On Sandboxing and Game Balance

I've been Gamemastering since 1978, and I've always run a "sandbox" (what I used to call "free flow") game.  The way I have done it all along is to create a framework in which the world's history, current events and Main Non-Player Characters (NPCs) exist independently from the Player Characters.  So the primary NPCs are busy going about their business in my world with our without the Players.  They have objectives and try to achieve them.  Some of them are good guys, some of them are bad guys.  Some are lawful, and some are chaotic, and most fall somewhere in between.

When the Player Characters enter the world, whose name is Elthos btw, they can, but do not necessary have to, cross paths with the major NPCs and/or their minions.  Sometimes the PCs may join forces with them, or wind up fighting against them, or sometimes just wave as they pass by.  Often times they wind up in opposition with the goals of the NPCs and so a conflict ensues.  These conflicts usually engage the minions who may be running missions in accordance with their leader's objectives.  The PCs will in these cases try to stop them.  Eventually, perhaps, word of these efforts reach the upper tiers and a Main Character NPC may become aware of the Player Characters, and begin to work against them actively.  Especially if they've been successful at thwarting their plans.

This all is as it should be and this style of play lends itself to being fun for both myself and the Players.  I never quite know how any given game is going to go, nor do I try to steer the Players toward any specific objective or goal as the GM.  That said, they may be in the service of a Main Character NPC who definitely will be trying to steer them toward a specific objective, but I take on that role of NPC as though I were playing the Character him or herself, so I don't consider it me as GM trying to steer the group toward an end goal for the game, and I certainly don't tell the Players which NPCs they should have an association with.  I leave all that decision making to them, and flow along with whatever they decide.

Sometimes this leads to difficult situations.  They may, for example, decide that some activity is so wrong and terrible that they must wage a campaign against it.  But the underlying story, the Main Character NPC, and the situation may cause them to engage with an opponent that is completely over their heads.  When the fighting starts, they may realize that they're totally outnumbered, or out classed by the forces they're opposing.

And this is where the problem of Sandbox and Game Balance come in.  Some GMs try to always establish a balance between the PC group and their opponents, in an effort to keep the game fair and avoid the twin problems of Too Hard and Too Easy.  While the goal is laudable, in some respects, it is also, I find, impractical to some degree.  The reasons why are as follows.

One, it is very difficult, depending on the RPG rules, to determine what the balance between two forces actually is, especially in cases where you have randomized initiative rolls to determine who strikes first.  I've surmised from my studies that getting an accurate calculation as to what percent chance either side has of winning would involve the use of binomial math, and after going back and forth on that for some time (I originally thought deriving a calculation would be relatively easy), I was informed by mathematicians far better than I that the only practical way to do it would be to run simulated combats between the two groups a hundred or so times and find out from that what percent of those runs results in one side winning over the other.  I'm pretty sure that most GMs don't do that.  And I'm pretty sure that most GMs actually just eyeball the thing and say something like "Yeah, 10 Orcs and an Orgre ... that shouldn't be too hard for these guys", and leave it at that.  The problem, as we all know, is that this is a hopelessly inaccurate process and the results vary considerably, especially if the forces arrayed have unusual powers at their disposal.

The second problem is that as GM of a sandbox world, I'm never quite sure I know who is going to actually be in a battle at any given encounter.  This is because the Players may split the party.  Yes, I know - everyone knows - you should never split the party.  But since I don't tell my Players what they should do, I simply go with whatever they decide to do, they wind up splitting the party (as often as not to their regret, but so be it).  So no matter how maticulously I might (and I don't) try to create "Balanced Encounters" there's really no way I will know if the encounter is harshly one sided or not until the moment the party encounters it.  After all Players may split the party exactly one moment before the actual encounter ("Ok, we'll bust the door down and charge into this room, but you guys go down the stairs and block any Orcs that may come up this way while we fight whatever's in there.")  This kind of thing has been known to happen.  So attempting to create Balanced Encounters as often as not simply doesn't work.  This is a result of allowing Players to do what they want, even if it is not necessarily the best idea in the world.  And yes, there's pros and cons to this approach.  Some GMs avoid it by running Railroad Campaigns.  And while some will argue that such a thing should never be, my own feeling about it is that it's ok, so long as the Players are ok with it and everyone has fun.  But I don't run Railroad Campaigns and I don't want to.  I like the free flow style, and I find it more exciting a way for me as GM to play the game.  After all, why should the Players be the only ones who get surprised?

So for me, I tend to sacrifice Balance for Freedom.  I let the Players do what they want to do, without much guidance (especially if they don't ask for it from any NPCs who might be able to offer advice).  But at the same time the Player Characters are at risk of encounters that can squish them like little bugs if they aren't careful.  Or be way too easy for them.  The way I handle that is by randomizing the encounters to a certain degree.  So while I know that a certain area of the scene (in this case the secret underground township of Whitewode) has a certain kind of opponent, I don't determine in advance exactly how many there are, or even what their exact composition will be.  I roll for it at the time.  To make this work for me, I usually generate the NPCs randomly in advance, but when the encounter happens I roll to see how many of those NPCs happen to be on the scene.  So what happens next is to some degree a matter of luck, but it also very much depends on how clever the Players are when they hit the encounter.  I have rules in my game (Elthos RPG) that allow for a variety of ways to attempt to cut and run in case the odds are overwhelming.  Sometimes they take advantage of that to escape before they get themselves killed.  Other times they plow in and hope for the best.  As it happens my Players are damn lucky die rollers for some reason, and most of the time they manage to get through and achieve their goals.  I've often been shocked by the incredibly good timing of their "Critical Hits".  But that's luck for ya.

At any rate, that's how I handle my Sandbox world in relation to Balanced Encounters.  I don't really try too hard to make the Balance, I just let things play out based on the luck of the rolls.  Sometimes it goes poorly for the PCs.  Sometimes it goes poorly for the NPCs.  But either way, the risk of calamity makes the game exciting, and when those incredibly lucky rolls do happen you can bet there are loud cries of delight and amazement around my table.  As there should be.  :)