sub-tables. 

Creating tables is easy - you'll start off thinking you'll never fill 20 or 100 slots or whatever, and by the end you'll be thinking about expanding the table to fit everything in. For the combat distinctiveness table, I just thought of all the ways a mapped-out combat or a tabletop skirmish wargame could be interesting (elevation differences, elite enemies, mounted enemies, hazards etc.), then moved onto thinking about memorable battles from videogames, movies etc. I put all my tables in the back of my notebook, for ease of flicking. As well as the combat distinctiveness table, have a table of d100 interesting NPC properties, another table for NPC personalities, a mechanic for Chaos magic (arcane magic's unruly, anti-social cousin), a section in my dotted notebook of 80 dungeon geomorphs (with stocking tables), and so on. The last double page spread in my notebook is taken up with 100 fantasy words that reflect the tone of my campaign. When I'm in need of a kick to get my improv brain working, I roll on the table, usually rolling twice and combining the results. 

As a bonus, once you've created a tool, alternate uses in different contexts often become obvious. For example, the combat distinctiveness table can be used to lend flavour to almost any locale with inhabitants, even if they're not actually fighting you - maybe this temple is built on different levels, or houses dangerous flora, or all the acolytes are equipped with a distinctive type of ceremonial weapon. The trick is to remember that you have the tools to hand when they become relevant, and not forget they're there.

8: Use (the right) random encounter tables

This is almost a counterpoint to Make your own tools. Before this campaign, I tended to write my own bespoke random encounter tables. However, this time, I found it extremely fruitful to use the original D&D tables, in all their slightly wonky glory. The reason this worked well was because they pushed me out of my comfort zone.

I usually think of random tables as a machine for filling in the indeterminate bits of the map. The GM can't possibly have each fold of the landscape, each tree and each tuft of grass noted in their mind, so they have to deal with things in the abstract, and use procedures to fill in detail. Hence I always used to design encounter tables to fit the region, including not only random monsters, but also events and locations, to work sort of like the procedural generation you see in games like Minecraft.

What I found in previous campaigns is that this sort of bespoke design tended to have me default back to the same tropes, like a sort of echo chamber. I think every GM has a set of these images that emerge as if from one's subconsciousness in the process of improvising to fill in details (see Improvising: Know your limits). Mine include pale sorcerers in ragged black cloaks, evil leafless trees, orbs of power and big sculpted faces on doors and things. Truth be told, the reason I didn't do a bespoke table full of these sorts of things is that I didn't have time. But after a few sessions of using the ones in the OD&D reference sheets (which I blu-tacked to the back of the 5e DM screen) I noticed that the encounters generated were so much more colourful. They prompted me to answer difficult questions - most notably: What on earth is a Balrog doing within a mile of a peaceable hamlet?

This sort of thing brings real life to the campaign world. Several encounters such as this, where I had to think outside the box to make them make sense (see Improvising: Make links), resulted in campaign-changing elements that weren't pre-planned. Plus, when you're doing this sort of behind-the-scenes work to integrate those elements, and allowing even spontaneous things to be consequential for the campaign, the players don't know what's pre-planned in your notes and what's generated on the fly. This sort of averages out as them taking everything at face value as part of the world, which is exactly what you want.

Tldr: Allow wonky encounter tables (coupled with improv) to challenge you and inject life and variety into the game.

Conclusion

That's about it. I could go on writing for days and days about the campaign - it's been a momentous experience for myself and, I think, for the players who I experienced it with. It lived up to everything I wanted from it, and I'm sure we'll be telling war stories for a good long time. And that's more or less exactly what I want from a sandbox campaign: A "war story generator" that creates experiences that bond people together and live on in shared memories.

I felt like I needed to write this post to draw a line under it, but in truth I don't think I'll ever really be done. With an open table and a persistent world, characters can always put in an appearance when players are in town. With the right attitude, the game is never really over.