Good story vs. connected situations (long but easy)

Started by
9 comments, last by Wavinator 22 years, 11 months ago
Got an odd idea based on semi-random encounters & would like to know what you think. It''s a bit long, but I''ve formatted to be easy to read... Planned Random Encounters You overlay a grid on a map. You then fill every cell on the map with numbers that represent possible encounters. The encounters are simply probabilities that get tested every time you move inside the cell. The grid numbers are based on player actions, who the player is, and events going on in the game world. Example: Red Faction is attacking settlements in a desert. They have a base in the NW corner of the map. The closer you get to the NW corner, the higher the numbers in the grid are. This means that you''re more likely to run into Red Faction the closer you get to their home base. Also, every so often when you stop at a settlement, you have a chance to encounter an attack in progress. Now, say you take out the NW base. All the numbers in the surrounding area for Red Faction drop. From the player''s point of view, Red Faction was crushed and scattered. They may, over time, concentrate in another area, however, and start up again. Reactive: Making This Like an RPG Story? It''d be great story (maybe) if Red Faction leaders decided to hunt you down. Now, you could actually do AI that had the leaders moving on the map, chasing the player, getting into near misses, ambushes, etc. But you''re not guaranteed to have a great story that way. RF leaders may miss much of the time. Or repetition may set in. Worse, the AI may not be up to handling all it needs to do to logically pursue. What about simply spawning a wide number of set piece situations according to the cell the player is in? They''d involve a cast of familiar characters in different situations, be based mostly on player actions, be spaced out using timers to control for pacing (which you could ramp up or down), and lead to a very non-linear yet seemingly connected story. Boot Hill The Red Faction leaders want a showdown at high noon. They''ve kidnapped the player''s best buddy, and are holding him hostage near the water tower. All the player knows is that his buddy is missing (a specific situation we triggered). When he goes into town for supplies, the RF leaders will be waitin''! Ever notice all the coincidence that happens when a GM games an RPG or when a writer writes a story? People are in the right places at the right time, and everything happens for a good reason. We want to imitate this, but keep the all important sense of player freedom. How It Might Work Maybe the answer to this is simple:
  1. Create a huge bunch of possible situations. The situations correspond to things you can trigger (based on player actions, or NPC actions). They involve NPCs, actions the NPCs will take (could be a script), what the NPCs have, and where the situation takes place. In our example, we''ve spawned a few NPCs with sniper rifles, a head honcho waiting in a saloon, some neutral NPCs that act as roadmaps, and the buddy bound and gagged in the empty water tower.
  2. Associate the situations with grid cells on the map. We only spawn things that make sense with the location. No Red Faction spawns, for example, if we know every Red Faction guy is dead, or that we''re in a place they could never be.
  3. Expose some (but not all) of the information to the player: For instance, if RF''s numbers are high in the town grid cell, a friendly NPC could warn the player. "Careful! I seen a lot of RF hombres lurkin'' about. Heard tell they was looking for you!"
  4. Every time the player triggers a probability, load the map with stuff then and only then (even if this is as simple as seeing a enemy hanging out). This cuts down on AI and managing what should be where.
  5. Give the player devices, allies, and information about what''s on the map (a spyglass to see who''s in the next cell, or warnings like the flickering of a campfire in the distance...)
  6. Be sure to let players get into and out of situations at will (like retreat, hide, evade) and make sure they can most of the time encounter situations only when prepared (meaning you might only hit them with a flood only if they had a boat, or whatever). This theoretically allows you to control game balance precisely, so that every situation is fine tuned and fun.
  7. Finally: Relate the situations to the past history of the game (say, like decrementing all the numbers for Red Faction after the base attack). You can keep a history of what''s gone on in the past, and this allows you to keep the player interested because you can vary situations. You can even key in on player responses and bias situations toward what the player likes, assuming you can somehow determine this.
Why This Is Better Than A Story Humans make patterns and draw conclusions. We can''t help but connect the dots. Given a random grouping of patterns, we''ll automatically see a shape if the information before us isn''t too confused and scattered. The same principle might apply here. By having connected, semi-random, set piece encounters whose probabilities we can tweak, we gain both connected, seemingly logical events, a high degree of player choice, and replayability. We create the illusion of cause and effect, and if we present the info to the player (thru NPC conversation, skill use, etc.) in a good way, they may get the feeling of an overall story without the restrictions of little or no choice stories traditionally impose. What do you think? -------------------- Just waiting for the mothership...
--------------------Just waiting for the mothership...
Advertisement
I like it (no big surprise there). Modeling subjective reality instead of objective reality.

A sort of influence map for encounters, hmm?

The only suggestion I would make is to make the ''encounter squares'' fairly large to reduce the chance of repetition.

I like the ideea. It seems so simple, yet so powerfull.

It could be used not only for player encounters, but for NPC-NPC encounters, which in turn could advance the game plot in dynamic ways. If at tile x/y, there is a 5%probability of Red Faction encounter and a 20% for White Faction encounter, there is a 1% probability of a Red-White encounter. If such an encounter occurs, and it ends up in a fight, then at tile x/y there will appear a say 50% probability of an aftermath encounter. If player goes around there, he may find dead Reds and Whites, signs of battle, etc.
I honestly don''t know if you''ve been taking inspiration from some of my proposals or if your thought processes have been independently converging towards the same framework for a game, but I agree.

_______________________________
"To understand the horse you'll find that you're going to be working on yourself. The horse will give you the answers and he will question you to see if you are sure or not."
- Ray Hunt, in Think Harmony With Horses
ALU - SHRDLU - WORDNET - CYC - SWALE - AM - CD - J.M. - K.S. | CAA - BCHA - AQHA - APHA - R.H. - T.D. | 395 - SPS - GORDIE - SCMA - R.M. - G.R. - V.C. - C.F.
Bishop, sorry that I have to say this, but I think you are giving yourself too much honour there

The only usefull thing you said was "I agree". The rest was, well, just you pampering yourself
*dwarfsoft considers rounding up the Game Design Corner regulars [Wav, JSwing, Naz etc] and plugging their brains into a Game Design Supercomputer to steal all of their marvelous ideas*

hmmm..

*dwarfsoft reconsiders and decides to just keep stealing their ideas from the Game Design Corner*

All I can say is that I am amazed at you Wav... It will be a sad day that I don''t turn up in the ''Corner'' and not see a new Wav post...

-Chris Bennett of Dwarfsoft - The future of RPGs
Thanks to all the goblins over in our little Game Design Corner niche
Thx for the great feedback guys!

quote:Original post by JSwing

I like it (no big surprise there). Modeling subjective reality instead of objective reality.


Haha, now I have a fancy name to call it. "The SRMS, subjective reality modeling system..."

quote:
A sort of influence map for encounters, hmm?


Yes!

quote:
The only suggestion I would make is to make the ''encounter squares'' fairly large to reduce the chance of repetition.


Right, or keep track of past encounters (which needs to be done anyway) and avoid repetition that way. Good advice, though.







--------------------
Just waiting for the mothership...
--------------------Just waiting for the mothership...
quote:Original post by Diodor

I like the ideea. It seems so simple, yet so powerfull.

It could be used not only for player encounters, but for NPC-NPC encounters, which in turn could advance the game plot in dynamic ways. If at tile x/y, there is a 5%probability of Red Faction encounter and a 20% for White Faction encounter, there is a 1% probability of a Red-White encounter. If such an encounter occurs, and it ends up in a fight, then at tile x/y there will appear a say 50% probability of an aftermath encounter. If player goes around there, he may find dead Reds and Whites, signs of battle, etc.


Diodor, great idea!!! I *did not* think to apply this same philosophy to NPCs except in the most generic sense.

But now I see you could apply this to key NPC characters as well as large groups. This would create sort of a semi-random history, and all you''re really doing is keeping track of squares that have been influenced and a table of possible events.

Awesome!

--------------------
Just waiting for the mothership...
--------------------Just waiting for the mothership...
quote:Original post by bishop_pass

I honestly don''t know if you''ve been taking inspiration from some of my proposals or if your thought processes have been independently converging towards the same framework for a game, but I agree.



Hey, you know what they say about great minds, right? Now, as long as we don''t start arguing about implementation details, I think we''re home safe.

quote:Original post by Ronin_54

Bishop, sorry that I have to say this, but I think you are giving yourself too much honour there


Jeeez, am I gonna have to seperate you two? Don''t make me turn this car around, now!


quote:Original post by dwarfsoft

*dwarfsoft considers rounding up the Game Design Corner regulars [Wav, JSwing, Naz etc] and plugging their brains into a Game Design Supercomputer to steal all of their marvelous ideas*

hmmm..

*dwarfsoft reconsiders and decides to just keep stealing their ideas from the Game Design Corner*


Yes, but just think how awesome we''d be networked!!! Except I''d be constantly whining about how most RPGs are medieval, which would ''cause Naz to blow a gasket, and the whole thing to explode! :D

quote:
All I can say is that I am amazed at you Wav... It will be a sad day that I don''t turn up in the ''Corner'' and not see a new Wav post...


Gee, thanks! I''m feelin'' all warm & fuzzy. :D And here I was thinking, what with every other post being from that d*mn "Wavinator" that I was starting to wear out my welcome.





--------------------
Just waiting for the mothership...
--------------------Just waiting for the mothership...
Implementation suggestion: use fuzzy state machines and an influence map.

Wherever there is a central location/base/hotspot - like the Red faction base in your example - attach a fuzzy state machine. This represents not only the planning of the Red Faction command, but the behavior of the Red Faction as a whole.

This AI will determine what sorts of Red Faction activity goes on in the area. That is, it sets the Red Faction encounter table.

Then use an influence map to measure the likelyhood of a Red Faction encounter in the area.

Use one fuzzySM for each important faction and a common influence map.

This topic is closed to new replies.

Advertisement