Event Driven Adventure Design

[komradebob]

Pre-Gen characters with set hooks and connections and goals?

A lot of this discussion of event driven design seems to center on a situation that as a GM, I used to hate a lot, namely that there was a conflict between the idea of player freedom, and plot of adventures. Since I was a WW fan, this one was especially trying. WW modules ( or whatever you want to call them...) were set up with the idea that a situation was present, plots were in motion, and various NPCs were madly scheming away.

Oh, yeah, and then there were some PCs.

They were sort of created however the players wished, and the GM was supposed to cobble together some sort of connection to the story line and involve the PCs. Very often the players, when presented with events would proceed to do the equivalent of open a flower shop. You know, basically take the assumed freedom and do what they felt like.

I hated it.

It often meant either scrapping the adventure as written ( which I often liked), or completely railroading the hell out of my players (which they didn't).

The thing that makes it tricky has to do with rpg tradition. With old-school dungeon crawl gamism, there is a certain amount of participationism. PCs can have a lot of leeway, so long as they go into the hole and kill critters and loot their stuff. Even if the players rebel a bit and go elsewhere, as long as they eventually follow the pattern, everything is fine.

Interestingly, I found part of the answer in LARP designs. Not the sort of generalized stuff built for on-going campaigns ( MET/WW, for example), but in one-shot designs. LARP designers working on the idea of one-off events often seem to involve players in a different way from TT RPG design. There is no assumption that PCs are "some guys in the vicinity of the action", but rather that they are important key players. Players get a lot of freedom to interpret their character, but not as much in initial design or plot hook creation. My impression of one-shot/event LARP design is that it is closer to improv theatre in some regard, with restrictions. To get back to WW, as an example, consider how different most WW modules would be if the players were assumed to take the role of one of the NPCs as a pregen, as opposed to creating their own character and shoving it kicking and screaming into the setting. Any of the "...By Night" or "Rage Across..." modules would have a very different feel, IMO.

So, in summary, yes, I think event based design is good. I think that it would be more successful with integrated characters, picked by the players, that have abiding connections with the adventure. I don't feel that wide-open character design with huge player freedom fits well with it. For that kind of feel, I think participationism should be left behind ( and I do consider participationism to be a positive term...), and a more "playing bass" style should be adopted.  So...

1) Pre-gen them. Let the players play important connected characters, with background and goals, allowing for interpretation, or...

2) Skip all the plot stuff and concentrate on whatever the players want. Build everything around those characters, developing enemies and allies as appropriate.

Both are good, but an admixture of the two is bad. Trying to have both is a recipe for frustration on everyone's part.

[contracycle]

I sympathise with your goals but thiunk your approach is unfortnate.

Lets look at "event driven" games as described.  this is comrpised of a series of events, causally linked to varying degrees, that occur in the game world.

the problem here is that they are still deatched from both players and characters.  Presumably, if neither players nor characters existed, these events would go ahead anyway.

Thus we kind of find ourselves back at the starting point: designing the environment in which play will occur, rather than the game that will be actually played by players.

Event driven design as described above would have a line of continuity from thet stats of game to the ernd of game, in the game world.  Bu the players may only be on, or cross, or in proximity, of that line from time to time.  Anytiem they are not on the line will require improvisation.

Hence, I have begun to think that this ideas is misplaced.  Instead we should be designing scenes that necessarily contain the characters.  A scene-based continuity would necessarily have PC's present at all points on the line.  consturcting events merely postpones the moment at which we have to develop an actual scene to do plot exposition.  I think we should start with those scenes, and design exposition directly.

[Kedamono]

Thanks Contracycle for your comments, but, your points... well, miss the mark.

Lets look at "event driven" games as described. this is comrpised of a series of events, causally linked to varying degrees, that occur in the game world.

the problem here is that they are still deatched from both players and characters. Presumably, if neither players nor characters existed, these events would go ahead anyway.

The problem is that your point above is true of any published adventure, regardless of how it is used. I buy module Q and now I have to convince my players to send their characters through it. It doesn't matter if module Q is linear, event driven, encounter bucket, situational, or a big collection of adventure seeds, the players will be detached from it.

Hence, I have begun to think that this ideas is misplaced. Instead we should be designing scenes that necessarily contain the characters. A scene-based continuity would necessarily have PC's present at all points on the line. consturcting events merely postpones the moment at which we have to develop an actual scene to do plot exposition. I think we should start with those scenes, and design exposition directly.

Two points: What you have described above is Linear Design. The old linear express, all aboard! Scene followed by scene, never branching or varying, the Plot must move on.

Secondly, how do I as a game manufacturer/designer, create an adventure that is tailored to you PCs? I can't. No one can.

Now, I do realize that your criticisms are based on a half assed description of a concept that is largely in my head, and I haven't stated certain assumptions about Event-Driven Adventures (EDA):
[list=1]
[*] They will not work for "open systems" because there is no way to design a game for radically different types of PC.
[*] I have been assuming that the RPG the EDA is designed for is focused on a certain theme and also focuses the PCs into certain types of characters.
[*] The EDA would continue this theme and maintain the same focus as the RPG.
[*] EDAs should have at least two "hook" events, events designed to bring the PCs on board and off to the next event.[/list:o]
So for a Pulp era RPG, I would create an EDA that involves the nefarious plans of the evil Doctor Fudd and his mutant Wabbits. The hook event would be Doctor Fudd siccing his wabbits on the PCs to take them out of the picture. (Or at least one of the PCs. In a Pulp era RPG, there will many Monks and Hams, but only one Doc Savage.)

[John Kim]

With old-school dungeon crawl gamism, there is a certain amount of participationism. PCs can have a lot of leeway, so long as they go into the hole and kill critters and loot their stuff. Even if the players rebel a bit and go elsewhere, as long as they eventually follow the pattern, everything is fine.

Interestingly, I found part of the answer in LARP designs. Not the sort of generalized stuff built for on-going campaigns ( MET/WW, for example), but in one-shot designs. LARP designers working on the idea of one-off events often seem to involve players in a different way from TT RPG design. There is no assumption that PCs are "some guys in the vicinity of the action", but rather that they are important key players. Players get a lot of freedom to interpret their character, but not as much in initial design or plot hook creation.

I think this isn't surprising, because in both a dungeon crawl and a LARP, the GM (or LARP organizer) has extremely little power to affect the situation.  A GM running a pre-written dungeon crawl is essentially a glorified accountant during play, reporting the map and room descriptions, and rolling dice for the monsters.  He has no control over the pace of the game.  The LARP organizer is similar.  There is no such thing as NPCs per se -- every character needs someone to play them.  So once the players head out, they run the game themselves.  

Hence, I have begun to think that this ideas is misplaced.  Instead we should be designing scenes that necessarily contain the characters.  A scene-based continuity would necessarily have PC's present at all points on the line.  consturcting events merely postpones the moment at which we have to develop an actual scene to do plot exposition.  I think we should start with those scenes, and design exposition directly.

I'm a little confused.  So you are suggesting to instead have a series of scenes which the PCs are necessarily present at?  Isn't this linear adventure plotting, a la Deadlands or Torg?  

Lets look at "event driven" games as described. this is comrpised of a series of events, causally linked to varying degrees, that occur in the game world.

the problem here is that they are still deatched from both players and characters. Presumably, if neither players nor characters existed, these events would go ahead anyway.

The problem is that your point above is true of any published adventure, regardless of how it is used. I buy module Q and now I have to convince my players to send their characters through it. It doesn't matter if module Q is linear, event driven, encounter bucket, situational, or a big collection of adventure seeds, the players will be detached from it.

Well, there are a few ways around this.  
1) The module can have a bunch of modular sections designed to be tailored for the PCs.  For example, "The Great Supervillain Contest" for Champions is designed as a sort of meta-adventure.  You take the module and plug in pre-existing villains from your campaign (i.e. the PCs' Hunteds).  

2) The module or series of modules may be designed to be integrated into the campaign.  So you as GM buy and read the module prior to the players creating PCs.  A typical example would be "San Angelo" for Champions.  The players create their characters as residents of that city, and the GM could then incorporate various plot hooks into the PCs.  

3) The adventure can force events onto the PCs without their choice.  A good example would be the "Prisoner of Zenda" adventure for Multiverser.  The PC is thrust into the center of the action by being mistaken for the prince, whom he looks exactly like.  The technique of "Bangs" is similar to this -- i.e. they are events that happen to the PCs.  You should check out http://www.geocities.com/doctorpeace/well.html">The Well of Souls, a HeroQuest adventure, for examples of the technique.  

Are you familiar with the term "Bangs"?  It's from Ron's Sorcerer RPG.  I use the term "prods" for roughly the same concept.  The term "prod" is in contrast to "hook".  A hook tries to draw a PC in a given direction.  A prod just tries to get them moving, but doesn't specify direction.

[contracycle]

The problem is that your point above is true of any published adventure, regardless of how it is used. I buy module Q and now I have to convince my players to send their characters through it. It doesn't matter if module Q is linear, event driven, encounter bucket, situational, or a big collection of adventure seeds, the players will be detached from it.

No, that is not a property of purchase, it is a property of design IMO.  Just becuase it is that way does not mean that it has to be that way.  You seem to have entirely missed the point I was trying to make.

Two points: What you have described above is Linear Design. The old linear express, all aboard! Scene followed by scene, never branching or varying, the Plot must move on.

I didn't say that, did I?  I'm describinf Actual Play, and asking you why your proposal designs things that are in fact irrelevant to the game we actually play.

If there is a scene for a movie that is never shown to an audience, then it cannot be said to be part of the story.  Similarly, everything, absolutely everything, of which the game is composed must appear in actual play.

I didn't rule out branching or options or anything along those lines; all I pointed out is that IMO design must concentrate on the actual activity of play, not on the background rationale that informs the GM.  The two are not the same.

Secondly, how do I as a game manufacturer/designer, create an adventure that is tailored to you PCs? I can't. No one can.

I see no reason for this to be true, as it happens.  It has to be said I can;t give you a recipe, but it seems to me that RPG's should be producable in much the way that a play is producable.

[shlo]

I aggree with contracycle saying we must concentrate on actual play.

But maybe we don't aim for the same thing when designing a scenario. Correct me if I'm wrong, but on the one hand Kedamono would like to write a "what if" scenario where many branches are possible -- giving a rather different story for the same plot -- and where PC would personalize the scenario, and on the other hand contracycle (and I) would prefer writing a scenario consisting of a serie of scenes, a regular story, and let the players -- and the GM too -- find their own ways and reasons to go from one scene to another, with the scenario giving some indications on what those ways and reasons could be.

While I understand the "what if" approach, I think it belongs to total improvisation and can hardly be prewritten, since we're talking of dozens of solutions. Note that the "scene to scene" just asks the players to be 100% present in the story but doesn't necessary tell neither on which side they have to be nor what their goals must be, and that's why I don't think it is linear. I see "scene to scene" like a more story (*) focused version of EDA as defined by Kedamono. In fact I would say "scene to scene" is EDA too, since I always used "scene to scene" and think of it as event designing. The difference with EDA as seen by Kedamono is that it relies a lot on timing and parallelism, so I suggest an appellation like Timed Event Adventure Design (TEDA) for example. Unless you disagree with my opinion that "scene to scene", or "event to event", is EDA too, of course.

I would be more than happy to see a functional (T)EDA method, provided it doesn't ask the designer to write dozens of branches, and I will keep on reading this thread and help Kedamono if I can. =)

Shlo.

(*) Here we have to consider the story as hollywood does: a story is a collection of strong scenes with whatever pretext to connect them. Don't skip those scenes!

[NN]

Id like a middle way between Kenamano's "what if" and "Shlos "send 'em from scene to scene", please.

How about a scenario with

-a plot that can tolerate a varying degress of 'success' by the players
-a scenario flow chart, with 2 to 4 broad outcomes per scene/encounter/event
-a few side plot-loops
-a choice of endgames - 3 or 4 'climaxes' depending on the branches took.

[Kedamono]

Id like a middle way between Kenamano's "what if" and "Shlos "send 'em from scene to scene", please.

How about a scenario with

-a plot that can tolerate a varying degress of 'success' by the players
-a scenario flow chart, with 2 to 4 broad outcomes per scene/encounter/event
-a few side plot-loops
-a choice of endgames - 3 or 4 'climaxes' depending on the branches took.

Hmm, this is closer to what's in my head than what I've written. :-)

Especially the flow chart, or in this case a Gantt chart.

Another thing is that people assume that the branches will go off in entirely different directions. In fact some events will still occur, even if the NPCs have to approach it from "branch B".

For instance: Black Bart knows that at the end of the month, the Wells Fargo office in Dodge has the government payroll locked up in their safe. He needs some dynamite and plans to get it from (steal) the Johnson's Surveying company. Then at midnight him and his gang will break into the Wells Fargo office and blow the safe.

Well the PCs thwart the attempt to get the dynamite from Johnson's warehouse, so Bart instead sends two his gang to the abandoned Dutch Creek mine to see if there is any explosives there.

The PCs ask about and find out that the Dutch Creek mine closed down a year ago, but old man Clinton bought a boatload of dynamite, that he never got to use.

The PCs arrive at the Dutch Creek mine after the hombres, Pecos Pete and Deadeye Dan. The two desperadoes hear the PCs and waylay them in the mine and using a stick, trap them inside. They escape, and stage their own ambush at the Wells Fargo office where Bart would have been if he had gotten the dynamite from Johnson's Surveying.

What would the events be?

Event 1: The PCs find out about Bart's intentions to rob Johnson's Surverying's Warehouse.

Event 2a: Black Bart and his boys break into the warehouse and make off with a crate of dynamite.

Event 2b: Black Bart sends Pecos Pete and Deadeye Dan to the Dutch Creek mine to get what's left of old man Clinton's stash of explosives.

Event 3a/b: The PCs can find out what Bart's goal is.

Event 4b: Pecos Pete and Deadeye Dan arrive at Dutch Creek mine and enter it to find the dynamite.

Event 5a/b: Black Bart breaks into the Wells Fargo office.

Event 5c: Black Bart decides to rob the coach that will be carrying the payroll...

6a/b: With the payroll in hand, Black Bart gets the Hell out of Dodge...

6c: Black Bart ambushes the payroll wagon in Blood Gulch...

So you can see that while there are several branches, you end up only with a couple threads to worry about.

[M. J. Young]

I keep trying to wrap my head around this, because I'm sure I've designed scenarios that are event driven, but I can't think of how to do it. So let me put forward an example or two, all Multiverser games I've run.

It's The Last Starfighter. The idea is to get the verser to take the place of hero Alex Rogan, but of course there's no guarantee. I drop him near the trailer park where the video game is, and make sure someone drops a quarter in the parking lot. Hopefully he takes the bait and uses the quarter to play the game, and then Centauri comes and offers him a job.

Maybe, though, he doesn't. In that case, Alex Rogan will play the game, and Centauri will take him, leaving the player character here. But we know that there's going to be a Zandozan coming eventually, looking to kill Alex Rogan, who comes back at some point. So there's still an adventure here, it's just a different adventure.

Or maybe the character leaves the trailer park altogether. That's fine; we'll just have another Starfighter video game wherever he goes, and it won't be Alex but someone else who becomes the starfighter if the verser doesn't, and again we'll have the Zandozan assassin, so we still involve the player.

Whether or not the player character goes to Rilos, Zur will attack the hanger. If he's there, that may be the end of the story. If he's not, it's either because he never left earth and is instead part of the unfolding of the other half of this story, or it's because he got a ride back and now has the assassin after him.

After that, Zur will breach the frontier. What that means depends on where Alex Rogan is and where the player character is, and that's got to be worked out.

I've also run If Looks Could Kill. In this, the character is somehow going to wind up mistaken for the spy and provided with briefing and equipment. He's going to see the cute girl headed for the casino. She's planning to kill the villain, because she's the only person who knows he's the villain; he's supposed to protect the guy, according to his briefing, and find out who is killing the people that the guy has already killed. The villain has a plan to poison all the heads of Europe so he can take over, and he'll follow his timetable pretty closely; but the girl is trying to destroy him. What happens depends on what the player character does in the middle of this.

It has just clicked in my mind. A few months back I did an article, http://www.gamingoutpost.com/GL/index.cfm?action=ShowProduct&publisherid=81290&ProductID=81549">Game Ideas Unlimited: Antagonists, in response to questions about designing an adventure. This was the first of several articles on adventure design, but it seems to be very much on point here.

What I recommended there was to create your villain and work out his plan, complete with a few contingencies that this particular villain would undoubtedly prepare. Then play the villain. He's behind the scenes, most of the time; the player characters don't know he exists, even, initially. But somehow (and this also must be planned) there is a moment in the villain's plans where the players will become aware of something. It may be that he needs something he will have to take from under their noses. It may be that something went wrong somewhere. Whatever it is, the players are alerted to the existence of the plot, and they can begin chasing down the details. The referee's job from there is to play the antagonist as well as he can, with an eye to carrying out the initial plan as closely as can be managed.

That strikes me as an event-driven sort of scenario. Is that what you're after?

--M. J. Young

[contracycle]

What would the events be?

Event 1: The PCs find out about Bart's intentions to rob Johnson's Surverying's Warehouse.

Event 2a: Black Bart and his boys break into the warehouse and make off with a crate of dynamite.

Event 2b: Black Bart sends Pecos Pete and Deadeye Dan to the Dutch Creek mine to get what's left of old man Clinton's stash of explosives.

I'm going to have another go at describing the difference I see.

Look at these three events: how much Actual Play do they constitute?

Even 1, maybe 10 minutes.  Event 2 doesn't happen in actual play at all - it happens Off Screen.  So in structuring these events we have not really structured the actual play we are going to enter in any meaningful sense.  In these two events, we have only one SCENE: "the characters find out that Black Bart has nefarious plans".

Event 4b: Pecos Pete and Deadeye Dan arrive at Dutch Creek mine and enter it to find the dynamite.

Event 5a/b: Black Bart breaks into the Wells Fargo office.

Event 5c: Black Bart decides to rob the coach that will be carrying the payroll...

6a/b: With the payroll in hand, Black Bart gets the Hell out of Dodge...

6c: Black Bart ambushes the payroll wagon in Blood Gulch...

None of these are scenes at all.  They all describe the whirring of invisible greased clockwork, and not one of them happens on stage, unless it is presented as a cutaway.  So maybe the GM's makes 5 minute speech about the raid or whetever, or the survivors drift into town.

So now with all these causally linked events, including several branches, we are therefore able to prepare for about 20 minutes of Actual Play.  Is this actually worth the effort?  The one scene we have here is "the PC's discover that Black Bart has carried out his plan."

My approach would be something more like:
Scene 1: the characters learn that Bart has nefarious plans
Scene 2: the characters interact with some of the locals
Scene 3: the characters learn that Bart has carried out his plan
Scene 4: etc...

From this perspective, the underlying logic of causality driving the conflict is rendered secondary to the actual business of play, the presentation of action right here at the table.  My experience is that even with branched structures like that proposed, I still find myself feeling under-prepared, because in fact my preparation has been for something other than actual play.

Equally, the scenes are structurally related, not causally related.  So in fact I can prepare two sets of stuff for Scene 3 depending on which plan Bart has actually implemented.  If causal event 5a/b occurred, I would prepare a scene in which the characters are woken innthe morning by shouts and screams as bodies are discovered in the WF office.  If 6c occurred, I prepare a scene in which a lone straggler staggers into town suffering from his wounds, or something.

Anyway, I hope that adds to the discussion.

[NN]

MJ, that site seems to be knackered.

I think that a possible trap is to make the villains plan too linear and too specific (villain does A to do B to do C to win ultimate goal D).

Instead it would be better presented as: Villain wants to achieve D, and heres an action 'tree' of how theyll try to get there.

Also, dont make D all-or-nothing.


nick

[clehrich]

My experience with something resembling EDA was very negative.  Let me explain why, and you can tell me whether this bears any relation to what you have in mind.

The campaign was a space thing, with the PCs running a small ship for one important federation/trade-empire place.

The GM scripted up a kind of timeline of NPC events, with a certain amount of flexibility in it, but mostly it was a fixed timeline.  Basically the various NPC factions were up to their various things, and barring major interference from the PCs they would get up to those things.  He then asked himself, "Okay, so how do I get the PCs to run across this stuff?"  And he dropped hooks.

So then we made up characters and off we went.  First up, we were ordered to a particular location, so running into the plot was certain.  But we weren't told, "Okay, now you have to do this next thing," but rather figured it out ourselves.

Now what happened was that when it got right down to it, there really was a "right answer" for where to go after a particular scenario, because if we went to the wrong places, given travel times between systems, we could end up right off the plot timeline and miss all the important events.  So the GM started dropping broad hints about where was the right place to go next.  Sometimes there were two or three different possibilities, because there were enough things going on with the timeline that if we made any of these few choices, we'd be sure to run smack into some big events.

Mostly we necessarily stuck to the order of scenarios he'd imagined.  But the further problem was, some of our characters (and some of our players) just didn't find the whole big plot interesting or important at all.  We just didn't care.  And nothing in our character designs had ensured that we ought to care -- we were given free rein there.

Now this meant that at any given branch-point, a number of us (me perhaps especially, because I'd put a great deal of effort -- encouraged by the GM -- into designing a complicated character with conflicting goals) would feel that the character wanted to go one way, the basic logic of the situation another (which conflict right there was just fine), and the GM was making clear that there was a third direction that was "correct."  Or alternatively, if the character wanted to go one way and the situation seemed to want to go another, the GM's timeline background made me say, as a player, "Well, clearly my character has to come up with a reason to do X, because that's what's necessary for the campaign to continue as scripted."  So a lot of the campaign for me became "Let's figure out what the GM wants us to do."  Which sucked, as I'm sure you can imagine.

In the end, we had the big climax, but an awful lot wasn't really clear: we'd spent much of the time figuring out what the GM wanted, not what was really going on.  So then the GM delivered a little lecture in which he explained his whole big plot, isn't that cool, blah blah.

For me, this was the #1 drive behind trying to design a system in which the GM simply doesn't have much control even if he wants it.  The GM here was trying to tell a story using the PCs, but at the same time he wanted the PCs to discover and tell the story.  This is called the Impossible Thing Before Breakfast, and believe me, it stank.

This is NOT railroading, as I read it, because the GM didn't actually railroad us through scenarios.  But that background timeline ticking away there required that (1) we care greatly about his plot, even though we couldn't know anything about it at the outset and thus might well not care; (2) we correctly interpret the hints and such dropped by the GM so we didn't go completely afield; and (3) we not as a group decide, "Nope, we're going off to do our own thing," because it was clear this would have pissed off the GM who, after all, had done a lot of work on that timeline.

So my question is this:

How do you propose to get around this potential difficulty?  I'm not convinced that there is no way around, but this is for me the obvious danger.

[Marco]

How do you propose to get around this potential difficulty?  I'm not convinced that there is no way around, but this is for me the obvious danger.

Well, I wasn't there but I know what works for me (and I wouldn't consider the games I play in exactly EDD but they still involve complex situations that impact the characters and which the characters impact--mostly situations that are not of the player's devising).

What I do when I run games like this is:

1.Give the players an idea of what characters are appropriate to the adventure. Set out tone, general direction, and some character concepts. You can see examples of this in my game writeups in Actual Play.

2. When I get the characters, I define and refine the situation to them. This is done with a little back and forth between me and the player (or with the GM if I'm the player).

3. Make sure that the circumstances of the game interest and engage the players and characters (i.e. whatever's going on is relevant to both of them).

At that point you should have a framework that just about any common design will work with--even "linear" assuming that you aren't *wedded* to it (i.e. your notes can say Act 1, Act 2, Act 3 and run with some reasonable reliability so long as you understand that that's just an organizing principle and not a mandate to force the characters around).

As I've posted here (earlier in this thread), there are techniques on creating "fault tolerant" situations (I linked to the essay) that make the possibilities of the PC's disengating from the plot minimal and recoverable (outside of pure PC disinterest which is a problem exterior to the scenario design at this point).

FWIW, having done these, I consider myself the 'author of the scenario' (Edited to add: and therefore, in a real sense 'the story' of the game) and having played in them, I consider myself the 'protagonist of the story.'

-Marco

[John Kim]

What would the events be?

Event 1: The PCs find out about Bart's intentions to rob Johnson's Surverying's Warehouse.

Event 2a: Black Bart and his boys break into the warehouse and make off with a crate of dynamite.

Event 2b: Black Bart sends Pecos Pete and Deadeye Dan to the Dutch Creek mine to get what's left of old man Clinton's stash of explosives.

Look at these three events: how much Actual Play do they constitute?

Even 1, maybe 10 minutes.  Event 2 doesn't happen in actual play at all - it happens Off Screen.  So in structuring these events we have not really structured the actual play we are going to enter in any meaningful sense.  In these two events, we have only one SCENE: "the characters find out that Black Bart has nefarious plans".

OK, I'm guessing here, but I think John (aka Kedamono) intended Event #2 to potentially be Actual Play.  i.e. The PCs find out about Bart's intent.  Mostly likely, they'll go and try to stop him.  So then there is a fight (or something) at the warehouse.  If the PCs don't go or if they go but fail, Bart gets the explosives, then 2a happened.  If the PCs succeed and he doesn't get the explosives, then 2b happens instead -- i.e. Bart gets explosives by another means.  

To me personally, it feels rather constrained because the limited branchings mean that all choices still get funneled into roughly the same later outcomes.  i.e. It doesn't matter if the PCs stop Bart at the warehouse, because he'll just get explosives by another means.  They can eventually make a difference (i.e. forcing Bart to try for the stagecoach rather than the bank vault itself), but it still feels limiting to me.  

A different but related technique is Clue Trees from Millenium's End.  There are limits to ASCII art, but I've put the clue tree from the sample adventure in the main ME book at http://www.darkshire.net/~jhkim/rpg/theory/plot/cluetree.html   A clue tree does not consist of NPC actions.  Rather, it is made of pieces of information -- which are primarily PC actions.  For example, "search the crash site" might be one node in the clue tree.  This information may lead to several other clues.  So each node is definitely related to the PCs, but not every node will definitely be reached by the PCs.  

John's tree seems more like a timetable.  A traditional timetable is literally a linear set of events in time which will happen if the PCs do not interfere.  This doesn't mean that the PCs are expected to or force to not interfere -- quite the opposite.  The PCs are expected to break the timetable and stop the villain's plans.  The timetable just tells them where to find the villain.  One problem with the timetable technique, though, is that once the PCs do interfere with the villain's plans, the original timetable is liable to become invalid.  From that point, the GM has to improvise.  So these represent two different ways of creating a tree structure out of a traditional, linear  structure:
Linear plot  -->  Millenium's End Clue Tree
Timetable -->  Kedamono's event-based tree

This is NOT railroading, as I read it, because the GM didn't actually railroad us through scenarios.  But that background timeline ticking away there required that (1) we care greatly about his plot, even though we couldn't know anything about it at the outset and thus might well not care; (2) we correctly interpret the hints and such dropped by the GM so we didn't go completely afield; and (3) we not as a group decide, "Nope, we're going off to do our own thing," because it was clear this would have pissed off the GM who, after all, had done a lot of work on that timeline.

So my question is this:

How do you propose to get around this potential difficulty?  I'm not convinced that there is no way around, but this is for me the obvious danger.

The solution to #2 seems pretty simple: make it obvious.  If you want the players to stay on the tree, the challenges should be in dealing with the events, not in finding the events in the first place.  Some games deal with this by GM aggressive scene framing (as suggested by _Feng Shui_ and several later games) -- i.e. rather than giving players the option to dither, you instead just declare they are at the next event.  A lighter-handed way of doing this is to give enough information.  

#1 and #3 seem like railroading to me -- just a more passive aggressive variety.  i.e. "Do whatever you like, but I'll be hurt if you do the wrong thing."  I would think that the thing to do with for these is to get player buy-in early on.  If it's a published module (as John is discussing, I think), then perhaps the players can pick the module.  Alternately, I earlier suggested several ways that a module can more directly involve the PCs.  What I wrote was,

1) The module can have a bunch of modular sections designed to be tailored for the PCs.  For example, "The Great Supervillain Contest" for Champions is designed as a sort of meta-adventure.  You take the module and plug in pre-existing villains from your campaign (i.e. the PCs' Hunteds).  

2) The module or series of modules may be designed to be integrated into the campaign.  So you as GM buy and read the module prior to the players creating PCs.  A typical example would be "San Angelo" for Champions.  The players create their characters as residents of that city, and the GM could then incorporate various plot hooks into the PCs.  

3) The adventure can force events onto the PCs without their choice.  A good example would be the "Prisoner of Zenda" adventure for Multiverser.  The PC is thrust into the center of the action by being mistaken for the prince, whom he looks exactly like.  The technique of "Bangs" is similar to this -- i.e. they are events that happen to the PCs.  You should check out http://www.geocities.com/doctorpeace/well.html">The Well of Souls, a HeroQuest adventure, for examples of the technique.

[Joe J Prince]

I see two differing viewpoints emerging from his idea

Determinism vs Free(player)will

I think an event based rpg campaign can work very well, as long as it addresses the premise - and Shared Imagined Space - and as long as there is a workable Social Contract between participants
Basically - according to the group's desires.

However, most players value freewill very highly and dislike the plot overpowering their character.
So the events need to be vaguely defined until the scenes incorporating them occur - then they can be connected to the PCs and premise.

A good way to go with events would be to examine Joseph Campbell's 'The Hero With A Thousand Faces' and try and build from the mythical exemplars he describes - or examine 'Story!' the screenwriting template (though I know less about this).

JJ - IMHO as always