The idea of "monster generators" was discussed on MUD-DEV recently, and I've been thinking about the possibilities such a feature might provide. For those not familiar with Gauntlet, a "monster generator" (unsurprisingly) is a "thing" which generates monsters - the stronger the generator, the stronger the monsters it produces.

Within a mud setting, I envision the generators being scattered randomly around the landscape, with the appearance of a small settlement. Over time they would grow from village into town and eventually into a city - and send stronger and better armed monsters out into the surroundings. Players could fight these monsters (and indeed might be forced to, should the settlement grow too strong), but could also weaken (or even destroy) the settlement itself (although this would be significantly more difficult). When this occurs, I would suggest that a new settlement be created somewhere else in the world.

Each generator would be based around a specific race, although it might have different troop types - for example a dark elf settlement might also include driders and various summoned nasties, while an undead "settlement" could have several types of undead. Different generators would fight each other, although it could be possible to have alliances between generators of the same type. This could allow regular player-run cities to get caught in the middle of NPC wars, which would be rather fun IMO.

It's a fairly simple concept, but I think it could work quite well. I'd be interested to hear other peoples thoughts on the matter (as long as your thoughts are more than two words in length:P).

I played a mud based a lot like this not too long ago.

Gateways were scattered around the areas (always in preset positions unfortunately) and each gateway was labelled with the type of monster it generated.  So you might see a 'Blood Skeleton Hellgate' for instance.  For extra spice, four grades of each monster could come out, from particularly easy to quite hard.  At the top of the spectrum were 'unique' monsters, who had names and were impossibly difficult.

One of the weaknesses of this mud was that players were unable to destroy these gates.  I think when I saw they were a feature of the mud I had something very much like gauntlet in mind, so it seemed to me like destroying the generators should not only be possible but necesarry and not terribly hard.

Along those same lines, I would have liked the generators to, well, generate more.  Monsters trickled out of the things, which made it feel not much different from traditional repops.  Thinking of gauntlet again, I was expecting relatively easy monsters to pour out.  You would be confronted with great numbers but the monster themselves would be quite easy.  Naturally it would take some balancing to make such a thing difficult (wouldn't want stock rom monsters who miss anything 10 levels higher than them).  If one is going to do the generators at all though, I think this is a crucial point.  If a player can't experience the flow of monsters and see them popping out at the gate, it is really little better than a repop.

I like the idea that they spread too... perhaps the player run city pays hunters quite well to roam outside the city and control nearby populations.  If not enough hunters can be found or they don't properly execute their job, direct defense of the city might become a necessity.

i think this si the best idea i have seen posted in awhile, even the old forums didnt have that many good idea's but this ne is the best, some ideas for it.
eachnpc settlement has a growth rate this way make it from 0-100% and you can control how strong they get, like a %50 rate they would make a new room once every 3 days or so.
a settlement would be like 5 rooms, a village would be like 40 a town 100 and a city 500, this way it would give more control over it.

So imagine the scenario...your guild have found a bugbear village, and have built a defensive parimeter all around it to stop anyone getting in or out. When the bugbears try to get out, they are killed and skinned, and their hides turned into guild armour. Any players trying to tresspass in are warned off or killed outright. This alone would take a huge amount of manpower.

Several game weeks pass, and the number of bugbears coming out are getting fewer and fewer, until eventually no more come out. A scout is sent inside, where he discovers that the creatures have starved to death - they weren't able to get any food, so the whole settlement died.

Of course depending on the race, they might turn on each other, or rally for a final attack (ie every last living member of the settlement charges the weakest point of the defensive parimeter).

I suppose the guild could get around that as well, by dropping off food, although depending on the requirements of the race (eg needing human flesh) that might not be feasible. Just as the settlement would need to collect wood or stone for new buildings and metal for new weapons, so it would need food for more members (as well as maintaining the existing ones).

You could also have a natural migration of settlements - some races are particularly well known for not staying in the same location for long. You could "guard" the settlement with your guild, but what happens when they pack up and move on? You'd have to fight every last one of them, and you'd no longer get the exp for destroying the village. Other settlements might eventually just "die out" or from age or disease, or be force out by bad weather conditions.

Besides which, there would be a fair number of settlements/generators, so even if a few were guarded, it shouldn't prove too much of a problem.

Of course, this is dependant on the size of your playerbase. But, assuming a mid to small sized mud, the players would not be able to camp <i>all</i> of the settlements. Then they would of course lead to a natural progression of camping the ones that offered the best ratio of returns vs investment. You'd get natural selection leaving the the easiest controlled villages to run rampant, and the other settlements to be 'farmed'.

Yes of course though, you are right in the fact that if you build in decaying conditions for your settlements it will solve many of your problems. But players will work out a system, sooner or later, that will keep their mob farm operating at full potential.

The best solution to this problem of course would be to work out a way for the mobs to 'burst' out. To try and break out of their 'prison'. But I fear that this would eventually become a autonomous response, and become boring.

But what if the returns of a settlement were based it's size?  Guarding the settlement would stop it from growing in the short term (it would have nowhere to expand) and actually make it shrink in the long term (all of its resources would be going into "breaking out") and thus the other settlements would start offering better returns.  The larger the settlement the better the returns - and the more players required to guard it.  But by guarding it they'd end up making it shrink again, which means that if you wanted to "farm" a settlement, you'd need to do it over a short period of time before moving on.

Alastair raises a number of interesting points, and generally parallels my own thoughts on the matter of monster generation as a function of a kind of NPC society.  I'm sorry if this strays off the path of the gauntlet-style monster generator =).

I wrote a about my thoughts of individuals in the societal unit.  In that post, I described three needs: Survival, reproduction, and societal.  I believe those needs are what may solve a few of the problems and questions raised here, such as camping and the nature of specialization within the settlement.

So you've got a society of mobs sending raiders and scouts to the surrounding area.  One of them happens to drop some nice equipment, and a player takes notice...  Now players try to farm the settlement.  Hardly an orc raiding team can leave without being accosted and obliterated by those bloodthirsty players.  Those poor, defenseless orcs.

I think the solution to this problem lies with the fact that the settlement doesn't care about its raiders, and is entirely self-sustaining.  Because the settlement/generator keeps sending raiders to certain doom, players are rewarded for camping.  And because a settlement/generator can keep sending raiders without worrying about resources, players will be rewarded.

So we need to make it so the NPCs in our settlement understands that sending raiders out is a death trap, and cannot keep sending raiders without an influx in resources to create them.  We can accomplish the first through a survival need, that is the prospective raider does not wish to take that job because of the high mortality rate.  But that kind of forward-thinking is hard to code, so it's most likely that we'll accomplish this through a social need, where there are different weights on the different jobs that an NPC would choose to take.

If parties of raiders are constantly wiped out, the society could do two things:  Increase the amount and strengths of the raiders in hopes of wearing down the players, or stop sending raiders entirely.  I believe the first would be preferable if the society had bigger and badder types of enemies that it could send out.  But if the players still persist, or the society is too small, I think that we should decrease the societal need for a raider, thus decreasing the overall number of NPCs from that settlement that choose that job.  So a player would likely see a steady rise in the toughness of raiders, then a drop in the number of them.  This effectively removes the reward.

If that wasn't enough for you, we can try tackling the problem from a resource perspective.  Not alot of muds have a good resource model, however, so I don't think this solution would be as simple as simply giving weights to types of jobs for societal needs.  Nevertheless, it does provide a few interesting consequences.

Say that the society is set to produce 10 worker orcs, 20 fighter orcs, and 5 master-fighter orcs a week.  It costs the society a bit of food and iron/steel (for weapons and armor) to create the workers, a bit more to create the fighters, and lots to create the master-fighters.  When a player camps -- the quality of the equipment dropped should probably be proportionate to the amount of resources that went into creating the unit killed.  Master-fighters would naturally drop better equipment.  But if the settlement only has a finite amount of resources, they may not be able to produce the same quality of resources for their fighters because of farming.  Thus, the quality of the equipment goes down and the reward is removed.

But this also leads to strategic planning on the part of the players.  Do they take a note of that fledgling encampent that's next to a rich iron vein, possibly coming back to it later once it's started producing top-notch soldiers?  If the top-notch soldiers are too powerful for them, do they cut off the line of workers carting iron from the mines, and does the society react to this by placing guards in that area?

The more we can make these generators do, I think, the better =).

Personally, I wouldn't go as far as removing manually-built settlements entirely - at least capitols should need some "local" feel to them to add depth and flavour.

About the specialization according to society model (and race), another layer of immersiveness would be to add a faction framework on the whole game.
What are factions?
Preset groups of players and monsters, which can represent a combination of:
- nations
- races
- classes or equivalent
- alignment or equivalent
- trade
- religion etc...

A player's faction influence any kind of interactions with the game world: how much he's charged in shops, wether killing a mob will earn him respect or a bounty on his head, (and where), the type of quests he can get, the type of reactions in dialogue.

In our place, the setting involves strongly present nations, based on races. You could hence further imagine trans-national factions with dual loyalties.

Finally, this could also open up additional types of interaction for players, aside from "go destroy the such-and-such hive":
- As already suggested, your faction might rather require that you defend a place
- If the engine supports it, you could support a colony by assisting in the resource-gathering - for profit, of course.
- Capture-the-flag type missions where instead of destroying a hive you capture it for another faction. In that case, you could further envision that player groups have the option to change the type of hive to suit their faction's preferences - a change somewhat limited by the surrounding terrain type, of course, but you could for instance replace orc caverns by dwarves etc...

This all interests me in the context of the whole challenge of giving players more things to do in game terms than merely camp & slash - without the constraints of free-form or staff-driven RP.

As an aside, I love such discussions. Reminds me of the old and regretted Game Commandos </nostalgia>

By the community sending out better equipped, stronger mobiles, you will make players want to camp a settlement more. The cycle will only be shorter. The players will actively attack the settlement until the settlement increases it's defeneses, and thus the players will reap bigger rewards for their attacks. Once they've demolished one settlement, they move on.

This too will change the dynamics of the farming. Why would they camp a tougher group, when they know that by camping and raiding a weaker group the payoffs will be better in that they have less challenge, and maybe a little smaller amount of resources farmed.

You will make it so that one of nessecity must join a guild to be able to attack a settlement. Simply because the guilds are blocking the lower level settlements because of the payoff ratio. Leaving the stronger settlements to grow stronger.

Mind you, I'm not saying it's not a great Idea, I'm just playing the Devil's Advocate here.

Ashon, you raise a number of good points, and I'll attempt to extrapolate solutions and clarifications here. Please let me know if there are any holes in my logic =).

Two points: I believe that the cycle should have definable lows and highs -- the cycle wavelength will be as long as the amount of change necessary to incite the societal need for better defenses, times the amount of time it takes to gain the necessary resources. In my mind, this might take quite some time (RL weeks or months, depending on your game flow). But I do see how it could become a problem (all other things not considered) if your game flow is very fast.

The second point: Sending out better equipped, stronger mobiles is most certainly a good thing. How does one have NPC wars if the NPCs don't adapt to their opponents tactics? This problem would likely to be easily solved under a good system. If the settlement invests more resources but has the same return, I very much doubt that a sound AI system would keep sending better equipped raiders. By then the survival need should overtake the societal need, and nobody would take those jobs. Or the societal need for those jobs would fall with no success.

Strategic potential with such a system is wonderfully high. Imagine a king that knows he is hated by the surrounding nations. He might wage a proxy war by sending small units to attack the outlying settlements, so that when his enemy's army rolls through, they are weakened by the bolstered settlement's attacks.

Then we simply make it so it is not in the player's best interests to destroy a settlement. If a player depended on settlements for supplies and equipment of their own, they might wish to protect their favored settlements, or hometowns as KaVir noted. Players would likely have a vigor for defending their hometown, and quite possibly the settlements that their hometown depends on for trade and defense. If settlements are neutral and underpowered, I doubt that players would have any qualms with destroying them. But if they're destroying them so that their own nation gains more territory, we lend a great strategic credence to their continued survival.

I'm not so sure about this. Maybe if your system is designed poorly, this would happen. I can't see how one would necessarily have to join a guild unless your world is very tiny. Even then, stronger settlements could get so strong that they branch out and smash the lazy camping guilds. If growth were exponential, players would be forced to face the larger strongholds lest their own nations be overcome.

This may be a bit of a side-line, but it's an example of different ways in which ‘developing mobs’ can be put to use in a game. Both the examples below are up and running in our mud, although the second isn’t quite finished yet:

The first example is an abandoned spaceship, in our Future Dimension. It is drifting around helplessly in space, and at first when you enter, it seems pretty empty. But there is this Room - (I think it’s the Freezer) - with a large WARNING! sign on the door. If you open that door - (which no player can resist doing, because of the sign) - you let loose a swarm of giant mutant cockroaches (the initial number loading is related to your level). The minute they are released they start swarming out of the door and breeding like rabbits, and soon the entire ship is infested with them, unless they are disposed of.

They start out as pupae, and after a given time develop into mature cockroaches, which of course are much more powerful, and which in turn immediately start to breed more pupae. So getting rid of the pupae first makes things a bit easier. But to stop the bugs from breeding totally, you need to find and kill the Mother Queen.

It is all done with scripts, which also control the max number of cockroaches. This is important, because the bugs are confined to the spaceship, and if allowed to develop uncontrolled they would eventually crash the mud, when someone enters an overpopulated room. (This happened with a couple of early attempts at a similar concept).

It is all rather simple, the cockroaches do nothing but breed and fight, we haven't coded in any ecology, since it's not really needed for the purpose. But the second feature is quite a bit more advanced.

This is the possibility of breeding cattle, in the Old West Dimension. The cattle mate, give birth, mature, grow old and die in a set life cycle. When a bull and a cow meet, there is a set percentage chance of them mating and the cow becoming pregnant. When two bulls meet, they fight till death. Only mature animals can mate and conceive, of course. There are also some predators that attack the calves, to keep the population down. And if the cow is in the same room, she naturally defends her offspring. We have the same concept working for horses and sheep too.

There is also the option to fence your herd in, since the ranching is based on a grid zone. In this way you can control the population, make the animals breed faster and protect them from predators. The natural life cycle prevents overpopulation, and we are experimenting with having the animals die of starvation, if some lazy player confines his herd to a too small corral. This will be based on the sector type, the animals will be able to graze more times in a field room than in a mountain one, before they run out of food.

The animals can also be coded to roam about in search of food and water, moving more often when they are hungry and thirsty than not. (But I am still not sure if this is worth the effort, it would be a nice thing to have, but there is always the question whether it actually adds something to the Game, beside the satisfaction of having a working ecology. I think very few players will understand and appreciate the mechanics. And after all, it's for the players we create our stuff).

It’s done with a mixture of code and scripts. The basics are hardcoded; like the mating and the animals following each other so that they stick together in a herd. When one animal moves, the entire herd moves, (with the odd exception of a few strays, which also is coded in). The life cycle is done with simple scripts; after a given time each mob loads a new and older one, before purging itself, and at the last stage loads a corpse, which decays like any other corpse.

The main objective with this feature is for the players to breed the cattle, horses and sheep for their hides, wool, meat and milk, and they can also be herded to a cattle market and sold, with the aid of a shepherd dog. The shepherd dog is the only way to really control the herd, and it answers to some simple commands, like ‘herd <direction>’ ‘sit’, ‘stand’ ‘attack <mob-name>’ etc.

So far this is only up and running in our test Port, because we still have to add some handicraft skills, so the players can use the products they get from their ranching. (Tanning, carding, spinning, weaving, leather-making, tailoring etc.). The breeding part is all finished though, the cows hump to their heart’s delight and the population seems pretty stable.

No holes, just ideas and extremities....

If you set it as defined by Kavir, where they increase the strength after losing resources and to protect themselves, a guild could result in guerilla tatics to shorten the cycle.  If the needs are First and foremost food, and then defense.  by attacking the food, the need for defense will be increased...  the players are influencing the cycle.


I'm sorry, I did not mean to imply that stronger mobs were bad.  But imagine this scenario, a guild finds a settlement of goblins, with an ample source of coal and iron, and little food.  The guild boxes in the settlement, raids the only food supply and cuts off the food.  The goblins start raiding the players, sending out armed warriors.  After not defeating the players the mobs send out more 'feelers' looking for more food sources.  The Guild kills the feelers.  This time stronger armed goblins hunt the guild.  While the hunt is going on, the players 'drop' food in the settlement.  While the guild keeps the mobs from having a food resource, but still feeding the settlement, the settlement will be sending out stronger goblins.  or is that not the way people have been proposing the system?

Yes, but I'm looking at the settlements full of goblins or other mobiles, not nessicarily player settlements.  Not every settlement can be a hometown, or if they are, of importance, to protect.

I think a random starting location would add a great deal to the discovery and sense of exploration available to new players, although it might mitigate the effectiveness of any special newbie-friendly areas of the game. A player could be eased into the world gradually by having a hometown situated far from any border skirmishes. Or they could be swept into battle quite early, as the king's recruiters come to bolster the army's ranks for an upcoming campaign.

Personally, I believe that each of the choices a settlement faces should be "equal" in that all of them are quite desirable. Reactionary AI tends to weigh which desires are most immediatelly needed when coming to a decision. So in the absence of player (or other NPC) attacks, the system would more than likely choose to gather food and expand the settlement. I see these attacks as a means by which players (and other NPCs) can keep settlements from getting too large -- the settlements only get as large as the ability to provide food for the entire settlement as well as defend it from attack.

But the time that players spend influencing a settlement is, in my mind, only a fraction of the time that the settlement will be influenced by the needs of food-gathering, expansion, and war with other settlements. So I think that even a large camping effort, while dramatic at present, is not enough to make the settlement a "soldier factory", so to speak ;). And if the world is large enough -- for every minute a player is camping outside of a small settlement, another one is growing and gaining technology at a much faster rate due to the lack of competition. I hope that this will provide enough of an incentive for players to not stay in a place for too long.

Sounds like an interesting scenario =). What if the system were designed in such a way that food must be constantly supplied? A small food supply would diminish the amount of good warriors that a raiding team would have available to itself. While I don't see anything wrong with players helping to provide a settlement with food, it would get fairly tedious -- even impossible for anything less than a bot -- given the appetites of some of those raiders =).

But I do see your concerns in this area. Perhaps we could give more incentive for players to move around by giving technology gain (and thus equipment forging skill) a boost to those peaceful settlements that have the time and resources to research?

But what if those goblins have a nation all their own? One that can easily overwhelm any prolonged farming campaign. Then we might get into giving them the ability to forge alliances with other, more militarily capable nations in exchange for resources and research...

In addition, depending on the type of settlement, the AI's reaction to needs may be vastly different. Also, in terms of blockading, both the terrain and the AI behaviour may influence the effectivity of the blockade.

For instance, imagine a drow settlement which in fact grows underground. There are perhaps two or three access points, with only one clearly visible, the others hidden. Blockading is made easy here, but a paranoid AI model would send out all it has to break a blockade at the first sign of trouble.
On the contrary, a human settlement located on the surface in the middle of a vast plain might prove extremely difficult to blockade for players: if the settlement proper is eg on a 7x7 grid, you need players to blockade 15 squares, not a small feat. From that, depending on the society model, the AI may just defend its position (perhaps there can grow food inside the settlement's borders to hold out a longer siege), try to break the siege, or even call for help (definitely works very well in PvP environments).

This sound too much H&S for me. I wouldn't design such a system as solely an xp / equipment farm as the target.

Beyond that, having settlements matching existing playable races certainly adds an interesting twist to the notion of nations...

Or, if the blockade is too powerful, it might just block up the entrance entirely (or rig it with traps and wait for the PCs to get curious and come inside to investigate why no more drow are coming out). In this case, it could easily switch to a new access point - or even dig new ones. Such a settlement could prove extremely difficult to block off entirely, as it could easily break off into new settlements without the players even being aware of it. If the AI was smart enough, it might even leave a few drow as a decoy before moving off somewhere else. Imagine camping outside the Drow city only to suddenly receive shouts of help from the rest of your guild - because the drow had launched a counter-attack against your guild HQ!

In addition, a well-situated settlement could hold out even longer - for example imagine a settlement beside the ocean, which could continue to trade with other settlements even when under seige (or "camped out") by PCs. Or the dwarven settlement which existed both above and below ground, and could simply continue expanding downwards (although admittedly at a slower rate). Even a settlement beside a river might be able to sneak supplies in and out.

If you wanted to get really smart, the settlement might try sending out diplomats to bribe some of the players, or assassins to take out the guild leaders.

Then there's magic, of course. An "undead settlement" (ie a powerful necromancer/lich controlling an undead legion) wouldn't need food, although it wouldn't be able to expand much without new bodies. Other societies might have mages who could create magical portals for bringing in supplies - not only would the PCs need to block off the settlement (hardly an easy task), but they'd also need to magically ward it off. Now the settlement only has to strike at one location and hold off the PCs long enough to break the ward in order to get supplies through.

In addition, there are other problems associated with blocking off a settlement - famine has been mentioned already, but the other big problem is plague. A group of staving diseased goblins are unlikely to put up much of a fight, but when half the guild goes down with the plague it's going to be a serious problem.

If the settlements match with player races, you could also introduce bounty and reputation systems. A guild which blocked off a drow settlement could get a "bad rep" with the drow - and drow PCs might earn a reward for killing members of that guild. There's invariably going to be some particularly strong PC members of each race (and some weak members of each guild) which means that any guild which deliberately camps out around a settlement is going to start losing guild members to PK. If the guilds are race-based (and particularly if the settlements are sometimes player hometowns) such activities are quickly going to turn into PvP wars. Camping an entire settlement is going to be hard enough, but when that settlement is the home of another guild it's suddenly going to become a LOT more difficult.

Here's an idea for ya.

In SMAUG and others, ye can use those handy little obj_progs.

So, simply create an obj_prog that creates monsters at certain intervals.

Then allow the monsters to move around and occasionally drop a new portal object (perhaps a random type of portal, which could spawn bigger and better monsters). Monsters could spawn more portals, hell, even portals could spawn more portals (and maybe teleport them to a random location).

Wandering monsters can drop portals, the portals can spawn higher level monsters, who in turn continue dropping yet higher portals, and so forth and so on.

Could also spawn some non-portal objects, for the community aspect, since, in reality, a settlement doesn't really change the environment, it adds to it (at least in most medieval settings).

Now, add an option for players to be able to destroy PORTAL type items, and BAM, you got yourself a fairly easy to implement monster spawning system. In fact, some mobs could be peaceful, and just build their settlements without killing everyone - maybe even attack mobs from less friendly settlements.

I think I'll go work on this now, and see how it turns out.

That's pretty much exactly what I did with the gremlins in GodWars, back in 1995 - plus the top level gremlins would lay mana-restoration non-hatching "eggs", which encouraged players to farm them. However it's a very simplified approach, and not really the same as the settlement (or "monster generator") concept.

Of course that's not to say it's a bad idea, but my interest here concerns a central settlement AI which manipulates individuals members of that settlement in a believable social model, providing an opposition "race" rather than a mass of free-roaming individuals.

Desperate times breeds desperate measures.

So the PCs begin to farm the settlements, setting up blockades keeping the inhabitants at a certain power level. Seems like the perfect time for the settlement to call in the calvary. Wether its a goblin settlement that is being hounded calling in help from another goblin settlement that seemed to have gone unnoticed, or it could be a natural (or unnatural) alliance is formed with another race settlement.

In the case of kingdoms sending out settlers to expand the territory, should one of those settlements be hounded by guilds then the entire forces of the homeland are available to break up any camping.

Maybe some custom mob and obj progs could give a realistic simulation of a real community mind.

Some sort of control file, or a status file, could give monsters a general idea on whether or not to spawn new gates and so on.

I believe the reason for this thread isn't so much the implementation, as much as it's a brainstorm about the idea (or the post would have been made in another forum).

To have something relevant inhere:
Having read through the thread, I'm positive this is all quite possible. However, everyone seems to think of grid maps. We currently don't use grid maps at Cruel World, but it has crossed my mind. I can't really see how to avoid it, if anything like this should implemented. Any thoughts ?

*cough* not all participants in this thread code on Dikuratives...

They way I was shaping this in my mind was that the AI itself would be the grid server, and keep track of the changes of any building (barren ground->hut->house->ruin etc). That way, this wouldn't impact wether the rest of the MUD itself is using grid maps or not.