Mob Migration – An element that should be part of all AAA MMO’s

This thought process works the same regardless if there are quests or not, in PVP or PVE. 

One of the reasons a game gets stale to me and does not have the longevity that it should, can be accredited to the lack of mob migration.  Here is an example:

IN GAME: “I am on my new first time in the game character and the whole world is mystery to me.  I love that feeling of new discovery.  Now a have a quest to kill 4 fire beetles and I wander around complaining to myself that the dumb NPC did not at least give me the location of the beetles….#@$%.  Finally I find them and complete the quest.  Whoot, now I am happy and forget about my complaints.  I also notice that the beetles dropped a crafting material that I will save for later.

Later that day I decide I want to try crafting and need more of that crafting material so I go back to where I found the beetles and farm them until I have the materials I think I need.

A month later when I am on my alt I know where the beetles are and go to that location.  Now the quest becomes trivial and no one needs the materials due to its low grade and abundance. “

As you can see the game is starting to get stale.

Now let’s replay that scenario with mob migration.  Now the beetles move around in the area where the quest is given.  As you can see not knowing exactly where the mobs are going to be adds to a game’s longevity and it stays fresh longer.

Taking it a step further.  Let’s say the zone around the quest giver is tier 1 and the subsequent areas further out are tier 2 and 3.  Now the beetles can be in any one of these zones with their appropriate level increases. 

IN GAME “I’m running around for an hour trying to find those stupid beetles and now all the mobs are orange con to me.  This %$#^ game really sucks.  I finally find them and they are all several levels above me.   Argggh this is going to suck.”

And I now have a choice, try to finish the quest with the higher con mobs or wait till later and see if they migrate back to a lower con area.  Ahh, the possibilities.

Now let’s take this mechanic all the way to end game where my guild crafters are looking for Tier 10 (top tier in this example) beetle materials to have any chance to get us the gear we want.  The rarity of these mobs is created by the fact that they are only max tier when in these highest zones and would drop lower end mats if they wander out of the zone.  So it is now or never, disreguarding the fact that I could wait for the next sighting.  Also my faction is not the only ones who hunt them.

IN GAME: “I finally spot seven of the elusive max level fire beetles and they are all three star group mobs.  Not only do I have to put a call out to the guild for backup but due to the rarity we may want to let the alliance know.  But I have bigger problems at the moment.  There is already a group of Albs pulling one of the beetles.  Bastages!  So I wait till the reinforcements arrive.  Just before we are set to attack a group of Hibs show up and start pulling a second beetle from the other side of the spawn.  Spineless Hibs would rather get the materials than fight the Albs (shakes head).  Finally my allies arrive and a horn blows and a cry goes out…CHARGE! That was the last thing I heard before we tore into the Albs and the fight for resources began.  As a Middie we instinctively knew who the target was without anyone having to tell us.  In the end the field was red and there were bodies everywhere.  And I don’t remember if we killed any beetles”

So even mob migration can be a key element in PVP and definitely help games stay fresh to all players.

Implementation of Mob Migration - How it can be done and why.

Mob migration can be problematic for many reasons.  But here is a brief synopsis of my thoughts.

For starters, what would prevent someone from following the rare beetles after they leave the max zone and report as soon as they re-enter the zone and become the prize again.  This would require the second part of mob migration.

 An example mechanic could work something like this.  Every time the server migrates from day to night the mob server would go down for randomization and when it comes back online the mobs are all moved around in their respective habitats and zones.  So with the setting or rising of the sun the mobs would all shuffle around.  And in between the mobs all vanishing and returning may be a 5-10 min period where there are no mobs at all. This would create a good opportunity for community dialog.

Now this mechanic would create some unique game play challenges especially if you have just 9 of the required 10 mobs to finish a quest.  You will have to find them all over again just to get that last mob.  Especially interesting is the fact that they may come back as a completely different con to you.

This principle can be used not only for animal / monster mobs but for NPC nomads that tend to break camp and move with their prey.

So let me know what you think and if you have ideas that may make this even better please elaborate.  CU is in its early stages of development and we all want to make this the best game ever and a blueprint for all MMO’s to follow.

Dark and Light proposed doing this a few years ago, but never got it working. They were going to have mobs/npcs change locations based on seasons.

You could take it a step further, and have some resources grow at different rates based on climate/seasons....trees, plants, food etc...

Currently in MMO's:

1. Query table

2 Spawn mobs

3 Run idiot roaming AI to take 3 steps and turn 90 degrees right repeat until Diablo 4.

Simple fix:

Mob Group Agent -

1 Run only on startup, run all relevant pre-load to prepare spawn

2 Querry Possible Locations table

3 Throw dice 3 times

4 Check if matches with possible locations table

5 If not, return 3, else forward  (or just f'kin random values matching possible values table)

6 Save 1st location to temp table

7 Repeat for each Mob Group Agent in Area(ThisRegion)

8 Run check, Saved Location 1 - distance to - Saved Location 2 is more than 1500 distance units. If not, repeat above for Saved Location 2

9 Query table for Mob Spawn Distance Info

10 Calculate Mob1 lateral distance + angel from Saved Location 1 based on Mob Spawn Distance Info

11 Repeat for all Saved Locations

12 Spawn Mob Agents.

While it could be done, there is a steep cost that you don't seem to have considered.  In order to draw anything, you have to have the relevant assets for it loaded into video memory.  If mobs don't move, then you can carefully consider where to place them such that the same art assets get reused a bunch of times in one area, but don't get used at all in most other areas.  If mobs move around a lot, then loading assets in time is much harder to do.  Off-hand, I see three ways to do it:

1)  Running the game off of a solid state drive is part of the minimum system requirements--which means that most of your prospective players can't run the game.

2)  Make the game very heavily zoned, with lots and lots of loading screens to give you time to check which mobs you need for the next area and load them.  Run for 30 seconds, stop for a loading screen, run for 30 seconds, stop for a loading screen, and so forth.

3)  Use few textures and low quality graphics.  It's a lot quicker to load art assets for a character into video memory if there are fewer assets to load.  This will net you graphics that would have looked rather dated 5 years ago.

Any one of those three would get the job done and make what you want possible.  There might be other options that I haven't considered, most notably by making how mobs can migrate very, very restricted.  For example, sometimes this area has mob A and that area has mob B and sometimes the other way around, but no other mob migration can ever exist for those two mobs or those two areas.

But magically making it so that mobs arbitrarily can move around without having some serious drawbacks is not practical.

It's a pvp game, little/no pve at all, were not even sure there will be any mobs.

Also, since when in CU a AAA mmo?

Originally posted by Zinzan

It's a pvp game, little/no pve at all, were not even sure there will be any mobs.

Also, since when in CU a AAA mmo?

There will be mobs. Not much, and they may not drop any crafting material, but there will be some. At least the 3 dragons will be there.

Technical it is not that difficult to do.

Just have N copies of the spawn table for each zone, and randomly pick a set to use (i.e. if beetle appearing in zone 1, then not in zone 2).

The question is whether players like that much RNG. And whether people like to fight things, more than to find things.

Don't think of it as mob migration - think of it as building your world on a foundation of simulations rather than static content.

Originally posted by Quizzical

2)  Make the game very heavily zoned, with lots and lots of loading screens to give you time to check which mobs you need for the next area and load them.  Run for 30 seconds, stop for a loading screen, run for 30 seconds, stop for a loading screen, and so forth.

It doesn't seem to me to be a fundementally different problem than managing the art as a player flyies through the world from zone to zone - it's just that parts of the world are zoning past even when the player is standing still.

Originally posted by Quizzical

While it could be done, there is a steep cost that you don't seem to have considered.  In order to draw anything, you have to have the relevant assets for it loaded into video memory.  If mobs don't move, then you can carefully consider where to place them such that the same art assets get reused a bunch of times in one area, but don't get used at all in most other areas.  If mobs move around a lot, then loading assets in time is much harder to do.  Off-hand, I see three ways to do it:

1)  Running the game off of a solid state drive is part of the minimum system requirements--which means that most of your prospective players can't run the game.

2)  Make the game very heavily zoned, with lots and lots of loading screens to give you time to check which mobs you need for the next area and load them.  Run for 30 seconds, stop for a loading screen, run for 30 seconds, stop for a loading screen, and so forth.

3)  Use few textures and low quality graphics.  It's a lot quicker to load art assets for a character into video memory if there are fewer assets to load.  This will net you graphics that would have looked rather dated 5 years ago.

Any one of those three would get the job done and make what you want possible.  There might be other options that I haven't considered, most notably by making how mobs can migrate very, very restricted.  For example, sometimes this area has mob A and that area has mob B and sometimes the other way around, but no other mob migration can ever exist for those two mobs or those two areas.

But magically making it so that mobs arbitrarily can move around without having some serious drawbacks is not practical.

You're ignoring the simple theory of dynamic loading and consider zones as something that governs all and only one can exists at any one time. 

Create Zone Environment as qubical boxes.

Connect Zones to others from their edges.

Create Possible Mob Spawn Area Tables (server side) as limited areas witin the Zone Box. If wanted, create them all the way to the edge, and continue with another one in the next Zone Box.

Set Player Agent View Distance to 1500 distance units

Load all textures used in This Zone

Show textures that are in 1500 distance units.

Load and show all Mob Spawn Tables + textures that are in 1500

Preload Mob Spawn Tables + textures that are in additional 1500 units. 

Drop from memory Mob Spawn Tables + Textures that are further away than 3000 distance units

When Zone Border Distance <3000 units, Load Adjacent Zone Textures.

When Zone Border Distance >3000 units, drop all Adjacent Zone data

.

Also forget the idea that "Zone" means same for user than to the code. Barrens in WoW is a Zone for player, it might be as many as 4-8 zones from the engine's perspective. At best in any single time, client can have 4 zones loaded and on screen, but by dynamically dropping them, never more than that. 

At special zones that you expect high traffic, you can make situational rules. Such as you can only enter a Capital from one direction, thus you never need to load more than Capital + 1 Zone, and never Capital + 3 near the corner you're in + change those depending on what edge of the capital you are. 

You can often see to another zone as "back drop" further away than 3000 units, but u dont need all its textures, just need to create a "back drop box" version of the zone environment at very low texture quality, perhaps even only as single Item and use it as single element until you get close enough to load the actual textures. 

Originally posted by Aelious
Sorry, going to believe Quiz on this as he seems to know what he's taking about.

Quiz's posts the past few months have been overcomplications of otherwise simple processes. Lots of words and jargon doesn't mean it's a better solution.  In that one, he overcomplicates a process that has been handled fine in MMOs since NetImmerse(DAoC).

I too am slightly sceptical on the impossibility of implementing mob migration. Already many games have patrolling mobs and npcs and changing spawn tables. As long as spawning, patrol route, time and speed are known, I would assume it to be possible to increase the distances and the amount of patrolling mobs. Though I admit I'm not an expert in the technical aspects of loading objects to memory.. I would assume that as long as the server knows well beforehand where everything is and when it spawns if it's dead, it can send the client a command to load them in advance. My view might be too simplistic though and I am not posing as an expert. Would be great to hear some more explanation on the difficulty of implementing this.

As to why do this, I would also point out migrating animals, npc groups and weather effects could be added. Would do some wonders to immersion imo. I'm not talking of simply spawning in different locations, but actually moving ingame.