The Dungeon Dozen

I would like to take a moment to recommend The Dungeon Dozen by Jason Sholtis, the mastermind at The Dungeon Dozen blog.

I love books that help get your creative juices flowing.  The Dungeon Dozen (Volume 1!) belongs on your shelf along with The Dungeon Alphabet, The Random Esoteric Monster Generator, and The d30 Sandbox and DM's Companions.

Like all the best inspirational books, The Dungeon Dozen has artwork that is worth the price of admission by itself, and is a meaty book filled with tables and ideas.  In fact, it is easily among the best "random tables" books I've seen.  
If this book doesn't inspire hundreds of hours of game play at your table, you must not have bought it.

This really is top-notch work.  You can get it here.  And at the same time, you really should be following the author's blog.

Ghoul Friend

In the comments to the last post, Wyatt Allworthy wrote:

Something only tangentially related to undead causing plagues that I wondered if you had any experience making work as a ref. I don’t know how it’s done in DCC, but in A/D&D you had the situation of undead like Ghouls with a paralytic touch, or Wights, etc that drained levels. These creatures could be encountered even by low level characters, in confined crypt like places, where they might have no exits to evade them. A ghoul had a speed of 9” and characters loaded down plate armor, let alone equipment and loot would be limited to a speed of 6”, as the speed of the slower members. How can a party survive something that paralyzes its lead members just by a touch, which they will almost assuredly fail their saving rolls to fight it? Only a tiny number of these ghouls would overwhelm a low level party, almost assuredly, every time they were encountered.

I know that back in the day, parties had larger numbers of players and possibly lots of hirelings and henchmen, is that the way to manage it, or is there some way to make a 6-man special forces style team of adventurers competent to handle paralytic touch undead (let alone level draining undead).

Thanks for any insights on this one, it’s a puzzle for me.

First off, low-level characters are unlikely to be loaded down with plate armour in any game that I am running.  That’s simply a matter of expenditure – plate armour is expensive, and there is not enough “spare” cash for this particular expenditure.  Loaded down with loot is a lot easier – in a question of “keep your loot or keep your life”, smart players choose to drop the loot.

I like ghouls, and I do use them at low levels.  I have been throwing ghouls at 1^st level PCs ever since reading the evocative play description in the 1^st Edition Dungeon Master’s Guide.  As a player, I have encountered ghouls at 1^st level as well.  In one memorable 2^nd Edition campaign, the DM (the esteemed Jesse Donahue) lured the party into an un-dead haunted swamp, where the easiest way to survive was to run and hide, then run and hide some more.   At the same time, I was running my megadungeon, The Dungeon of Thale, in Venice Beach, California, and there were roving ghouls on the first level.  I think they got perhaps one or two characters over a long period of play.

But, then, these characters weren’t clunking around in plate.  That heavy armour affords you one sort of protection (better AC) while denying you another (making it hard to run away) is a trade-off that makes for interesting choices.  Dungeon Crawl Classics does that one better, by making heavier armour subject to more devastating fumbles as well.

An illustration I drew based on Jesse's game

Things that I have found adjust the odds against ghouls are teamwork, good use of magic, having a cleric on hand, and having an elf or two in the party.  In DCC, you should also consider burning that Luck in order to make your saving throw, especially if you are the last PC standing.   In many games that I have run, ghouls shun sunlight, and will not willingly enter it or an area of continual light.  Having some areas that the PCs can retreat to, while leaving them with a serious problem that still needs to be solved before their own food runs out, can be fun for all concerned.

I’ve run James Raggi’s Death Frost Doom to great effect, using the DCC ruleset.  How you deal with a horde of ghouls and zombies is a major part of that adventure.  At first, the players thought the answer was “you fight them”…but that is not a very viable answer in Death Frost Doom.  Sometimes, in a good adventure, fighting should not be the best option.  Sometimes, it should be a suicidal option.

If you go poking around crypts and barrows at night, you should expect to run into the un-dead.  If it is possible, save your explorations for daylight hours.  At least that way, you may be able to retreat into the sun.  As you explore, consider how you can use the areas you have already examined to your best advantage.  Mindless creatures, especially, may be lured into traps that you discovered and bypassed.  There might be choke points where a few can hold off many.   There might be places where a barricade can hold foes so that the archers can do their work.  Never underestimate the value of a spear or a pitchfork when you can hold your opponents so that they can’t reach you.

Even so, sometimes, you have to let the dead devour your fallen so that you have a chance to get away.  And sometimes the ghouls get you.  It is completely okay to have the entire party wiped out after mere minutes exploring the Barrowmaze.  Those are the risks adventurers face.  

Hooks and Win Conditions

It is strongly my position that my job as GM is to supply players with context, from which the players make choices, and then I adjudicate the consequences of those choices.  This adjudication, which includes both success and failure, as well as every grey shade between, creates a new context from which additional choices are made.

The players' choices do not come from a menu.  Every ruleset embodies certain default choices within a framework of rules, but that does not mean that players cannot have their characters attempt anything, even if that "anything" requires an adjudication from outside the rules or modifies the rules themselves. The players are not guaranteed to succeed, and I will keep the rules in mind, but if the players come up with a reasonable means to fuel a spell with a major sacrifice, in keeping with the game context, why wouldn't I allow it?  The "reality" of the game milieu trumps the "reality" of the ruleset.

What does this have to do with hooks?  Well, adventure hooks are sample win conditions that the players can latch onto in order to set goals for themselves, allowing them a sense of completion once some goal has been met.  The adventure hooks given for any scenario are not the only possible win conditions for that scenario.  If playing G1, for instance, the players might simply wish to rob the giants.  They may wish to subvert them, turning them from one evil master to their own uses.  They may merely need to get to the Hidden Chapel of Elder Weirdness in order to complete a magic item they wish to create.

Creating and offering hooks is a part of the creation of context for the game milieu.  Selecting from hooks, rejecting hooks, and reforging the information from hooks to meet some new goal are all part of the process of choice, and that lies entirely in the players' court.  When creating portions of the game milieu, the wise GM considers how those creations can be used, and what win conditions the players might accept to bring a session or group of sessions to a satisfying close, but the GM should not impose win conditions.

Yes, the GM is justified in believing that most players will accept win conditions, such as "survive", when placed into a situation where survival is threatened.  However, beyond such very broad goals - and sometimes, even then - players are surprising.  Exactly how long will you strive to reach the Grail in the collapsing temple, Indy?  Even though you will probably die if you wait too long?  What do you value more?

Recently, in the Comments section of this blog, a situation was discussed in which a GM had set up a campaign, wherein he imagined that his players would be attempting to stop the undead causing a plague. Two players had other ideas; they devised their own goals, and their own win conditions. If I was running the game, I would not have ended it for this reason.  I supply context, the players make choices, and then I adjudicate consequences.  Rinse, repeat.

Also in the Comments section, a situation came up where the players and GM discussed the goals of the characters prior to the campaign beginning, and the GM in question was unable to see how this limited the choices of the players.  The players and GM discussed and came up with an initial context, the players made choices as to how they wanted to approach it, and then....well, those initial choices delimited what choices could be made as the game went forward because they sharply differentiated between the context of the players and their characters. A condition of the campaign world was determined to be natural ahead of time by all involved, even though the condition of the campaign milieu was that many people believed it to be supernatural in origin. Characters wouldn't experience certain facets of the implied milieu because, having accepted the de facto hook, they no longer had the full range of options that would naturally exist in that milieu had they not.

In a bit of coincidence, both of these situations involved a plague, and both involved undead. In the first, two players decided to treat the plague as though it were a naturally occurring event (although the GM wanted them to go hunt vampires), and in the second, the players knew it was a natural event which they could profit from.  There was no point in investigating potential supernatural causes or cures, and no real decision making involved in determining whether or not sorcerous types were responsible (and hence no question about the ethics or advisability of using magic, burning purported witches at the stake, and so on).

In both cases, a predetermination of what the players were supposed to do took away what, to my mind, are vital elements of player choice.

Hooks present options.  They are not intended to be straight jackets.  Win conditions are ultimately chosen by the players, not the GM, and a group of players can operate in the same game even with very different win conditions.  Sometimes even opposed win conditions.  The players decide that, not the GM.

Supply players with context, from which the players can make choices, and then adjudicate the consequences of those choices.  It is beautiful in its simplicity.

Is there room in the boat?

I am thinking about creating a megadungeon for Labyrinth Lord (as well as one potentially for DCC), but with excellent products like Barrowmaze, Stonehell, and the Castle of the Mad Archmage available, is there still room in the boat?

What do you think? Is this an idea worth pursuing?

Silent Nightfall.....Epic? Final Session

On Tuesday, my players finished their sojourn to Silent Nightfall. In the end, caution was selected as the better part of valour, and all survived. Normally, I don't answer player questions about the backdrop of the game because, hey, if they want to know what foozling the begummmertz is going to do, they need to figure that out in game. In this case, we did a question period afterwards, and were able to determine that some, but not all, of the PCs could potentially have died.

At 8th level, DCC characters potentially have the resources to make light work out of Silent Nightfall, but the set-up and the unknown prevented the players from taking too much for granted. Between levels 2-4, probably up to 6, is ideal for this location. But you can make even 8th level characters sweat.

Failing Forward

This post comes about in response to some questions asked by YagamiFire, to wit:

What are you feelings regarding the current trend of making failure "attractive" as an option by virtue of "failing forward" in game design? To what extent does this impact the legitimacy of a challenge? If "the game must go on!" eliminates some failure scenarios because the players might find those scenarios undesirable...would you say that that undermines the concept of failure? Should failure from a PC stand-point be undesirable from a player standpoint with undesirable consequences that accompany it?

Good questions, and I will do my best to answer them.

It is my belief that the GM should not "guide" the action to a specific outcome. On the other hand, I do believe that part of good design is seeding a location with enough material to allow unexpected things to occur, for both good and ill. It is my firm belief that the early TSR designers, for instance, did not expect every last bent copper piece to be found, and thus seeded treasure in excess of what would usually be found so that, if a player happened to think to look inside the giant lizard's gullet, there was a chance of actually finding something.

I am well aware that there are some who imagine that every scrap of treasure in a published scenario is intended to be found, even though the only quote available on the subject, in Module B1, says exactly the opposite:  "Although monsters will inevitably make their presence known, treasures are usually not obvious. It is up to players to locate them by telling the DM how their characters will conduct any attempted search, and it is quite conceivable that they could totally miss seeing a treasure which is hidden or concealed. In fact, any good dungeon will have undiscovered treasures in areas that have been explored by the players, simply because it is impossible to expect that they will find every one of them." (p. 24).

There is nothing inherently wrong with offering the chance to "fail forward" in a scenario. Having unexpected good come out of failure can actually offer sweet moments in the game...but the word "unexpected" is key. Like the expectation that finding cool treasures spawned the idea that they would follow you around until you did discover them, the idea that some form of good could come out of failure ceases to become surprising when there is reason to expect that most failures will be anything but that...failures.

What JRRT called "eucatastrophe" - the feeling that, when catastrophe is assured, sudden hope changes everything - is a powerful feeling, but it only works when catastrophe seems inevitable. It doesn't work when failure is expected to be "failing forward".

From the previous series, it should be clear that I think that the maximum good, from a player's standpoint, comes from being able to play the associative game. The sort of metagaming that comes about from deciding what forms of failure are off the table works against this. It also works against the idea that the player's choices have any value - if the choices lead to failure, so be it. I have written in the past that the GM should never include a consequence for failure that he is unwilling to live with, and that is because players must be allowed to fail if they are ever to experience true success.

Balance of Power Part V: Function and Dysfunction

One of the oldest problems in philosophy is the question of evil. Why is there evil in the world? Especially if you believe that there is Someone in charge of the universe, what is the purpose of evil?  We are not just talking about wrongdoing here, but also illness, predation, the need to kill other things in order to survive, tragic accidents, and sheer bad luck. Surely an all-knowing and all-powerful Someone could arrange it so that these things simply do not happen.

We have all seen this thrust arise in gaming over the past decade - railing against PC death, save-or-die mechanics, campaign or adventure premises that seem to be one thing but turn out to be another, etc. In some game systems, characters can only die if the player chooses to put that option on the table. The rise of Challenge Rating (CR) type mechanics from 3e D&D onwards has led to an expectation that "challenges" be "balanced", where "balanced" all too often is taken to mean that the PCs should succeed without any undue loss. And, the advice now seems to go, you should consider fudging rolls or statistics to ensure that the expected outcome occurs. There is a sense that some believe that a saving throw doesn't represent a last chance at survival, but is rather something that should be repeated until an encounter is over or the character wins. Medusa doesn't simply turn you to stone - she slows you down to make the fight harder, but ultimately you triumph!

All of this boils down to the same philosophical problem: Why is there evil in the world?  With an all-powerful GM to look out for the PCs, why should the players ever fail?

None of us wants bad things to happen to ourselves, or to those we love, but at the same time most of us gave up watching programs or reading books where nothing bad ever happens long ago. Indeed, it is hard to imagine anyone making it through grade school without demanding more solid fare. I have found that few even make it past the age of 7 without developing some desire to have real problems occur in the fiction they are exposed to. This is not to say that they want Gollum to devour Bilbo Baggins, or Smaug to catch the hobbit burglar out, but the destruction of Lake Town rings true, as does the death of Thorin, Fili, and Kili seems right.

The GM has several jobs to do - provide a game milieu that makes sense to him, and that he is interested in running. Provide the players with context so that they can make choices, and determining the consequences of those choices. The players have a job to do - make choices within the context available, and role-play their characters.

If railing at the universe worked, then we would all rail at the universe whenever something bad happened. There are players who follow this principle in rpgs, because sometimes "railing at the universe" (through the agency of the GM) does work. A GM who changes rulings due to such railing does harm not only to his own enjoyment of the game, but also to the enjoyment of the other players. Rewarding railing such breaks the fourth wall, and effectively punishes players who accept the universe as it is.

We can imagine a GM who both encourages rules disputes, and then uses those disputes to split the party between "supporters" and "non-supporters" of the GM's position....but why would we do such a thing? First off, encouraging rules disputes perforce limits the associative game by forcing the players to think in terms of rules, rather than in terms of the fictional "reality" of the game milieu. Secondly, rules disputes automatically split the game participants, whether it is the intention of the GM or not. If they did not split the game participants, there would be no dispute.

I am not encouraging the GM to be a dictatorial monster - if you are that GM, your players are right to leave your table. What I am saying is that the GM has a responsibility to be the referee...to judge the rules as impartially as he is able to do. Trying to foist off that responsibility onto the players helps no one. Yes, discuss why rulings were made at a suitable remove from game play. No, do not encourage rules disputes during the game.

There is a reason that so many early games emphasized that the GM is always right, and it has nothing to do with stroking the ego of the GM. It is because - in the hands of a competent GM - that is the way traditional rpgs work best. And your goal, if you GM, should never be to be less than competent. A GM who uses "The GM is always right" to feed his own ego, or to make the game suck, isn't made better by encouraging rules disputes. A GM who is doing a good job, to the best of her ability, is at best hampered by rules disputes, and at worst hamstrung.

In order to be functional, any relationship must meet at least two criteria:

(1) Is power in the relationship shared fairly?

(2) Does each person in the relationship have the necessary rights needed to meet his or her responsibilities?

As to (1), either the players or the GM can end the game, but only all the players together have the ability to end a specific campaign, although the GM has the power to do so.  Without the GM's materials, the other players can create a continuation of sorts, by inventing (or buying) their own materials, and without the players, the GM can continue to use the same game milieu with others, if he can attract new players.

The GM has the majority of the say, because the GM does the majority of the work. If you expect someone to do the majority of the work, but have no increased share in power, then you are actually advocating a dysfunctional relationship. The world is full of people who advocate dysfunctional relationships. Usually, they advocate them for other people, while blithely ignoring their own advice, or they advocate them in their own favour.

IME, in most games there is no question about whether the GM or the players are going to walk, as long as condition (2) is met. Really, most people are able to talk out issues and make compromises, and most people are able to respect the work of the other people at the table.

As to (2), I have written - a lot - about what the GM needs. Let us turn for a moment and look at what the player needs. The player's primary responsibility is to play a character and make choices within the game milieu. That means that, unless there is some form of external compulsion involved that makes sense within the context of the game milieu - and even that should be used in very, very, very, extremely very limited amounts - the player gets to choose what the character does. Period. End of the sentence.

The GM does not get to tell the player what his character would do. The GM does not get to demand that the player approach the game with a specific goal or mood in mind. The GM does not get to demand that the players work together. All of these things fall outside the GM's purview. Only in the rare case, where an issue external to the game is being played out inside the campaign milieu, should the GM intervene. Demanding that the players choose a single goal that they work together is the GM version of being a rules lawyer, or demanding that the campaign milieu works in accordance to your expectations.

Now, if you are not interested in letting players play the associated game, or the GM play the dissociated game, you can come up with different forms of functional relationships. And, as a player, if you can find a GM who wants to run what you want, or if, as a GM, you can find even a single player, you should always run the game you want the way that you want.

But, for me, there is always a chance that bad things will happen in the game world because it is necessary for the associated game that it be so. It is also necessary for an interesting game. My job, as GM, is to provide interesting context and consequences that follow rationally from your choices. Your job, as player, is to play your character and decide what your character will do. I will respect your job, and I expect you to respect mine. We share this game. We share the power, based upon what we contribute and what our jobs are. We share the success or failure of each session. I am not out to screw you over, but it is my job to make sure you can make choices that do screw you over, just as it is my job to make sure that you can make choices that result in your coming out on top.

Gary Gygax talks about this in his Insidiae, and I quote here from pages 50-51:

It should be hammered home by now that the role of a game master differs significantly from that of a fiction author. The job of the game master does not involve revealing to the players the private thoughts or motivations of NPCs and monsters, nor will a good GM dictate what the players’ characters feel or how they ought to act – because he doesn’t know that. In general, a player should not be forced to explain his character’s actions, or to justify his actions to another player even if asked, unless the character’s normal demeanor has drastically changed, or the action threatens the entire party’s success or survival. Likewise, the denizens of a campaign world are known by their actions, their natures and private thoughts kept secret by the GM – unless learned by guile in play, ripped from them by magic or torture.

Also, no single antagonist or creature should become more important to the plot than the heroes. In other words, the game master should not make any NPC absolutely central to the unfolding story, because nothing controlled by the GM is more important than the development and advancement of the PCs through their interactive play. It is apparent, then, that the game master is far removed from being a “third person omniscient narrator”. Sure, he might be omniscient in regard to the details of his chosen milieu, but because he cannot know the future actions or thoughts of the PCs, he cannot be called a “story-teller” in the fullest sense.

Finding people who want to play, if you let them play their characters, and you don't punish the rest of the players by rewarding the weeds, has always been easy in my experience. Likewise, finding a GM if you respect the position, and if you don't act like a weed, has never been difficult. If you feel like you are coming to the game "cap in hand", from either side of the table, you might consider trying this yourself.

And that's the end of this series.

Unexpected This Is: Planned vs. Improvised

This post comes in response to a series of comments left by Vanguard, in which he is promoting a viewpoint that I simply do not believe to be accurate.  It is a conversation that we have had before, and the circularity of the argument may give "burning wheel" a whole new meaning.

I am not going to respond to the thrust that follows this amazing game, that both is a traditional role-playing game and yet is so unique that having played dozens of traditional role-playing games for decades doesn't give one any insight into general principles.  Nor am I going to discuss Vanguard's particular campaign milieu using that setting, where lack of prep doesn't cause inconsistencies, and the inconsistencies only deepen the mystery.

There is no point.

What I am going to do is disentangle two threads that Vanguard is trying to weave into one:  namely, planned vs. improvised play on the one hand, and shared world-building on the other.  And then, barring more input that isn't part of the same circular argument, I am done with being derailed for a bit.

Planned vs. Improvised Play

Every GM needs to improvise.  It is impossible to prep a game so completely that all contingencies are accounted for.  Even were you able to do so, the odds are good that your game would lose a spark of vitality by so doing.  Having to come up with answers in the moment is a spur to creativity.

BUT...If you put 100 prospective GMs in a room, and made them all create material on the fly, and also create material using some method of actual prep (brainstorming ideas, refining, and then actually going to the effort of writing them down), at least 95 of those GMs will have done better with prep, and at least 95% of the prepped work will be better than the aggregate of the on-the-fly work.  This assumes only a single scenario - the advantages of prepwork become more dominant as scenario compounds on scenario, and a campaign milieu is more fully realized.

This is because of several factors.

(1) Human brains are lazy. We have default answers. If we don't force them to do more, they will come up with the same or similar material repeatedly.

(2) Human brains are slow. Those easy answers come a lot quicker than harder (and often times, more interesting) answers.  The game is played in real time.  The odds are good that the easy answers will predominate.

(3) Human brains are forgetful.  When you bother to write out your material ahead of time, not only can you see connections that you might otherwise have missed, not only can you ensure that you add appropriate "footprints" to the elements that you add to a scenario, but you have something to refer to which can aid your memory of what has gone before.  You might not remember that the Duke of Duck had a limp; that doesn't mean that your players won't.

Factors (1) and (2) mean that on-the-fly material tends to be less inspired that the material that comes from careful prepwork.  Factor (3) means that logical inconsistencies also tend to arise.  In terms of The Walking Dead, these three factors together explain why the abilities of the walkers seems to change from episode to episode, why the characters seem unable to think of basic ideas that would occur to anyone in the same situation (ex., Rick is waiting for people in an area with many abandoned cars.  After hours of waiting, Rick runs out of gas, not having checked his fuel level earlier, and not having the good sense to siphon some gas from the many vehicles available.  He then decides to abandon the vehicle rather than siphon a small amount of gas from each of the other vehicles so that they can all continue to drive.), and why Rick is apparently unaware that a prison lies within a couple of hours of his house, despite his job as Sheriff.

Making it up as you go along gave us Hitchiker's Guide to the Galaxy, so if you have had only good experiences, or you think you are Douglas Adams, by all means continue doing what you are doing.  Most people, though, benefit from prepwork.

Of course, if the game milieu cannot be effectively prepped because it is going to change whenever the players make decisions, there is no point to doing the work.  Otherwise, it is the "work" in prepwork that causes most people who avoid it to avoid it.

I have been doing this for a long time, in lots of places, and with lots of people. I have run games with good, minimal, and no prep. I have participated in games run by others with good, minimal, and no prep. In well over 30 years of gaming, I have never seen no prep trump minimal prep, or minimal prep trump good prep.

If your mileage varies, go with your experience instead of mine.  If you want me to believe that not prepping is just as good as prepping, well, extraordinary claims require extraordinary proof, and I have yet to see any evidence at all.

Assertion is not evidence. It doesn't matter who is making the assertion. I know that I keep harping on not taking my advice if it differs from your experience, and this is why. Your mileage really may vary. Someone saying something - me, Vanguard, anyone - does not make it so.

Shared World-Building

Let us imagine three different GMs are prepping a game, and all three want their players' input on what to include.  I have been all three of these GMs are different points in my gaming career.

(1) The GM is devising a city to be the hub of his DCC campaign. Let us call this city Shanthopal. He asks his players, "What would you like to include in this city?" and takes notes, but does not guarantee anything. By so doing, the GM gains an idea of what his players would like, and the players gain some idea of what may be there, but no concrete idea of what is there.

(2) The GM is preparing a 3rd Edition D&D campaign, in a region he calls the Lakelands. He invites players to devise the details of their characters' backgrounds, including information about where they may have come from, but cautions the players that what they come up will be what their characters believe to be true. It will not necessarily be true.

(3) The GM is preparing a city for a 2nd Edition AD&D campaign. Following the advice in one of the 2e books, the GM invites the players to aid in creating the city, with the understanding that what they decide to be true will be true.

In all three cases, the players began with an initial sense of excitement and ownership.  In the third case, the players ended with a sense of meh as they were forced into positions where their associated game was damaged by the dissociated game they had played earlier.  In short, it is less fun to "discover" vampires in the sewers if you have placed them there and are actively manoeuvring the game into that "discovery" than if there happen to be vampires in the sewers and you discover them the hard way.

It is not difficult to extrapolate to the situation where the players also choose what will and will not be allowed into the game. One of the best series of sessions that I have run was using James Raggi's Death Frost Doom, along with some time travel, and the DCC ruleset. If the hard-hitting contents of these sessions had to be vetted prior to play, the scenario would not have been interesting at all. Sometimes we want material that pushes our comfort zone. For some of us, knowing that we don't get to decide what the world is, only how we respond to it, is the point of the game. The exploration of the game milieu is in turn an exploration of ourselves, of our strengths and limitations, of our ability to deal with the unexpected or the unwanted. It is not that we want to lose, but rather that we want to know that it is always possible to lose. It is not that we want to face extremely hard choices, but to know that those choices are always out there in the wings, and that the game can turn on a dime.

By these criteria, a game of Medieval Walking Dead, where you are never confronted by the decision to either help or ignore a sorcerer being stoned, not knowing whether the walkers are due to his magic (as the mob is screaming) or not is a pale sort of game indeed.  Unexpected and unintended consequences, and the need to make important choices with limited information, are, IMHO, part of a good role-playing game.

Examples (1) and (2) I would advocate. Example (3) I would not.

A Poisonous Meme

Read this quote:

For the purposes of the game, the world consists of what players encounter. Anything they do not is waiting to be discovered.
Finally, players are limited by what they can discover by both their in-game decisions and time (as in, we play once a week for four hours).
In either approach to GMing (planned vs. improvised), the kind of things they discover are necessarily constrained. For the former, it's restricted to whatever the GM put there. For the latter, it will be something within the scope of the shared fiction.
When you're playing the dissociative game, you have every option, sure. Once play begins? No, not if you're sticking to your notes. Likewise, when we sit down to build the situation, we have every option as well. Once we play, however, we are firmly within the prison of our own making. 

This might make why I broke down the players' and the GM's game slightly more clear. From the GM's (dissociative) perspective, the world consists of that which is prepped. From the players' (associative) perspective, the world consists of what they know, what they imagine they know, and what might be. Like a quantum wave-function, it only collapses into a single possibility when it is checked.

This is a primary function of the divide between the associative and the dissociative game, and one of the major things that the GM provides the players that they could not gain by merely making up the game milieu in turns while playing. Until the wave function collapses, the scope of the game from the associative side is wider than that from the dissociative side, and the game is so designed that some wave-functions can only collapse in the case of either (a) an event occurring within the game, or (b) the players partaking in the dissociative side of the game to at least the degree described in example (3), above.

Take our poor sorcerer being stoned as an example. The possibility is there from the associative side of the table until it occurs. That it does not occur does not take it off the table. Even its occurrence immediately opens up a new wave function that it may occur again under different circumstances. The GM need not even have considered this possibility for the wave function to exist; it exists merely because the players consider it. And players tend to put a lot of thought into the parts of the game milieu that they do not yet know, because their doing so increases the chances of their succeeding at their goals.

From this wave function arise additional wave functions. What if the mob turns on one or more PCs? What if the mob is wrong? What if the mob is right? No mob need appear in the GM's notes or in actual play for these wave functions to arise.

The associative game allows the game to be much, much larger than what is played at the table.

When you understand this, you also understand the limitations inherent in shared world-building. The more the players know about the world with certainty, the more wave functions collapse, and the more limited the associative game becomes. The poison in the the meme that "no limitation occurs" is that, if you accept it, you might find yourself trading away something without even being aware that it existed in the first place.

In Conclusion

Vanguard says

I should note, I'm deeply amused by one thing in this conversation. You're arguing here that on-the-fly GMing doesn't lead to a believable consistent world (which is not true), but criticiszing this approach as more akin to fiction writing. Isn't verisimilitude, which is what you're going for, part of that same equation? 

This is not actually true.

I am arguing that shared world-building of the "hard" kind - example (3) rather than examples (1) or (2) - damages the associative game.  I am further arguing that, if the world is being written or re-written in accordance to the players' choices or expectations in situ, then the players are no longer exploring a world; they are exploring the GM's evaluation of their choices.

The outcome/consequence of choices being based upon pre-existing conditions/context is a hallmark of role-playing games. The outcome of choices being based upon the needs of plot is a hallmark of fiction. If the writer does his prepwork, the fiction can be good fiction. Most writers who do not do their prepwork create bad fiction.

"Golly," said Sheriff Rick.  "How did I forget that Federal Penitentiary was here?"

While the GM can (and IMHO should) strive for verisimilitude as part of the creation of the context and consequences of choice, no one at the table is creating a story. The story is what comes after the game, when events are distilled into story form.  What is happening during the game is context-choice-consequence, repeatedly, creating a pattern of expanding and collapsing wave-functions. A story is static; the game is not.  Good prepwork allows for context and consequence that rationally and consistently follow the available choices. Bad or no prepwork leads to situations where everyone thinks sorcerers caused a plague, but no mobs are out looking to lynch sorcerers.

Balance of Power Part IV: Putting the Games Together

Some caveats upfront:

(1) If you have found yourself disagreeing with every post in the "Balance of Power" series so far, you probably won't agree with any of the posts to come.

(2) This series of posts is in reference to traditional role-playing games. In philosophy, there is something called the anthropic principle, which stems from the point that if we are here to observe the universe, the universe must exist in a condition that we can be here to observe it in. IOW, any universe observed by intelligent creatures must be a universe in which intelligent creatures can exist to observe it.

There are all kinds of versions of this principle, but one interpretation suggests that the universe we see is the result of our collective beliefs and expectations. I.e., if enough of us believe in unicorns, unicorns will not only exist, but they will always have existed.

A traditional role-playing game assumes that the players are dealing with a world whose basic properties are in some fashion set. In other words, the players, through the medium of their characters, explore that world, and their growing understanding of the principles by which that world works, including knowledge of peoples, places, etc., lead to increasing success within the game.  Just as we must adapt to the real world, while attempting to use our increased understanding to alter our environment to our benefit, the PCs adapt to the fictional milieu, and attempt to change it to their desires and/or take advantage of its properties.

If you are playing a game in which the PCs' beliefs and desires shape the world around them, or where the world remains formless except in the immediate field of view, where there effectively is no "real world" within the game because flux occurs either due to lack of prep or the inherent nature of a world where the presence or absence of a guard behind a door is based upon a character's convictions, while that game may be incredibly fun to you, it is not a traditional role-playing game.

Outside the context of this discussion, I don't mind if you want to call it a traditional role-playing game. Inside this discussion, I would appreciate the acceptance of terms for the purpose of discussion.  If it makes you feel any better, it would be no different than describing what I am referring to as "traditional role-playing games" as "apples" or "balmaranas"; the term is only used for clarity within the conversation.

(3) There is certainly no obligation to agree with me, and your comments are welcome. Assuming that I have thought through every objection you might have, that I am automatically right, or that I know exactly what all the repercussions of any idea might be will never be a prerequisite to commentary here!

Okay, then.

In the type of game that I am discussing, the players require that there be something to explore and interact with.  Although it is impossible (and, in fact, undesirable) to eliminate the information disparity between the GM and the players, the players seek to reduce the disparity in order to make more effective choices within the context of the milieu.

That information disparity also means that the players require the ability to trust the GM, and the GM requires trust from the players.

While much of the dissociative game is played alone, in the creation of materials, and while the GM may move timelines forward without players being present, the GM requires players to bring the whole to life. Without an associative element, supplied by the players, a traditional role-playing game is nothing more than preparation for the game. NPCs cannot supply this - there is no discovery in "exploring" that which you have already created, and if you have not yet created it, you are probably writing fiction rather than playing a game.

Writing fiction is a fun pastime; it is not a traditional role-playing game.

Likewise, there is a reason why solo play - for example, generating a dungeon using the tables in the DMG while you play through it yourself - falls rather flat. Without both sides at the table, the prospective player-GM is like Gollum, who thinks that great secrets must be hidden beneath the Misty Mountains, but discovers only darkness and a sort of half-existence gnawing old fish bones.