[pve] dynamic Monster activity, implementation

[sushibandit]

Hello all, I've been thinking about the way PvE may be implemented in Mortal. The typical linear progressively-more-difficult-areas is an annoyingly simple system, one I hope the developers will choose to toss aside like so much garbage. It makes quite a bit of sense that monsters will become stronger the further away from player/NPC strongholds they are, but the way that most games statically implement this 'tiering' system breaks immersion quite a bit. I'll be explaining how and why a dynamic system should be implemented in this post.

Please read the entire post before replying, and please comment constructively only (constructive criticism is welcome, idiotic criticism is not). Discussion is also welcome and encouraged.

~~~

Here is my 2c on how to implement a dynamic, player-driven (important), exploration-rewarding PvE 'difficulty tiering' system. The following are a few rules that I believe should exist in a dynamic environment game:

1) Player activity shall affect Monster activity.
2) Friendly NPC activity shall affect Monster activity.
3) Monsters shall intelligently react to external stimulus.

Now to go into a bit more detail on each rule. Bear with me, this is the long part :P

1) Player activity shall affect Monster activity.

No brainer here, when players have been killing monsters in an area for long periods of time, the monsters should start realizing that fewer of them come back from (insert activity here, foraging, etc..) a certain area. This should either encourage the monsters to move to another area or organize a strong scouting party to investigate the area in question (see #3).

This obviously relies on intelligent monster AI, as a nomadic tribe of weak but agile lizardment would most likely just move camp while a slower, more thick skinned orc tribe would probably end up sending out a combat party.

We'll also need to assume that monsters are in an area for a specific purpose, such as harvesting edible plants, gathering wood, hunting small creatures, etc.. This assumption can lead to dynamic and interesting power struggles between groups of players (want to build a fortress near an excellent hunting ground) and the Enemy NPC presence that wants to maintain their source of sustenance.

2) Friendly NPC activity shall affect Monster activity.

See the first paragraph of #1 above. The same reasoning applies to Friendly NPC activity. Obviously this activity should be limited in scale so as to allow the player base to have a larger impact on Monster activity.

For example, 'NPC Town A' is a trade-based town, and relies heavily on the import/export business. So they pay guards to patrol the major routes into/out of the town during certain hours or for certain groups of merchants. It makes very little sense for Monsters to randomly spawn on the road when 'NPC Town A' has been regularly patrolling the route.

This also leads to the possibility of (see #3 below) intelligent monster planning, ie. forming a raiding party to attack a specific caravan, or to wipe out the patrol at a certain point so that a Monster group can safely move through the area, etc...

3) Monsters shall intelligently react to external stimulus.

This ties into both #1 and #2 above, but merits its own 'rule' in order to break it down into more easily explained pieces.

With a dynamically designed system, we will need to assume several things about Monsters in order to set up a method for detailing their behavior.

For the purposes of this post, we will be assuming the following:

Monsters need food
Monsters are community-based
Monsters are not 'stupid'
Monsters are aware of and operate in relation to other Monsters

Of course, you cannot apply a single set of assumptions to every monster in game. For the most part, the assumptions above are intended to apply to various weaker, traditionally 'tribal' Monsters, such as Orcs, Goblins, Lizardmen, Spiders, various Insects, many species of Animals, etc...

For the purposes of this post, we will for the most part ignore the traditionally 'stronger' solo monsters like Giants, Trolls, Vampires, Werewolves, Dragons, etc...

Now, to apply these assumptions to Monster behavior via example.

A medium sized village of Goblins have camped in a mountain valley. They do so because of proximity to cover (forest), food (forest, mountain stream), and safety from other Monsters that live higher in the mountains.

A group of Players are working their way into the mountains in search of a place to construct a base camp for their guild, who plans on constructing a large fortress in the mountains. This group of Players chooses a spot fairly close to the Goblin camp. While scouting the area after constructing their camp, a Player finds one of the Goblins fishing, and follows the Goblin back to the Goblin camp, then returns to tell his fellow Players. As they are few in number, they decide to drive out the Goblins through guerrilla warfare.

After several Goblins are found dead with arrow and knife wounds, the Goblins decide to pack up and move further into the mountains, as they are unable to determine (anti-tracking skill? camp-concealing skill?) the location or number of the humans killing them.

And so on. In this example I've applied player activity to Monster behavior patterns based on the assumptions above. If it had been an Orc camp, perhaps the Orcs would have tried harder to track down the Players killing them, or simply decided not to move, or decided to start patrolling the area, etc... The Monster reaction should be intelligent based on circumstances.

Switching topic slightly, but still applying to Rule #3.

In regards to overall Monster 'difficulty'. I am a firm believer in the importance of player activity affecting long-term Monster behavior. Specifically assuming that Player guilds will be able to construct fortresses and such pretty much anywhere, the location and size of these fortresses should determine Monster 'layout' in the area.

Taking EvE Online's PvE system as an example (it is static however, unfortunately). Solar systems are assigned 'security ratings' in the form of numbers ranging from 1.0 to -1.0; the lower the number is the stronger the NPC presence in that system. You will have very weak enemies spawn in 1.0 space while you can have very strong enemies in larger numbers (even named) spawn in -1.0 space.

I believe this system can work (slightly modified) in Mortal.

Take the following 'difficulty tiers' as examples.

Example of old system:
Town Area -> level 1-10 monsters (weak)
Road to Mountains -> 10-15 monsters (medium weak)
Mountain Valley -> 15-25 monsters (medium)
Path to Mountain Pass -> 25-35 monsters (medium strong)
Mountain Pass -> 35-50 monsters (strong-very strong)
Mountain Pass Boss Battle -> Boss (over 9000, etc)

Example of new system:
Town area -> 1-10 (safer)
Road to Mountains -> 10-15 (safe)
Mountain Valley (with player stronghold) -> 1-10 (safer)
Path to Mountain Pass -> 10-20 (safe)
Mountain Pass -> 20-30 (hazardous)
Mountain Pass Named Enemy (possibly) -> 30+ ? (very dangerous)

As you can hopefully see, the Monster activity/density is variable based on both NPC and Player activity.

Be aware however, that as this system is a dynamic one you should never be 100% safe anywhere. The goal of a dynamic PvE system is to allow you safer areas to PvE in, not to prevent you from getting ganked by a raiding party of Orcs that just so happen to be moving through the area (perhaps getting ready to ambush a caravan?). Your own safety is still reliant upon your skills as a player, not on some default 'max Monster level' based on the area you are in.

~~~

That's pretty much all I have for now. It can probably use a bit of polishing, but I wanted to type it out before I lost the idea.

I realize that the system above is quite complicated but I believe that a strong PvE system is just as critical to the success of this game as a strong PvP system is.

Please discuss, comment, etc.. but keep it limited to strictly PvE only. PvP has its own set of rules that should be kept to separate discussion.

[OMGPIE]

Interesting I like it but I would like to hear from the devs what kind of AI the mobs in MO will be getting.

[Traceur]

yes to everything above.

i would also add that "safety" should be part of it's reaction to stimuli: being able to account for NPC/player-population & to decide it's not a good spot to stay or hunt in (or to mount up an attack). this way they'd react to player cities and such as well.

i can only hope that the AI is good enough, because honestly i think it would be nearly impossible to do it by basic scripting.

[Shinzon]

I like this, it creates a chance of starting cities being raded by NPC's, and essentialy any zone can over time turn hostile to even the most seasoned players; and the true reaon to venture outside into the far away lands would be to scout for new resources, as should be; not to grind the lvl 60 Boars that drop a sweet dagger...

[Torgrim]

Very good idea and i like it alot, it all comes down to how good SV are in programming

[Shinzon]

Probably more on the bandwith avalable, I am still curious to see their solutions to allow for this kind of gameplay...

[Traceur]

bandwidth? wouldn't most of the behaviors the OP talks about be server-side?

[Shinzon]

Yeah it would probably load the server more, but it is still more data to transmit... Guess I was talking more about FPS PvP more, quite Off topic, scuse me...

[sushibandit]

Whew, replies. I was getting worried as of Sunday morning/afternoon when nobody had responded yet.

Originally Posted by OMGPIE

Interesting I like it but I would like to hear from the devs what kind of AI the mobs in MO will be getting.

The goal here is to prompt both discussion and suggestions about how we as players think the AI can be made better than other games. Yes it will be much easier to make suggestions regarding Monster behavior once the Devs give us some information about what they've designed so far, but my hope for this post is to generate some interesting ideas for the Devs to (hopefully) see and act on.

Originally Posted by Traceur

i would also add that "safety" should be part of it's reaction to stimuli: being able to account for NPC/player-population & to decide it's not a good spot to stay or hunt in (or to mount up an attack). this way they'd react to player cities and such as well.

i can only hope that the AI is good enough, because honestly i think it would be nearly impossible to do it by basic scripting.

Your comment regarding the Monster-perceived 'safety' of an area is definitely covered under Rule#3 above. I would like to add, however, that Monsters should not have a 'super 6th sense' in regards to detecting Player activity in the area. By that I mean that the Monster activity should not be based on a raw number like 'number of x type killed in x area over x time' but on the Monster's communication and observation skills. If Players raid a Monster village, the survivors (assuming there are any) should be spreading out and warning other Monsters of the same type/affiliation of your presence, which the other Monsters should react to based on their behavior patterns and knowledge of your party's strength. IE; orcs would probably form up a counter-raid party while weaker Monsters would most likely start packing up their villages.

And yes, It would quite probably be nearly impossible, if altogether impossible, for what I'm suggesting to be accomplished with generic script-based AI. At least according to my understanding of how AI works.

Originally Posted by Shinzon

I like this, it creates a chance of starting cities being raded by NPC's, and essentialy any zone can over time turn hostile to even the most seasoned players; and the true reaon to venture outside into the far away lands would be to scout for new resources, as should be; not to grind the lvl 60 Boars that drop a sweet dagger...

Firstly, I'm hoping that the MO Devs won't implement a ridiculously unrealistic looting system to the extent that boars might drop a dagger (unless a Player came along and stabbed it before you ran across it).

I also see the likelihood of 'starting' areas, if they exist in the traditional sense, getting raided by Monsters to be very slim, mainly due to a constantly high Player count and activity, as well as highly visible NPC guard presence. But of course, based on how MO deals with 'starter zones' it may very well be possible for a group of Monsters to raid the area.

But I do completely agree with your comment regarding resources. Aside from exploring new lands to find new enemies, possibly even unique enemies, I believe that the primary purpose for exploration should be for resources such as Monsters, crafting materials, food sources, ground to build your house/guild fortress on, etc...

Originally Posted by misc..

wouldn't most of the behaviors the OP talks about be server-side?

Yeah it would probably load the server more, but it is still more data to transmit...

Everything I talked about in the OP, aside from Player activity, would be server-based. If it's too difficult to implement in real time, at the very least we should see the results of Player activity affecting Monster behavior every 'tick' or after a default set period of time. This time period should be relatively short, as it could make Monsters behave rather wonky if they are reacting to say... the results of several village raids all queued up simultaneously.

[IronFist]

All very good ideas, very long time to implant them into the game.

[OMGPIE]

Hopefully the beta will be released a while before the game so the people's ideas can be tested.

[osteo]

I luv the idea i could picture this to work out very well. This could make it tough for small guilds to get there cities going unless the small guilds investigated the area very well too make sure not too hard monsters were around.

Also if monsters are organized like this though like you brought up there would be some sort of boss what happens when this boss dies? do the ones that are left behind merge with other groups for a retaliation to the same area? Or do they just disappear?

What do you all think?

[Tanque]

Originally Posted by osteo

I luv the idea i could picture this to work out very well. This could make it tough for small guilds to get there cities going unless the small guilds investigated the area very well too make sure not too hard monsters were around.

Also if monsters are organized like this though like you brought up there would be some sort of boss what happens when this boss dies? do the ones that are left behind merge with other groups for a retaliation to the same area? Or do they just disappear?

What do you all think?

If they had such a good AI (lets hope so) it would depend on the inteligence of the group,maybe they would run away or do a suicidal attack,more inteligent groups should be able to quickly adapt to the situation and get a new formation. There are many things that could happen with such a great AI.

[Traceur]

maybe they could have a hierarchy - if one dies there's someone next in line? maybe it's always the oldest one of the group, or maybe more like the military... or maybe they'd fight and the winner gets to be the new pack leader?

i think it should be different for different species.

[sushibandit]

That could be a great way of dealing with a certain type of monster in an area you want to take control of for say, building a guild fortress. The monsters would have to be fairly intelligent, enough so to work together in an organized fashion.

Say this type of monster has a shamanistic leadership, meaning the one eldest shaman/etc... leads the group. Since you want the area and you are too weak to destroy the monster camp in open battle, you poison/assassinate/etc... that tribe's shaman and then watch the camp dissolve into chaos. Then you do the same for the next one, or perhaps take advantage of the chaos and attack the camp, or maybe you would destroy the natural resources within a certain distance of the camp and let the monsters migrate elsewhere, or any of a number of ways you could deal with an enemy like that in the real world.

That's the kind of PvE AI reaction I want to see in Mortal. Real time monster behavioral patterns influenced by player activity.

[Traceur]

maybe pack-leader would be a separate skills for the creatures? so if you just killed the oldest one, the new doesn't automatically gets access to all pack behavior tactics (AI scripts), he needs to gain them with time...

[MattPuerto]

Good ideas, maybe a bit over the top if you want to play in a huge explorable world but it would be great if any of your suggestions make into the game!