From the WotC Boards: Mearls on 'Aggro'

[TwinBahamut]

I'm not sure that all this posturing is really accomplishing much of anything, especially given that we've been told that the rules are not going to make it into 4E.

At the same time, a number of folks have said that it's "obvious" that these rules are a bad idea for both D&D and MMORPGS. It isn't obvious to me in either case, so I ask: why is it obviously bad game design to include agro-like rules in either a MMORPG or D&D. I ask this because World of Warcraft seems to be fairly popular despite having such rules.

Just Wonderin',

--Steve

I can't speak for the people who think it is "obvious", but if you ask me, the main problem with Aggro rules is the fact that they cause battles to be dangerously unstable. Such rules depend on the tank being on the top of the aggro list at all times, and that other party members deliberately hold themselves back to avoid being above the tank on the aggro list. Unless the whole party obsesses over aggro control at all times, and nothing occurs outside of the team strategy, then fragile attackers or healers tend to be killed suddenly. It reduces team flexibility and the chance of trying new strategies, and forces a game where everyone does the same thing all the time.

Anyone's guess is as good as mine as to why it works for WoW. The only MMORPG I have ever played is Everquest, so I know exactly how bad aggro rules can be implemented, and the problems they cause. Ugh, trains and the fact that Area of Effect attacks were suicide, and the high death rate, were more than enough to convince me how bad they could be.

 

[Wolfspider]

I'm not sure that all this posturing is really accomplishing much of anything, especially given that we've been told that the rules are not going to make it into 4E.

At the same time, a number of folks have said that it's "obvious" that these rules are a bad idea for both D&D and MMORPGS. It isn't obvious to me in either case, so I ask: why is it obviously bad game design to include agro-like rules in either a MMORPG or D&D. I ask this because World of Warcraft seems to be fairly popular despite having such rules.

Just Wonderin',

--Steve

I can speak to why I think that these kind of rules are a bad idea for D&D.

I think that it takes away control from the DM that he or she should have over what monsters do.

Maybe he's set up an encounter where the party runs into a group of goblins. With the kind of "aggro" rules that seem to have been considered for inclusion into 4.0 ("explicit rules . . . where monsters had to attack the fighter or paladin or a creature's tactics dictated that it attack the nearest foe," to quote the original blog), monsters would have a set pattern in the way that they attack party members.

For example, maybe goblins, being relatively weak, automatically attack unarmored characters. Now, that may seem pretty wise of weak creatures like that, but if this behavior is hard-wired, so to speak, into their Monster Manual description and has rules to back it up, then the DM cannot have the monsters behave in any other way. (Of course, the DM could always invoke Rule 0, but if the "aggro" rules are deeply ingrained into the new rules system, that might throw off balance and cause other problems.)

Maybe as a DM I want the goblins to attack the cleric because he is wearing the holy symbol of a deity that their shaman leader particularly despises. Sorry. They've got to go for the wizard.

Maybe as a DM I want the goblins to fear the wizard because they have seen what their own dark wizard master has done to those who have failed him. Sorry. The goblins have to go for the wizard despite their own fears.

It says so right in the Monster Manual.

Or might have. Thankfully, it now seems as if it won't.

Whew!

EDIT: Actually, the best explanation for why it's a bad idea comes from the original article itself. (Imagine that!) It reads:

[An "aggro rule"] restricts DMs needlessly. We don't want to tell DMs, "You have to do this." It's pretty lame to force DMs to walk through a monster script. It might be interesting for a specific monster (the clockwork knight programmed with three specific routines) or encounter (the zombies in the temple of Orcus attack good clerics above all other targets), but not as a core rule.

 

[Gloombunny]

At the same time, a number of folks have said that it's "obvious" that these rules are a bad idea for both D&D and MMORPGS. It isn't obvious to me in either case, so I ask: why is it obviously bad game design to include agro-like rules in either a MMORPG or D&D. I ask this because World of Warcraft seems to be fairly popular despite having such rules.

Two reasons: big burly guys with big sharp axes should be dangerous, not just damage sponges. And having monsters attack the guy standing around shouting insults instead of the guy setting them on fire is stupid.

These points are axiomatic as far as I'm concerned.

 

[Driddle]

I hope ever so much that they decide to include rules for camping on "spawns" so serious gamers can "grind" some easy experience.

(Did I use those terms correctly?)

Appendix B of the new PHB: http://www.worldofwarcraft.com/info/basics/glossary.html

 

[Cadfan]

Its not at all obvious why an aggro mechanic is bad for D&D.

There are two problems with creating one, though.

First, D&D has collision detection that is (or rather can be written to be) much superior to that available in current-day MMORPGs. Aggro rules eliminate the need to program a way for a character to physically block a monster from moving past him. D&D already has rules for that. so it doesn't need aggro as much.

Second, D&D is more complex than most MMORPGs in terms of available strategy. Monsters come in more varieties, with more tactical options available. Most MMORPG attacks boil down to "hit enemy, maybe inflict status effect." RPGs are a lot more complicated. What if there's a wall of fire that deals 5d6 damage between a wizard and a monster, and there's a fighter on the side with the monster? The wizard casts a lightning bolt and does 50 points of damage to the monster. The fighter hasn't done more than 10 in any given round. Does the wizard draw aggro? Does the wall of fire have an effect? How is it factored in? Etc. MMORPGs rarely deal with this sort of complexity, and when they do, they rarely do it realistically. A DM's judgment call is usually just as good or better.

Realistically, most DMs have an informal, mental version of aggro that they use as a rule of thumb in planning monster tactics. So informal aggro (and its unacknowledged brother in law, morale) will be part of D&D no matter what.

 

[GSHamster]

Two reasons: big burly guys with big sharp axes should be dangerous, not just damage sponges. And having monsters attack the guy standing around shouting insults instead of the guy setting them on fire is stupid.

These points are axiomatic as far as I'm concerned.

This seems to be a paradox to me.

If the fighter is just as dangerous as the mage, the monster should attack the mage because the mage is more fragile. If the fighter is less dangerous than the mage, the monster should definitely attack the mage because it poses the greatest threat and is more fragile. If the fighter is more dangerous than the mage, I feel very sad for the mage.

So why should the monster attack the fighter? Wouldn't the superior tactic in all cases be to attack the mage?

Put a party of skilled PCs against a similar party of NPCs. Watch what they do. Most of the time, they will target the healer and wizard types first, because that's the best strategy.

Yet that's not fun. Most DMs instinctively realize this, and have monsters use sub-par tactics, attacking the fighters because they are closer, and not targetting the healers/wizards unless forced to. They basically keep a mental "threat list".

3E tried to solve this paradox with Attacks of Opportunity, punishing the monster for ignoring the fighter. MMOs use threat mechanics* to determine who the monster attacks, which becomes the basic gameplay.

So that's the basic question. If a new DM asks you "How do I determine which PC to attack?", what would be your response?

* They also use threat because collision detection is computationally expensive, especially with latency. Another reason is collision detection would allow griefing on a massive scale -- simply stand in doorways and refuse to let people by. Or fight a monster that can't hurt you badly in the doorway.

 

[Celebrim]

Really? What should videogames be doing instead?

Errr... You say that like it is a rhetorical question. Ok, two can play that game. Isn't it obvious what would be better?

Ideally, the AI behind monsters should be constructed with some sort of theory of mind such that to the human observer, he consciously or unconsciously describes the actions he sees the computer's avatar's performing as being guided by a rational mind. We've been doing this for ages. One of the most famous examples would be the difference between Pac Man's AI, and the AI for the ghosts in Ms. Pac Man.

Most AI's constructed like that run short events based on examining the game state and have a bit of built in randomness in choosing what to do next. The goal is not so much in getting the AI to behave in the smartest possible fashion, is it is to get it to behave in a manner which is believable to the observer. Believable minds don't act in predictable manners and can't be consistantly manipulated.

Granted, AI like that is more complex (and hense more expensive), but you see that sort of AI outside of a MMORPG environment. The reason you see really really simple AI's in a MMORPG game is simple - performance. Simple scripts eat up few CPUs. And while I'm not an expert in programming MMORPGs, I'd guess that there is a boost in emmersiveness if for any given event, all the clients that observe the event can predict how the monsters will move next. So really, we are doing it this way not because it makes for a better game, or even to a certain extent because we don't know how to do it better, but because given the hardware/bandwidth contraints that an MMORPG opporates under a more sophisticated AI causes other issues.

 

[Gloombunny]

This seems to be a paradox to me.

If the fighter is just as dangerous as the mage, the monster should attack the mage because the mage is more fragile. If the fighter is less dangerous than the mage, the monster should definitely attack the mage because it poses the greatest threat and is more fragile. If the fighter is more dangerous than the mage, I feel very sad for the mage.

So why should the monster attack the fighter? Wouldn't the superior tactic in all cases be to attack the mage?

Put a party of skilled PCs against a similar party of NPCs. Watch what they do. Most of the time, they will target the healer and wizard types first, because that's the best strategy.

The solution is to change things so that it's not the best strategy (as 4e is doing), or assume that it is the best strategy and balance things around it so it's fun anyway (as Guild Wars did). Requiring that all enemies fought act in the same predictable but insanely stupid way is... not a good solution to the problem.

* They also use threat because collision detection is computationally expensive, especially with latency. Another reason is collision detection would allow griefing on a massive scale -- simply stand in doorways and refuse to let people by. Or fight a monster that can't hurt you badly in the doorway.

Is collision detection really that hard? Guild Wars doesn't seem to have any problem with it, and they don't even have monthly fees from all their players to support their server expenses. Player-griefing is easily solved by letting players not hostile to each other pass through each other.

 

[Celebrim]

If a new DM asks you "How do I determine which PC to attack?", what would be your response?

Put yourself in the mind of the monster. Stupid or mindless monsters will use very simple strategies that more or less amount to 'close with and attack whatever is closest to me'. Animals will use instinctive strategies (ambush, stalk, pounce, grab, trip, whatever), and will tend to become confused when presented with problems or challenges not in thier ordinary experience. They do this because they can't evaluate the situation around them. Smarter monsters will better assess the situation around them and make smarter decisions. A monster that is at least as smart as you should be making the decision you would make if that monster were your PC and you had the same information that is available to the monster. A monster that is significantly smarter than you are (say you believe you have a 16 INT and the monster has a 24 INT) should be played the same, but using all the information available to you as a DM.

 

[FadedC]

Errr... You say that like it is a rhetorical question. Ok, two can play that game. Isn't it obvious what would be better?

Ideally, the AI behind monsters should be constructed with some sort of theory of mind such that to the human observer, he consciously or unconsciously describes the actions he sees the computer's avatar's performing as being guided by a rational mind. We've been doing this for ages. One of the most famous examples would be the difference between Pac Man's AI, and the AI for the ghosts in Ms. Pac Man.

Most AI's constructed like that run short events based on examining the game state and have a bit of built in randomness in choosing what to do next. The goal is not so much in getting the AI to behave in the smartest possible fashion, is it is to get it to behave in a manner which is believable to the observer. Believable minds don't act in predictable manners and can't be consistantly manipulated.

Granted, AI like that is more complex (and hense more expensive), but you see that sort of AI outside of a MMORPG environment. The reason you see really really simple AI's in a MMORPG game is simple - performance. Simple scripts eat up few CPUs. And while I'm not an expert in programming MMORPGs, I'd guess that there is a boost in emmersiveness if for any given event, all the clients that observe the event can predict how the monsters will move next. So really, we are doing it this way not because it makes for a better game, or even to a certain extent because we don't know how to do it better, but because given the hardware/bandwidth contraints that an MMORPG opporates under a more sophisticated AI causes other issues.

Well that's not entirely true, MMORPGs usually have tremendous server power and are capable of (and do) run some pretty complicated scripts. They could in fact very easily get rid of tanking mechanics and just have the monsters behave perfectly "intelligently" killing healers and mages first and ignoring the fighters. It might actually be interesting to have a game like that, although it would require a massive redesign of the rules, requiring everyone to be able to take a beating and not having any class roles. But it's chosen at least in WoW to use a traditional warrior holding off the enemy system with classes. And plenty of people enjoy it, but no game is for everyone.

I hope ever so much that they decide to include rules for camping on "spawns" so serious gamers can "grind" some easy experience.

(Did I use those terms correctly?)

Your terms are correct, although a bit out of date as that was more of an EQ thing then a WoW thing. People tend to earn xp from quests in WoW.