Hmmm... This seems like a general truism of RP play rather than something meaningfully insightful about the play process of typical sandbox games or even your games. Of course the actual process of play tends to be messier at the table or the conversation between players and GM is not as rote as it's made out to be. That's basically tantamount to humans interacting over a game.
I think that this is why some posters struggle with your posts and would also like for you to self-deconstruct the "evocative," "metaphorical," or "romanticized" language because it obscures, whether intentionally or not, the underlying play process of "how it works." It can come across as "generic," "banal," or even "trite," which results in undercutting the message of what makes typical sandbox play work and how.
Well, I think part of it is we are not as interested in reshaping the core experience of play. For example DramaSystem, which I just mention because it is the game I am most familiar with that departs from general RPG approaches and really focuses on mechanics that deliver a particular experience, is all about reshaping and structuring the way scenes in games are introduced (and it is specifically structured around scenes, which most general RPGs aren't, or at least aren't in a concrete and explicit way: they may loosely resemble scenes). There is a whole procedure for establishing the scene, establishing who is in it, then characters needing to advance their agenda within the scene (and there are meta-currencies that work as carrots to drive drama). Laws clearly spent time thinking about the fundamental process and why he wasn't getting the drama he wanted in say a standard RPG. Most sandbox GMs are satisfied with the core process. For them, and for myself, the core process is working fine (and the core process might vary from group to group: but it is that fundamental exchange that most RPG books point to of the GM saying what the PCs see, the pcs saying what they do, and the GM ruling or saying what happens next---and obviously this is much more open and organic than that suggests because a lot of exchanges are fluid dialogue between characters). The things that matter to the sandbox GM are advice, tools, managing rulings, what not to do, etc. And I think that is because the big problem sandbox is interested in is avoiding railroads and avoiding GM as storyteller. I think the latter may be less true of old old school sandbox GMs (who maybe were running sandboxes prior to the 90s), but OSR sandbox GMs and people who came to it after the 90s I think were having the 3 pronged reaction against railroading, 00s encounter balance/structure and 90s storytelling.
For me it isn't a concern if the procedures are seen as trite, banal or generic. I would describe my style of GMing and the kinds of systems I gravitate towards as traditional, and there is an assumed culture of play around that. I think the focus for me is more on adventure structures, tools, making sure the table is functional, having fun, and long term campaigns.
Again I am not the best mouth piece, there are people much better than me at breaking down how to run a sandbox. I have always been much more intuitive and emotional in my explanation and descriptions of these things. All I can do is share what tools and approaches work for me (here I am answering some of the questions you raised in another post which I said I would get back to later). This all only applies to me but it is also stuff I have picked up over the years from the sandbox circles I travel in (I am definitely not a representative of 'pure sandbox'----my sandbox concept starts with Feast of Goblyns and that is very atypical):
1) Embrace that it is a game and embrace the role of surprise: I think this is really important. If the GM isn't being surprised, then I do think that is when you can start getting into the territory Pemerton is talking about of 'playing to discover the GM's notes' (which is how I used to describe my frustration with running those EL/CL based linear adventures back in 3rd edition D&D "I might as well just hand the player my notes" is pretty much what I felt after every game).
There is a lot here but I think important elements to this include letting the dice fall where they may, disconnecting yourself from your interest in the PCs survival and success, and disconnecting yourself from wanting the campaign to go in a particular direction
As an example, I had a session last night where the party was defeated by a psycho-path granny---its a wuxia scenario so she was powerful---who put them in coffins and dangled them over a chasm. Her method for killing was to set 'feast beetles' on the coffin which would eventually kill those inside (the process is elaborate but this is enough for the example). This was a ticking clock situation where I
decided based on what I knew about this character that she liked to torment people and would first kill the person the party had come to rescue (who was in a coffin beside them). So I marked down a bunch of boxes indicating the number of rounds it would take for the feast beetles to reach and kill her, then marked the round at which point the old lady would set feast beetles on the party (so I had a concrete sense of how much time the players had to escape before each of these things happened, knowing it was still flexible because the old lady could react to their actions). I also clearly noted the integrity of the coffins they were in, the integrity of the ropes they were bound by. Then I made sure I ran every segment of the situation by the book, and I kept consulting with the players to see if they thought a particular ruling was fair. If I didn't have a clear answer on what a ruling would be, because of how dangerous the situation was (they were helpless in coffins so I wanted to be as fair as possible) I talked with them about things like "What do you think the Target Number should be here" or "Is it fair in your opinion for me to require an Athletics roll for you to cling to the side of the cavern after you make your jump". None of the previous information is super important, but it is just being put down so you can see my process (hopefully I have laid it out enough). We basically went round to round, taking each character's actions step by step. It was more granular than normal because of the situation (if there were not such high stakes, this moment might have moved a lot faster, but I wanted to chart every step for fairness)
The important thing here is this old lady is a serial killer in the setting whose homestead and cavern complex I have mapped out, who I fleshed out before hand. I have about two paragraphs on her (which is the most I like for any NPC, though I will do more if I need). It is a simple cavern complex but the chasm is above a pool of water that leads into a cave where another former martial hero lives (another old woman in this case) who is imprisoned and under the effects of a special poison to make her subservient (she lives passively in the cavern making baskets and coffins for her captor even though she is technically powerful enough to beat the old lady who imprisons her, because her will is so depleted). I designed this whole arrangement as a nod to a couple of scenes in Condor Heroes and Return of Condor Heroes (where the main character (s) is cast into a dire imprisonement situation but finds themselves in the company of a great master. This is somewhat artificial. But I am okay with it from time to time. The key thing though is as this is playing out, while it would be great for the party to find themselves at the bottom of the cavern in order to
discover this old lady, that isn't the point of play at all here. I need to not care if they encounter her or not, so that means not doing even subtle things that would direct them to the fact that going down the chasm is possibly a safe, fun or good idea. It so happens in this game one of the characters had a really bad roll and fell into the water and they ended up there anyways. But the point here is when I run these kinds of situations, I need to be content when the players 'miss' something and never learn about it. Now it is living, so the two old ladies are 'still in play' in the game (there could be a time later when one or both of them become relevant again) but if the players just went up and escaped and never came back, that is fine by me. The other point of this scenario was I set up everything very procedurally and business like. When I talked to the players, I wasn't being dramatic or anything (my delivery is pretty dry actually). But I was intentionally making this scene the same as if I was noting stock or keeping an accounting book. Because I didn't want to save or try to kill the party. I wanted it to be as above board as possible. So there was a woman there they were trying to rescue. She was on a timer basically. If they broke out and freed themselves, they could have saved her, if they failed she would die (in this case they failed and she died). By the same token, one or all of them could have been killed by the feast beetles. The point of this kind of moment is I don't know how it is going to turn out at all (and the previous session before they arrived, it was the same, I didn't know they would end up in those coffins when they confronted the old lady (I knew that was her modus operandi, so she would definitely do that to them if she could, but I had no investment in a dramatic cliffhanging scene like that when the fight happened).
2) A world created: Definitely an established world designed with geography, structures (in the old school annales sense of the word---my world building is very Fernand Braudel inspired), pre-history, history, myth (and the reality of that myth), is the starting point. People all approach this differently. The Rob Conley posts I linked are a deep dive into pretty granular approaches to dealing with climate and weather (I am not as scientifically minded as him on that front). I liked having firm ground on which the living world concept plays out. And that means I usually begin with cosmology and some general concepts of where I want to go, then I start with a planet, followed by a continent or continents, then I start mapping out the early stages of people on that map (which will vary according to the specifics). So just as an example in Sertorius I placed the first humanoids on the map (who were all created by various gods), noted their languages, and then charted the changes in something like 500 year chunks (sometimes longer sometimes shorter). In this step I am doing one map for each era and figuring out migration, movements of language, development of things like early city states, until I eventually get to something like looks like a world at approximately the stage of the time of the Han or Rome. There might be a golden era in the past too (as there was in Sertorius where you had an orc and an Ogre Empire loosely modeled after the details of the Ramayana/Ramakien. In this case there was a cataclysmic event, the killing of a god, that unleashed magic into the world. The movement of languages is pretty important to me. I like to know how people and languages evolved in the setting over time. This helps me to name locations, and it also helps me understand how things like dungeons might be placed in the setting (I am not big into dungeon crawl but I do have a lot of underground tombs and limited complexes in the style of old conan stories). In Sertorius I spent a lot of time working out the languages. Not all of them got he same treatment (one was essentially just 'not latin'), but my 'not arabic' language actually got pretty deep because I studied Arabic and wanted to bring some of the structural elements that I liked to it, and I also wanted to play around with how things like official titles can vary by region. In Wandering Heroes of Ogre Gate, I focused less on the early movement of people, and more on the development of what became the empire of the setting (which is in the setting is a song Dynasty Analog). And this was done in the way I charted out the movement of peoples in Sertorius (should note that these maps are not in the sertorius books themselves as it seemed to expansive to commission them at the time, but in Ogre Gate I was able to include most of, but not all, my maps of the different eras leading up to the present empire). I also focused the underlying principles of the cosmology in Ogre Gate (which helped me as I made the setting). After I have all that, I start focusing on present day details: cultural elements (i.e. imperial exams, calendars, institutions, sects, laws, marriage, etc). In Ogre Gate there is the geography fo the world itself and the empire and political powers. There is a lot more than this probably but the overall point is to create a place that has foundation to it, that does't feel like it is shifting around the players at my whim or their's.
3) Living World: This is the heart of sandbox play for me. At least it is very crucial. Just to repost where I take this concept from, here is the Feast of Goblyns entry:
Now, this might be considered banal here to posters. I don't know. But for me this was a light bulb moment that radically changed how I played and enabled me, at least in the context of things like monster hunts (I wasn't running sandboxes at all at the time) to have a more lifelike, surprise filled, and open type of adventure.
But the living world to me is the stuff that exists in the created world. So this is where you really bring the things like institutions, sects, NPCs to live, this is where the motion int he campaign is in terms of things existing independent of the PCs. However on a slower scale some of it is stuff like geology (i.e. earthquakes, comets, floods, etc). It serves two functions, one is simply to have a world that exists and is in motion on its own.
So when I make my calendar I often plant specific future events on the calendar (earthquakes, assassination attempts, etc). I also have a table called the Table of Future events which I can roll on periodically. It has three stages after you roll on the initial table to find where the event occurs. The first is for monthly events. Usually this is something like "A new local magnate emerges", "Disaster Strikes a village or town", "A new song, poem or work of literature gains populatrity", or "bandits plague the region". Typically not terribly important and the table I find needs regular updating simply to improve the entries and refine them. But if you get "Roll on Table II" that brings you to Major Event or Development (this is a result of 2 on a 2d10 roll). Table II has results like "Two martial sects go to war", "Minor Invasion", "Key figure in martial world dies", "Natural Disaster", or "Minor Uprising occurs". This table has a result of TABLE III: HISTORICAL TURNING POINT; which is a result of 2 on the table (maybe higher chance than the real world but good enough for a game setting). Table III includes things like "new trade good discovered", "Moral Panic Spreads", "State Collapses", "Major Invasion", "Major Uprising Occurs", "Government Starts Major Project such as a canal or other works", etc. This stuff can sometimes be exciting, but mostly it is to keep the world moving and in a state of change around the PCs.
The other function is more directly connected to the players and that is so the world reacts to them, and is filled with characters who are not just sitting there waiting for them to go on an adventure and find them. If the players become involved in the martial world for example and start wheeling and dealing, they will start meeting characters who have goals of their own and things could pan out in any number of ways depending on what the players do and what the NPCs do (either as a product of their motivation on its own or as a response to something the players do). These characters form groups, move around, change plans, adapt, etc. The way this is done is primarily around making NPCs with clear motivations, charting out alliances and group relationships, tracking what these NPCs are doing as the players do what they do. So say they meet scholar Han and piss him off for some reason (like they don't like some scheme he has going on and put an end to it), but they leave it at that. He might go and hire three men to help him kill the party and come after them (or maybe he goes to the empire and seeks their help, informing officials that the players are wanted men. Sometimes I simply play such characters like I would a PC. I know Scholar Han is after them, so I decide what he is doing each session, and I decide what he does during the session. Sometimes I give this role to a friend not in the campaign to make the NPC more ferocious (in which case I let them make decisions about actions and resources between sessions). Other times I rely on my grudge tables (if I have six NPCs or Sects after the party, that might be unwieldy, and their grudges can slip my mind, so I have a regular table I roll on that includes anyone they have a grudge with and occasionally they simply come up as a result (and the rythym of the tables feels about right to me). Mind you, this is all in the context of a wuxia setting, so grudges are important, it is the only aspect of living NPCs, and even in the wuxia setting there are lot more kinds of relationships than this. This principle extends to sects, government,s etc.
4) Exploration: This is probably where a lot of the issues are coming from in this thread. This is a hard one because every sandbox GM handles it differently. Some use hex crawls (most use hexes either way even if not crawling). Some take a more open approach. When it comes to things like local explorations (dungeons say) you often see the classic gygax or moldvay approaches. Justin Alexander has a whole in depth exploration of hex crawls for example. Rob Conley has a discussion too where he talks about the ideal size of hexes (the baseline hex in his view should be about the limit of human vision, so characters can see to the edge of each hex effectively). I don't think there is a one true way here. I am, as you might guess, much more lose and hand wavy around this. I do all theater of the mind and I am not that into rigidly doing a hex crawl or tracking turns in a dungeon. I like things fluid. Still I have my tools. In my case Survival Rolls are a big part of it. Players can say they go wherever they want, but this will require a survival roll by one member of the party (presumably the person with the highest roll). Survival is an open skill and can be taken for different terrain (cities, wilderness, underground, seas, etc). For travel, it is basically one roll a day, if you fail you have an encounter. But on the local level, or through more dangers terrain, it may be more frequent. So a dungeon might be every ten minutes make a roll (and many dungeons would have additional encounter tables on top of this). With cities, I use wards or quarters as dividing lines, so rolling survival anytime you move from one quarter to another (and players can always roll survival in a city for things like trying to get information: though the point of contact with an information roll will usually be played out in 1st person). There is obviously a lot more to explication than just this, but it is a big topic so I will leave this here for now and can follow up if you have any questions on it.
5) Encounter Tables: This are a pretty important tool in most sandboxes and I have found there is a real art to putting together workable tables. Again they are tool, and most sandbox GMs consider tools optional. But I use them consistently int he ways mentioned above. I like to layer my tables and pin them to regions. And so the first table might be things more like local law enforcement, bandits, fated encounters, imperial officials, grudge encounters, local sects, etc. The next table will be more dangerous threats: sheriffs, large numbers of bandits, imperial agents, et. Rolls on these tables can lead to Local and World personality tables (where I put all my characters in the setting who are local, followed by a table for all the characters in the setting itself). On top of all this, I am always free to just have an encounter happen. And again I do drama sandbox, so sometimes my encounters will be more than just rolls on a table.
6) Rulings: This is pretty fundamental to most sandbox games I have been in. Obviously it depends on system, but the idea is, when the players propose doing something that isn't covered strictly by the rules, you elaborate on existing mechanics to provide a resolution for that request. This is so it doesn't feel like they are always pushing on buttons and that taking specific actions can have more specific outcomes. There are other ideas in rulings as well but this is just the simplified version. It isn't unique to sandbox, like most of this advice, but it is important.
7) Scenery smashing: A good sandbox game runs itself, and one way that you allow that to happen is by letting the players tear apart your world: kill NPCs, take over things, form alliances, form grudges, start their own organizations, get embroiled in conflict with groups, destroy institutions, build institutions, join institutions, etc. This is I think very important. Again this might seem pretty banal, but when you lean into allowing players to weld their power how they like, even if that means going in wild directions you don't expect, it starts to give the campaign its own scene. And I think the GM needs to maintain a mindset where 1) he or she doesn't have strong expectations of what is to happen---you might have plans because the campaign was going in a certain direction, but your 'stance' (in the martial sense, not the GNS sense) should be neutral, relaxed, flexible. You should be avoiding feeling any resistance to the players saying that want to go this way, do this thing (with obvious exceptions like things you simply don't want to happen at the table for moral reasons or something),etc. The idea is to honestly think through what the consequences would be and go with it. Even if the players do something outrageous like amass serious power (which you shouldn't just hand to them but if they legitimately obtain it), instead of feeling the campaign is negatively impacted and needs to get on course you adapt and realize 'this just means there are new complications and activities for them to deal with and pursue). Again, it may seem banal, but it is something we learn to resist and so it can take effort to do
8) Organizations: This is covered in living world but deserves some focus. I like to give numbers of members to my organization, stats for disciples, stats for senior disciples, stats and entries for 'named members', full descriptions of the organization, its purpose, its headquarters, its leader. If I can I sometimes provide names for all of the members if the group isn't too big (and the names often serve as clues to personality or motives). Then I usually map out the different groups, noting whose in conflict, who has alliances, what goals they are perusing etc. Often my campaigns tend to focus around organizations and institutions.
Hopefully that answers some of your questions. My mind is a little fuzzy today so I rambled a lot.
