Some of you might have picked up on it, but I'm a veteran of the US Army. One thing I noticed about the Army was it's propensity to screw up. The armed forces in general fail all the time. The one thing the Army is better at than failing, though, is learning from failure. Every training exercise and mission are wrapped up with an After-Action Review (AAR). Broadly speaking, these are three things that went well, and three things that could have gone better.

I think DMs can learn a lot from that approach, and I try to do autopsies of each campaign I finish as I finish them. I figured I'd share the lessons I learned from my last campaign here, because I think they're highly applicable to anyone trying to run a sandbox game using 5E or one of its hacks.

For those of you wondering if this is just me touting a really good campaign, no. In fact, last Saturday we all agreed to just stop playing because everyone was pretty tired of it. I'm taking a month off of DMing (something I've never had to do before) to take a break after it. It has not gone well, and that's why I wanted to share what I learned through it.

For those of you wondering if I need a shoulder to cry on, no thanks. I have a sympathetic bottle of bourbon and an even more sympathetic heavy bag I've been working over, so I'm clearly handling my emotions like a champ.

No, I'm posting these up here because I think it'll help DMs avoid the pitfalls I'm discussing. And because screaming into the Internet is how I deal with stuff. Please do not bash my players; constructive criticism is welcomed, but let's keep it civil, y'all.

(Feel free to bash me, though.)

Spoiler: Campaign Basics
Show
This game was run in Esper Genesis, a hack of 5E that uses (per the developer) "the same room with slightly different furniture." Minimal rules differences, designed to deliver a Mass Effect-style experience. I had five players, all of whom have run a 5E PC from 3rd - 20th levels in a game I ran. There was a good spread of short and long rest classes, spread among martials, half-casters, and full-casters.

The conceit was that we were playing as humans (or genetically engineered near-humans) just before First Contact, exploring a newly discovered sector of space called the Tartarus Sector. At some point, I knew the PCs would come into contact with hostile and benign alien civilizations, and how they handled those would reflect back on humanity as a whole, triggering war or diplomacy (or both).

The core gameplay loop was designed as:
  1. Party hears rumor about an interesting thing in a star system
  2. Party goes to star system
  3. Party explores star system
  4. Party resolves threats in the star system, or flees
  5. Party returns to port


XP was milestone. It derived from discoveries, meaning things that your character did not know that required risk and effort to find out. You needed a number of discoveries equal to your proficiency bonus to level up, and my design expected at least one discovery per session.

The campaign lasted for 35 sessions over the course of a year, reaching level 15. We did not do a session zero all together; rather, I met with players in ones and twos in the weeks leading up to the campaign to run preludes for their characters, essentially "dropping them off" right at the first adventure hook.

The sandbox itself was overdesigned. I generated 30 star systems, several with multiple sites of interest / conflict, using Kevin Crawford's excellent Stars Without Number. I baked downtime into the game by requiring solar recharges (a la Battletech's Jump Drives). These meant that you needed 1 - 5 weeks of recharging (depended on the star you were using to recharge) per about 20 parsecs of movement. Building this thing took months of time and effort, and I can't tell you how ready I felt to use it!


Lesson One: Do Not Try To Run Exploration-Centric Play If Your Players Don't Care About Exploration

It is a truth universally acknowledged that for every DM who wants to present a world with causality, consequence, and then run it as an impartial referee, there are like two hundred players who have had a rough week, don't want to think too hard, and just want to play through a cool story.

Yeah, I'm butchering Jane Austen. Sue me.

Now, it should be noted that these players are not wrong. I might find running for them boring as hell, but they aren't playing the game wrong.

I had already run my players through a pretty epic campaign (in scope, anyway) that was so player-led they didn't want to take downtime. At all. So they already knew my position on plots in RPGs (namely, that they are dumb and should be avoided), knew my DMing style (I consider myself an impartial referee, portraying NPCs as loyally as I can to their core principles), and knew how I adjudicated actions. So I figured they would be up for an exploration-based campaign. I literally pitched them a sci-fi game or a western, and they asked for a space western. I mean, when you get handed XP rules that are based around discoveries, you figure that finding sh!t out is the core premise of the game, no?

Well, apparently that was a surprise to some of them. During our last session, I solicited specific feedback on how everyone was feeling, and how they were enjoying the game. To paraphrase one player, "I kinda felt like there was no plot." Now, I could have blown my stack at that, because we had literally had a pre-game conversation around "I don't run plots," but I felt like that response was missing the player's point.

He was having a hard time caring about the campaign, because it was, literally, not something he cared about. In Robin Laws' categorization, he's a Wallflower. He likes to show up, roll some dice, have some cool spotlight time now and again, and that's it.

He was a bad fit for this campaign, and I should have seen that coming. Skerples tells us:

Selecting players for a game is like selecting ingredients for a dish. Sometimes, you're making do with the stuff in the fridge. Sometimes, you get to pick and choose. Sometimes (as with drop-in games), a bunch of ingredients shows up on your doorstep and you need to try and figure out what you can cook with them.

Post: OSR: Behind the Curtain: Session 1 Examination, 07/12/22
This actually holds up pretty well here. I had functionally lost a player before I even started the game, and I should have been aware of how different his playstyle was from what this game demanded.

Lesson Two: Run Collaborative Session Zeroes

This one would have solved like half the issues I ran into in this friggin' game, y'all. I figured we had all made characters in isolation in our last game, and that had turned out fine, so this game would be fine, too. In retrospect, this is like saying, "Well, the last time I played Russian Roulette, nothing happened, so this time it'll be fine, too!"

What actually happened is that no one was invested in anyone else's characters. Everybody kinda just stayed silo'ed the entire game. They never gelled as a party, because no one had built a character that would care about the party. You remember how I said I was really proud of building in downtime to the game, with weeks spent in space as they recharged? I figured that would be great fodder for RP-heavy scenes as PCs processed the weird alien crap they just saw.

In actuality, what happened was this:

Me: "OK, you spend five weeks in space together. What are you all doing?"

PCs: .....

Me: "Anybody want to have a scene with anybody else?"

PCs: .....

Me: "The NPCs?"

PC: "Nah, I'm good."

It was painful, y'all. And so much of it could have been avoided. Another thing that popped up was that one character wanted to be a cool pilot with her own ship! And of course! Why wouldn't you? That's an awesome character concept!

I decided to just give her the party's ship, and that was what her prelude was about - stealing her family's old ship back from these shadowy government types that would come up later in the game, and to which everyone else had a tie, as well.

This was about as smart as asking Michael Vick to dog-sit for you. The player thought, "Oh, cool! For this first game, I can create a cool moment of tension for the party by telling them all that there's a fee if they want to go on the adventure!"

That is a legit source of tension. It ****ing infuriated some of the other players. She never cleared that with me, or with the rest of the group (which, why would she? I hadn't established that as an expectation), and it kind of blew up. She never wound up collecting on it, and it was pretty clear that the weird energy at the table was notable, but I did an even dumber thing here:

I didn't intervene immediately.

I should have. "Pay 50 gp to go on this adventure" is basically a flashing red sign that you've screwed up as a DM. But I didn't. I instead reached out afterwards, and so freaked out this player (who just thought she was creating a cool scene) that she told the rest of the party they could ignore her character, or throw her in the brig.

End result? I basically set that player up to fail. That whole situation did not need to happen. The key to building PCs that are interested in one another? Build that in from the jump, with a collaborative session zero. Otherwise, you're playing with fire - sometimes, it'll catch. Sometimes, it'll fizzle. This time, it fizzled so hard that it lowered the godd@mn room temperature.

Lesson Three: Too Many Notes Is Cacophony, Not Symphony

One other piece of feedback I got was around how overwhelmed the PCs felt. See, I had built in interconnectivity between all my star systems, so if the PCs spent more than a few weeks in a star system, a random encounter check could give them a new rumor about another star system. From the second session to about the ninth, the PCs (through just dumb luck and a lot of socializing), generated about three rumors a session. So that's about 24 rumors. So the PCs are looking at the quest board as it starts to balloon like Aunt friggin' Marge just aggravated a pubescent wizard, and can't decide where to go.

This is also a function of the overdesigned sandbox. By creating thirty freaking systems, they were constantly lost as to where to go.

Moreover, my core gameplay loop failed. They never had a mechanical reason to go back to the main port; they got their fuel from solar charging, so there was never any need to refuel. I never set up a system to track food or supplies, so they were free to wander pretty much aimlessly. The reason it felt aimless was because without meaningful constraints ("We can only pick two options!"), navigational choices feel meaningless. Because there was never an impactful cost, there was never a real choice. It was *Spin the DM's Twister Card of Quest Options* and go.

Analysis paralysis and aimless wandering? Pretty much the two biggest things that you'll hear as negatives for sandbox play. I knew about them, planned to avoid them...and still got slapped upside the head with 'em.

Now, for my next turn running this game, I plan to address this by enforcing fuel as a constraint. No more free recharges; if you run out of fuel, you can make a last ditch effort to get some helium-3, but that's dangerous and will definitely damage your ship. Fuel will be costly, and repairs will be even more costly. Doing so keeps the PCs hungry for the next score, and keeps them wanting to go on adventures.

Lesson Four: No One Cares Unless You Give Them A Reason To

Another piece of feedback I got was that the campaign lacked stakes. Again, I tried to tamp down on my temper and really listen to the player giving me this feedback - she told me that the scale made the stakes feel grander, but less meaningful. She talked about saving an entire planet (something they managed to do by the skin of their teeth!), but said that it really didn't resonate with her. Because she didn't care at all about the planet itself. Like, yeah, they saved a planet. Cool. But...they don't really care about the planet, despite just saving the people who lived there from an awful AI-created tyranny.

Now, I could have countered by saying that in a player-driven exploration campaign, the players are supposed to tell the DM what they care the most about, and then spend as much time as they want to dealing with that. But again, that misses the point - she's telling me that she's having a hard time caring about anything, so there's nothing for her to emotionally invest in.

Saving a planet didn't matter to her. Getting jackfruits to an emotionally stunted mob boss named Jack-Jack so he could fulfill his lifelong dream of selling people "Jack-Jack's Jackfruit"? That mattered, because she could invest in that. And so most of those 30 star systems (and like two months of my life) did not matter worth a darn.

Lesson Five: Sandboxes Give You A LOT Of Tools. Pick Three

The PCs also mentioned that they didn't care about most of the factions in play. If anything, they enjoyed mucking with one or two, but that was it. They liked hunting pirates, but did not care at all about the colonial resistance movements they freeloaded off of. Sandboxes can be designed packed with content. Players check out after like three things, or at least mine did. Focus on a couple of factions, people, groups, or places, and introduce a third as the tension escalates between the first two. Let them get to know only a couple of things, and it'll make it more meaningful when you rotate everything out.

Lesson Six: Encounter Design Is Garbage

Esper Genesis inherits a lot of the design flaws from 5E. You can tell where the designers have tried to fix a few things, and where they just threw up their hands and said, "Ah, screw it."

Encounter design is clearly one of those places.

In a sandbox, your players will likely have multiple long rests between locations, unless you are playing with gritty realism or slow natural healing rules. That means that 1) random encounters will need to be triply deadly to make a mark if they are focused on combat, 2) You have a wide range of level latitudes when you build your locations, and 3) Triple-deadly combats get boring fast. Play Tomb of Annihilation to see all of these in action.

I approached this by avoiding gating my locations by level. In a few cases, I didn't even have statblocks written down. It worked well; a combination of a lot of 1/4 CR gang members and a CR 13 shadow technocrat was a pretty solid encounter for a level 8 party. I also noticed that by randomizing their reactions to the PCs and avoiding using combat-focused statblocks, I was able to focus on RP a lot more. The party wasn't being rewarded for combat, and actively avoided it frequently. Ergo, a lot more of this game was focused on finding creative ways to avoid combat and make those sweet, sweet discoveries.

Lesson Seven: Letting Players Murder Their Own Fun Is Miserable

I screwed up early on and told the players that they could play through the quests they got in whatever order they got them in. I did this to remove a lot of that analysis paralysis, even though it violates one of my core principles about running games (that time should matter).

This was very, very dumb. It was about as dumb as calling Alexander the Great a cotton-headed ninnymuggins.

It immediately removed all time pressure from the game. The players treated the sector like it was Skyrim, and told me so several times. "We're doing this Skyrim-style; all side quests, baby!"

I let them, because I figured if they murdered their own fun, it would encourage them to stop screwing around.

This did not work. To quoth St. Gygax:

YOU CAN NOT HAVE A MEANINGFUL CAMPAIGN IF STRICT TIME RECORDS ARE NOT KEPT
I mean, the guy's a bit of schmuck, but on this point, he's on firm ground.

Letting players murder their own fun is miserable, both for the players (who are pissed they are "doing the right thing" and feel crappy about it) and the DM (who is just letting their world get trampled on). Don't do this. Just talk to your players, or lay down immediately how you want to enforce the world.

OK, y'all, I've got all that off my chest. Ask questions if you want. Laugh, learn, hug your players / DM (with consent!). I'm going to go blow some stuff up in Battletech. Thanks for reading along!

Note to the mods: if this needs to be moved, just let me know and I'll submit a ticket.