Sunday, July 20, 2008


Yes, the title says it all. Today, I want to talk about one of the biggest issues plaguing MMO's - the great divide between PvE players and PvP players.

In every MMO I've ever played or heard of, the story is the same.

Boy meets game.
Boy plays PvE.
Boy clocks game.
Boy plays end game raiding, and becomes "leetsauce" at the game.


Eventually boy gets a little bored one afternoon, and decides to venture into PvP. Then, 80% of the time...

Boy gets pwnt by PvP, and scurries on back to PvE, never to return again.

Why does this happen so often? Why is it that someone who has mastered every aspect of PvE is an absolute nub at PvP? Isn't it the same game?

Well, fairly much, no, it's not the same game. That's because monsters in PvE have really bad AI. I mean, terrible. Like, worse AI than you had in RPG's from the early 90's. To compensate for crummy AI, game developers jack up the HP, armor and damage of the monsters. This makes the game challenging again, but not in the fun way. More in the "ok we need 40 people for this raid, and if anyone messes up the aggro, we're insta-gibbed" kind of way. Sure, it's still a *bit* fun - many people live for this WoW-style PvE endgame scenario - but it's fun because of the camraderie and the cool items, not because the combat is particularly enjoyable.

Now, some games have (post-release) attempted to modify their AI to reduce this effect. WoW monsters will attack those healing the tanks, Guild Wars famously was updated to make its monsters run from (any) AoE, etc. But these are routinely simple additions that the player base takes all of about 45 minutes to work around.

Anyway, the AI is bad, so skills that would be crummy against real opponents are often the ones that destroy the monsters with ease. So, you get to PvP and use your good ol', tried 'n' true PvE tactics.... you send your tank in, wait until they "aggro" him, and then blast them with AoE spells. Except... they ignore the tank (who does pitiful damage), run straight in, and start smacking your healers around.

Uh oh. Your healers cast various protection spells, your AoE shift targets to try and make it difficult to attack the healers.... so the enemy changes targets, and goes after your similarly squishy nukers. They refuse to ball up, they remove debuffs you place upon them ASAP, they swap to unprotected targets, and worst of all, you're several men down because your tanks are not geared up to do damage, and are being ignored in this fight. You just can not win.

Is it your fault? Of course not! This is the stuff the game designers have made you learn to get through the high end PvE content. It's the best possible tactic in top level raids.

So what can game designers do to get rid of this effect, reduce the steep PvP learning curve and make PvP fun and accessible for everybody? I'm sure there are lots of solutions, here are some of mine:

1) Your monsters need good AI. I mean, we're at the level where computers can go toe to toe with the chess world champions, let's push the envelope a little people. Sure, you're limited by the real-time nature of the game and the fact that servers have a lot of stuff to do other than just AI calculations, but that's why you designed your game around having good AI didn't you? In fact, you designed your combat system with the specific goal of being able to predict good tactics for the computer to adopt without to much fussing around. You knew exactly how you wanted your combat to play out before you designed your first skill, and it made everything a lot easier.

Apart from good design decisions from the outset, either give your servers more grunt (can anyone believe an entire Lineage 2 server is managed by ONE computer?), or invent some kind of crazy server-client outsourcing scheme to get the clients to perform lengthy calculations for you. Yes, this screams "hack me", but not if you add random server-side checks of the calculations, followed by insta-banning if the server and client calculations don't match (ie, if the player tried to tell the server that all the monsters wanted to bunch up and sing christmas carols instead of attacking).

I don't care how you do it, MMO's need better AI to compete with regular offline RPG's.

2) With good AI, you won't need to give every single monster quadruple damage, 100 times more hitpoints, etc, etc. Sure, some monsters should be epic and feature such enhancements, but in general the game should play by the same rules that the players have to.

3) Know your game. Let me give you a mini breakdown of how 2 characters work in Guild Wars. Warriors are the main source of damage - they have the best single target DPS, they have great disruption, awesome spike ability, and are highly armored so that they can push deep into the enemy lines without too much fear of retaliation. Monks, on the other hand, are responsible for protecting the party, casting various enchantments to greatly mitigate damage, healing to reverse any damage that does go through, and removing debuffs.

Anyway, Guild Wars was released with a bunch of "premade" PvP character builds, which featured some of the most spectacularly shite combinations of skills imaginable. One such was the "Paladin", which featured a Warrior/Monk character (in Guild Wars you choose 2 classes for your character), with a mix of some of the worst damage dealing skills, and some of the worst healing skills. Not only were the individual skills bad, but putting them together in that way is silly. If a melee character has to stop to heal his group, the target he is chasing has time to run away. On the other hand, healing skills cost energy, and warriors have the worst energy regeneration in the game. Finally, the healing skills all promoted being cast on the warrior itself, which is worse than useless, since nobody targets warriors to do damage anyway, making them an utter waste.

There were, I don't know, 20 of these builds, which players new to PvP would of course dutifully take into battle, excited by the lure of being a "Paladin" or some other such garbage. The experienced PvP-ers would yell at them, the team would lose, and the poor fellow would trudge back down to PvE.

If the makers of Guild Wars instead had have included a decent warrior template (which their alpha testers could easily have given them, along with instructions for its correct use), anyone wishing to venture into PvP has a much easier time of things.

4) Don't make crap skills. There is a skill in Guild Wars called Spiteful Spirit. It is a long-lasting debuff that makes it so that whenever the target does anything (for example, swing a sword), AoE damage is done to them and their team-mates. It's a great skill in PvE as you can imagine, since the monsters are stupid and swing through it repeatedly, massacring their fellows who clump up to get to the "tank". In PvP it sucks horribly, since anybody who has it cast on them can just ensure that they aren't clumping up with their team-mates, making its damage laughable, or simply wait until their monk removes it. Only people very inexperienced with all the skills in the game will attack through it in a big clump of their allies.

So what is its purpose in the game? Good players don't use it, or get beaten by it. It is therefore a tool by which one noob can own another noob. Why do we need such a thing in the game? Is it because press releases look good when they say that the game contains over 500 skills? Whenever I see a statement like that, I roll my eyes. It's just not possible to create 500 skills that are all wildly different and equally useful. Maybe 20% of the skills in that game have seen tournament play, and probably no more than 10% at any one time thanks to various nerfs etc.

Don't make sucky skills.

5) Finally, as Crimson referred to in an excellent blog post recently, don't pit players of wildly different skill levels against each other. You need some way to measure player skill. It can be as simple as looking at how often a player obtains wins vs other teams of various ratings. Guild Wars features this in some arenas (Guild vs Guild, and 1v1 ("Hero battles")), but not in other arenas where it could be equally useful (Random and Team Arenas (4v4), Hall of Heroes (8v8) and Alliance Battles (12v12)). As a result, the latter are often a "noob farm" for the seasoned veterans of the game. Sure, this is fun every once in a while, but it gets old fast, and it's not fun at all for the losing side.

In conclusion, make monsters smarter, and PvE players will become good at the same game that PvP players are playing, and be far more likely to keep playing (hint: and give you $$$).


Garumoo said...

So true.

Another element is just how fragile most characters are - if you compare the typical amount of damage players put out vs the amount of health they have, and you can see how a moments distraction can result in you being dead. Most PvP encounters are over in 10-30 seconds.

The more enjoyable and memorable PvP fights I've been in have been the longer ones - lots of sparring and tactics and counter-tactics. Sadly, these have usually been up against noob locks or tankadins who were not particularly good at killing me.

mbp said...

Wow brave move CS to brand one of the most popular and powerful PVE skills in the game a "sucky skill".

Mind you as when I was playing GW as a confirmed care bear my biggest problem with Spiteful Spirit was that it should have been a Mesmer skill. WTF its the aoe version of empathy!

Melf_Himself said...

But mbp, you're forgetting the point I was trying to make. Spiteful is indeed a good PvE skill, but only because the AI is so bad.

If the AI were better, Spiteful would be a bad skill, and people making the transition to PvP wouldn't try using a bad skill.

Improving AI will make the game better, and the players smarter.

Tesh said...

Mark Rosewater over on Magic the Gathering design calls "sucky" cards "skill testers", especially in a drafting environment where card choice is paramount. I don't completely buy that argument, but there's a nugget of truth to it. All the high level players know to avoid dumb cards, and buying sucky cards in a booster pack isn't cool... but they can indeed be teaching tools.

That said, I think that it would be better to teach with tools that aren't so... sucky.

Then there's the PvE vs. PvP AI concern. I completely agree. The AI threat mechanic, so expertly showcased in WoW, is just... dumb. It's not smart tactically or strategically. If I take a step back and see that I'm fighting a boss monster who exhibits the AI of a rabid sloth, it really isn't all that fulfilling to destroy him. It's putting something terminally stupid out of its misery with a side chance of looting something good. That's why combat feels so scripted, and it really doesn't feel very fulfilling.

MMOs are ostensibly about playing with other people. Mind you, I love some great single player RPG gaming. It's just that MMOs as a genre need to embrace their strengths instead of trying to be everything for everyone. Intelligent AI would go a long way to preparing people for the inevitable PvP.

Melf_Himself said...

I used to buy into that "training" stuff, but not anymore. I've played too many games where I have to basically learn an entire new language (i.e. the name, icon, and function of, say, 500 skills) just to get to grips with how to play the game.

Compare this to being a FPS player, where your set of learned skills pretty much applies straight off the bat to the next game you play.

They need to minimize the learning curve a lot in RPG's in general. I think a logical combo system based on a few core skills would go a long way towards this - see the new "Warden" class in the LotRO expansion for what I think could be a step in the right direction.