And the NPC lived happily ever after...

Submit your own resources!

Rodney (OldRod) Burns   (Apr 10, 2008 at 18:16 GMT)

I agree with you - someone should have done it by now.

The biggest problem is (I think) getting developers to not cow down to the player base. Lets say you put in some unique content, and the first player to get to the content does the quest and then it's no longer available... now the rest of the player base will be up in arms and say "but it's not fair - I was at work when the patch came in, so I had no shot at the quest!", or dozens of variations of that complaint.

How do you get your developers to turn a deaf ear to those complaints and press ahead with a truly dynamic, unique world? I'm sure once players got used to the idea, it would be popular, but it would take a metric ton more content than what we see in current games to pull off effectively.

Probably the most realistic way to do this is to have the standard repeatable content for a lot of things, but have a good share of unique stuff too. That way players who play more get more opportunities to do unique "stuff", but people who are casual players still have plenty of stuff to do as well - it just won't be unique.

It's a tough job... but I hope someone does it :)

Nicholas Bogart   (Apr 10, 2008 at 18:22 GMT)

Amen.... I know I am still new to the game development world and I hack and slash around and get happy when fxshapereplicator works right(starts grumbling about groundcover still not working) but I know when I have the skills/time/money/resources what I want to see.

Your right... you should be able to walk up to an NPC and have to actually prove yourself by your deeds and not how many quests you ground or mobs you slaughtered to get in his good graces.

Anyways I like what you have written and its a cliffhanger for sure.

Ted Southard   (Apr 10, 2008 at 20:12 GMT)

Quote:

---

How do you get your developers to turn a deaf ear to those complaints and press ahead with a truly dynamic, unique world?

---

Well, that's the beauty of it: The system has to be designed like that from day one, so it is set up to always generate content. The fact is, "patching" a game isn't a viable way to get a dynamic environment. It has to push zone state to the users when they log in, and push state changes to users as well. Of course, this means two things (which are contradictory on their face):

1) You have a lot more data to dish out (and take).
2) The client side application needs to "sip" data.

Now, I'm not going so far in my game that anyone can do anything, or even looking to do something like networked terrain deformation (though it tempts me, but I have no good use for it). But there are certain things that I do want to do that means pushing state changes and data to the user at a higher rate than your average MMO.

One way to get around it is to get these changes across in two phases:

1) When the user logs in, the data and state changes for the zone that their character is in needs to be pushed. Depending on the data load we're talking about, maybe we can do all the adjacent zones as well, just in case the player's character is sitting ten feet from a zone border.

2) Related to the above, you have to have smart-loading of content. State changes can be pushed in mmotime (realtime + db access + lag) to players as long as they have the data to reflect that. The data they need to reflect that should already be on their harddrive from phase 1. Smart-loading of content also means that when players get near an adjacent zone (or maybe continuously- again, depending on how much data is being flung around), they start getting the data changes.

The above sounds pretty severe, but that's only if you make it severe. That is, the house that gets reduced to rubble doesn't necessarily need to be seen crumbling if you cloak the swapping of models with a good particle effect. So you have two models instead of three or more, or a really nice animation of the model falling apart, but that also consumes bandwidth as that gets pushed out to your players. But it needs to be designed in early on.

Something else I didn't touch on but should have is the sense that quests have to be fast-food: Served up and gulped down quick. Honestly, when was the last time a human could slay ten wolves without a .50 machine gun? When was the last time that someone going from Point A to Point B in a jungle was able to just jog the whole way, even if they never encountered anything?

I tried to relate some travel stories to people at the convention, but I don't think I got my point across as much as I would have liked. The point being that travel, in and of itself, is a hard business, and the less travelled that road, the slower going it's going to be.

When was the last time a ship in WoW hit a storm and became disabled, giving those with crafting skills an opportunity to fix it in order to be on their way? When was the last time a player found themselves face to face with a MOB that could kill them easily, but rather than attack, stayed in the cover of jungle and stalked them until it got tired of the game? When was the last time you found your player in an environment without the proper clothes or equipment, and lost hitpoints or endurance as a result?

It doesn't happen- except in real life, and that's what pen and paper RPG's had at first tried to emulate: The rigors of hard travel and high adventure, but in a world different than our own. And this is where MMO's fail spectacularly. You see a Jaguar? Go kill it. If I had followed that rule in Costa Rica last year, I'd have been tender-friggin-vittles. You don't fight a 200-400lb animal. Instead, you walk past it and stay calm (outside at least), and you don't run when it stalks you. I feel silly bringing that story up, it's not bragging because that cat could have killed me (and I was armed- it was the jungle), but the point is the complete disconnect from the kind of adventure that people want, yet don't know it. Indiana Jones, 007, any of those characters have obstacles they run away from, circumvent, or simply fail to overcome either directly or indirectly. If MMO's want to portray adventure correctly, then it needs to present the player with threats that it can get past without fighting, or run from, but in a way that makes travel from Point A to Point B a quest in itself, and the end goal (killing that Ice Wizard) just the icing on a very rich cake.

Jonathon Stevens   (Apr 11, 2008 at 01:24 GMT)

Thanks for the compliment Ted and your sponsorship was most appreciated! I want to let you know that I whole HEARTADLY agree with your points in this post, however to play devil's advocate you have to remember one key thing (that I happen to have said during my opening on day 1): Indies are in it for the love of the game, AAA is in it for the love of the money provided by the game.

In reality, the cost to benefit ratio must lie largly on the benefit side for AAA to implement it. The investors gave them money to build a game so they could make money, not so people would enjoy the game. I understand it, and if I had to do it to get my game funded, I'd walk away from the funding (no really.. i'm serious damnit!)

Having said that, if I got the money to build the game the way I DIDN'T want to, but that money led to me being able to build _my_ game, then I'd certainly sell out on the first one. Everyone's a whore for the right amount of money yes?

Ted Southard   (Apr 11, 2008 at 01:40 GMT)

I'd whore out for the money, yeah. But like you, I wouldn't do it for my game. As it is, I think if I can belt this tech out before I look for a team, and before anyone with money sees the project, then maybe I can have my cake and eat it too. Of course, the odds of being struck by lightning are far higher, so only the small eggs go in that basket.

Like I said, the AAA's have risk to worry about. We, as indies, don't. But one thing that struck me (especially on these forums lately) was that a lot of indies want to make the same kind of MMO as the AAA's do, with one risk -maybe- which is the classic AAA MMO formula. AAA's have to worry because failure can take down a company. Indies don't worry because, well... Let's face it, for most of us there's nothing to take down!

Anyways, I can rant all night about this, but I'm staring at a diagrahm that needs to get converted to TorqueScript for that cliffhanger conclusion in a couple of weeks ;)

J.C. Smith   (Apr 11, 2008 at 06:32 GMT)

Interesting discussion. We took a different route with The Repopulation by creating templates and then using the templates to generate the content for quests, to populate the dungeons and to generate the bosses special abilities. So rather than designing a quest for a particular NPC you design it for an NPCs personality trait, their mood, their profession. Each step can have four NPC requirements or item requirements which can also be generated by different templates. It winds up allowing for a lot more content to be created with a lot less manpower, which is important for indie developers.

One of the more interesting things you can do with that you give each NPC a personality, which never changes. But you also give them a profession, mood, cause and reason, which can change. In most cases the profession won't but we left the ability to have it change. So let's say you take a quest which asks for npc1 to be a Lonely, Single Male, and npc2 needs to be a Shy, Single, Female. That criteria could fit a large number of NPCs in the game world. If the quest was a matchmaker quest, at the end of that quest the relationships of those two NPCs could be changed, as could their mood, cause and effect. That would then open those NPCs up for different quest opportunities and the cause and effect fields would allow NPCs to speak about things that have happened to them in the past. It gives the players an opportunity to affect the game world. Using another example, if you have a quest to kill an NPC, and that NPC has any relationships, their relations could have their mood, cause and effect changed. Later another player could receive a quest to kill you, by his Mother whose mood is Vindictive, the reason because her Son Died and the cause is the player name. We've had this type of system running on our Alpha Server for The Repopulation for about five months now. Our quests are generated using these templates, and it makes the world feel a bit more dynamic. Since the quests are using templates with limited criteria the performance hit isn't that bad so long as you limit the total number of dynamic quests at a time (we allow 5) and you limit the check frequencies.

We employed similar systems to dungeons, using prebuilt layouts, but different texture sets and using DIF objects to block certain hallways randomly to change the overall layouts. There are spawn points throughout a dungeon instance. You can only enter if you have a quest for that instance and it uses the quest criteria to help determine how to populate it. Bosses, normal npcs, traps are all calculated with a random element. So when you enter into an instance you never really know what to expect. The dungeon layout, trap positions, NPC spawn locations, boss locations, object locations change every time. Same goes for boss special abilities. Their special abilities are determined by race and tier, and generated randomly when an instance is entered, which we hope will make guilds adjust to tactics on the fly during raids rather than just executing ready built templates.

The overall gain is that we can ship a game with 15 dungeon layouts, and about 100 quest templates, and have enough content to keep players busy for a long time. By contrast a big studio created MMOG might ship with 10,000 unique quests (though 99% of them are simple one step kill, or delivery quests) and 25 dungeons, with maybe 50 unique bosses. Once that content has been gone through though, users are just using spoiler sites for everything because it never changes and they become bored. Using a generated system you could focus instead of generating 100 solid quest templates (we currently have 34 in the alpha) and letting the generation system mix them up enough to keep the quests interesting. The big taboo with generated content has been that it is too simplistic, but it doesn't have to be. We have templates in alpha currently that are 10 step quests, and the ability to chain quests together to make them even more complex. Using a system such as this you can limit the overall manpower required to create a game while increasing the replayability.

In the end though I do believe that even large houses who are funded are trying to make good games. The employees aren't all just whoring themselves for money, a lot of them are in it because they want a chance to help shape and form a genre that they enjoy. Unfortunately for those studios its a lot easier to just copy what the last guy did rather than trying to be innovative. For the publisher they are doing what they think will be a safe bet. For the employees they are stuck integrating their employers vision of a money making scheme.
Edited on Apr 11, 2008 07:17 GMT

Ted Southard   (Apr 11, 2008 at 14:19 GMT)

@JC: I think we're on a closer page than you might realize, especially with quest templates and chaining the quests together. There are a few things that I'm doing differently, mainly to make the travelling to certain places a bit more difficult, but when reading what you have above, it's very similar at the base level, tuned of course to the kind of gameworld you guys are looking to have.

The NPC's in my game also have personality-style AI, though it's somewhat different than yours due to the conversation component that we're using, as well as some of the other system differences. Awesome to hear that I'm not the only one doing this!

Rodney (OldRod) Burns   (Apr 11, 2008 at 16:50 GMT)

@J.C: How do you keep track of all this data? Let's say there are 100 NPCs. So each player would have to keep track of each NPCs mood, profession, cause and reason?

Or conversely would each NPC have to keep track of all players? Mood could fluctuate by player after all. Curious as to the "nuts and bolts" of this system. It sounds fascinating, and it's something I've tinkered with, but never come up with a good solution.

Ted Southard   (Apr 11, 2008 at 16:57 GMT)

@Rodney: You can have either NPC's track PC's, or PC's track NPC's, but do it in the database, and you're not really using up that much RAM except what gets used for the time that it's needed. You can even do something like a "base mood" for an NPC with vectors associated with PC's so that if an NPC is mad and sees a PC that is a good friend, they don't act as mad to that PC, but can still address the PC they are mad at in an aggravated fashion.

Don Hogan   (Apr 11, 2008 at 18:39 GMT)

I really have to agree with the idea that dynamic content can be better than the prevailing school of thought. As an environment artist, I hadn't thought of using it with NPCs as you have described; but for a long time I've held fast that several simple sets of templates can be combined and leveraged to dynamically create a wide variety of content that appears far more complex.

It's especially intriguing to think about what JC is saying regarding the mixing of fixed and dynamic attributes for a given scenario / NPC. One example that came to my mind straight away was the idea that that technique could be applied to vendors as well. I.e. the vendor always buys reagents to make a product when its inventory is below max; but depending on what kind of supplier / customer you've been he might react in a myriad of ways - freebies, discounts, special (re)quests for unique products.

Fascinating stuff, I look forward to the next blog!

J.C. Smith   (Apr 12, 2008 at 03:04 GMT)

@Rodney: As Ted aludes to its all done in the database. Basically what we do is every 5 minutes there is a check. If your player has a full complement of 5 dynamic quests, it skips it. If they have 4 quests active then they have a 25% chance of the 5th, if they have 3, then a 50% chance, if they have 2 then a 75% chance and if they have one or none they will always get a new quest generated. It defaults to a Menial quest, which is any type of simple task, delivery style quests, checking up on NPCs, etc. It then scans through their skills set and based on that it has a chance to upgrade the mission type to a type of combat quest (Low Level Combat, Medium Armorcrafting, Underworld, Epic, etc) and after a few random calculations a missiontype is determined. It then searches the database for that missiontype. Currently we have 34 random quest templates in the alpha, so you receive a relatively small number of results back from that poll. I guess the easiest way to describe this is to paste in an example of a quest:




So if this is the quest that is chosen then the basic generation criteria for that quest is:

dialog.numnpcs = 3
dialog.randomnpc3profession = "Metallurgist"
dialog.randomquest = True
dialog.randomnpc1sex = "Male"
dialog.randomnpc2sex = "Female"

That tells is that this quest has to generate 3 random NPCs... NPC1 must be male, NPC2 must be female, NPC3 must be a Metallurgist. That search criteria could also have specified any of those NPCs to need a specific personality, mood, relationship, race, npcresponsible, type, race and a flag to see if they are locked into any other dynamic quests in the world. The reason for that last one is if that character has a mood change as the result of another quest, you would want them to exclude them from any other generated quest which could change their mood. The quest could also include information on if this was a Local, Regional, or Distant quest, or specify NPCs needing to be from a particular zone.

The mail information in the main dialog is information that is emailed to a player when they are given the quest. The Repopulation is a Science Fiction game, so everyone has a PDA and they instantly receive email. In our game you don't seek out dynamic quests from NPCs, instead they email you job offers (quests). Players can accept or deny a quest, and quests can contain time limits. If a player accepts a quest then it attaches that Dynamic quest dialog to their characters dynamic quest list and the quest remains active. If they deny the quest then you could either delete it (as we do now) or put it back into a shufflecard for another player (which could be done to minimize calculations if that ever became an issue). This is the text that is emailed to you. In that particular quest template what happens is an admirer (any male in game) wants to get the attention of a particular female (any woman in game). The first step is to find a Calibrium Diamond, which is a rare mined item. You will notice the track = "Calibrite Ore". If this quest is being actively tracked and you come into range of any Calibrite Ore clusters then it will automatically begin tracking them on your radar. This is an optional addition to any quest, you can track any NPC, object or point of interest and have the overhead map include a directional pointer to it. The quest dialog determines if this tracking is available or not. Once players find a Diamond, it changes their tracking to the admiring male. When they speak to him, he sends them to the Metallurgist that was generated for this quest, who will craft them a ring for a fee. They bring this back to the male who asks them to deliver it and the message to the female. When speaking to her, you have some diplomacy options, which we put in a few quests to raise that skill, and also an opportunity to say something shady which raises your underworld faction (and can open up underworld quests) at the cost of faction with the Commonfolk.

Using this type of a system you can design quests as though they were unique quests, but randomize the criteria. When you think of how many quests in MMOGs are simple kill, delivery, craft quests, just with a tiny bit of randomization you can duplicate those quests. Sure you lose the unique story (though you can always make the generated quests specific to a location if you want to add some backstory to what is going on there) but in a recent poll of WOW players they found that the vast majority of players simply click through without reading the backstory at all, they simply accept the quest, and look at what they need to do and then they do it.

@Ted: I look forward to seeing what you guys have in store. I've been hoping to see more randomization in these games for years. My day job is that I run spoiler sites for these games (since EQ'Lizer on the original Everquest, 10 years ago). When I first began doing these sites, I thought that it wouldn't be a long lived venture as eventually companies would make things random enough that spoilers would be useless. Thankfully for me, they never did. As a player though, I hate the fact that you can play a game like WOW or EQ 2 and get an in-game map and spoiler information for everything while you play. That guilds can look up boss mob strats and once one guild has done it everyone else just uses that same strat. Randomization and player created content can get rid of all of those things, if only more developers would choose to use it. Though I guess, perhaps its a good thing that more haven't as it makes those games stand out. Which is good for us I guess with a shipping date later this year.
Edited on Apr 12, 2008 04:37 GMT

Rodney (OldRod) Burns   (Apr 14, 2008 at 14:37 GMT)

What I've been contemplating is more of an advanced faction system. Rather than having a faction standing with a group of NPCs, I am planning to have a standing with an individual NPC. This would effect how that NPC reacts with a given player and what quests could be offered. Combine that with an NPC mood system like Ted described, that provides some variance in standard interaction, along with some random content the way JC describes, and I think someone could have a very strong system.

Ross \"Boaz\" Leland   (Apr 17, 2008 at 16:19 GMT)

There is a game already doing this. It just got into the MMO scene and is called Mortal Online. You do have a very valid point though. Not being the first one to kill an NPC that took forever to get to in quests is a bit discouraging, especially if you watch someone kill it right before you. A great way to get around this would be to make it a daily quest to slay the mob and only someone with the quest can spawn the mob.

If someone has already said this, sorry it's a bit early for me to read all the detailed replies. o_o;;

Ted Southard   (Apr 17, 2008 at 18:21 GMT)

I had a whole post written up about Mortal Online, but it got eaten when I hit the submit button. Since I have to catch a cab to the airport soon, I'll just say that after reading up on it at their website, what they seem to be doing is pushing out rare unique quests that can change the world. That's not the aim of my blog's rant though.

What I was trying to say was that all quests needs to be dynamic and unique, and even beyond that point, they need to be able to modify themselves and branch, just like real stories do. It looks like JC has made some great strides towards that end, and his project proves that it's not about the player doing some epic quest and saving the whole world, but players going around doing little things to slightly change that small corner of it a bit at a time.

While a lot of MMO's have good features and are great games, they all sort of miss that point and think that every player needs to be the only hero, when that's simply not the case at all. It's just that that's what developers push out because they think the players want it, and the players think that that's what the games are supposed to be about because that's what the developers push out. That's not the case for every MMO, but it's the case for a lot of them.
Edited on Apr 17, 2008 18:30 GMT

You must be a member and be logged in to either append comments or rate this resource.

Home | Products | Solutions | Downloads | Support | Search | About Us
Copyright © GarageGames.com 2000 - 2008 All Rights Reserved