When I was young — say between 15 and 21 — I pretty much exclusively ran sandbox games. Hex crawls, really. I’d make a map, usually a huge map, with some named locations on it and a lot of different terrain and then set off the adventure with some very basic kicker, like a rites-of-passage quest to get 12 Amusing Things. And then the game would essentially be driven by random encounters and me ad libbing Story off of randomly generated magic items and my colourful place names. This was very satisfying gaming.
I’ve run plenty of unsandboxy games too. What would a good word for that be anyway? Mission-driven gaming was my favourite — spies with an objective, that kind of thing. Episodic. Still plenty of ad lib since all I’d write down was the mission brief and then wing everything else. Sometimes the mission brief was crazy simple, a phone call perhaps, with someone hysterically wailing about carnage at the Michael Jackson concert.
Turns out it was werewolves.
So sandboxery isn’t ad lib. I can ad lib at least two ways.
So what is it? Am I sandboxerizing now with the Soft Horizon games? Let me tell you how they run and you can tell me.
A Soft Horizon game starts with characters, an organization they belong to and a kicker. The kicker is a problem the organization has that needs solving. It’s vague. Something like “a client got de-certified here; find out what’s up”. And then there’s a place that’s randomly created with just a couple of phrases. And then there’s the over-arching conflict of the plane. In The King Machine it’s the fact that the King is bad and the King’s bad actions are ruining a Good Thing here. There’s some advice about what that might entail. In Sand Dogs it’s the fact that there are tombs full of sleeping gods and improbable goods and people are literally dying to turn that into wealth.
And then as the players address their central conflict they screw up and the system generates new problems. As ref I pretty much just ad lib descriptive text around that problem and then go with the flow. All I have for plans is a sheet that has one or two ideas for “starting some shit” should things slow down.
So is that sandboxery? There’s no map (well there’s a relationship map). There are no encounter tables (though the game twists around a die roll that performs the same twist-the-plot function as an encounter table). It feels pretty sandboxish to me.
But then I’ve heard people say that sandboxing requires a lot of prep because you never know which way the players will go next. Well the system seems to do that for me just fine so is that not sandboxing? Do I need to prep a whole world? I never did that. Maybe I never sandboxed.
What the hell have I been playing all these years? Does it lack a category? Or are categories mostly bullshit? Or somewhere in between — maybe no category can really embrace anything but rather has some idealized play and then almost everything is clustered around the tails of that bell curve.
Could you maybe plot gaming on two axes, say Plot Planning and World Planning and find categories that way?

So where would we put an optimal sandbox game? From some of the things I’ve read we’re looking at:
Is that right? Doesn’t seem to cover all the talk about sandboxing but certainly some of it. Maybe the whole left hand side is sandboxish.
Where are your games on there? What would you call that? Are all hex crawls sandboxes? Certainly all sand boxes are not hex crawls.
Most importantly though, if I tell you my game is a sandbox and you buy it and then disagree, are you going to be upset?