Monday, July 28, 2014

Complexity and Depth Part 2: Conflation Station

A few months back I wrote up a blog post about the ideas of complexity versus depth. Complexity at its core is how many rules do you have, and depth is what is the meaningful possibility space enabled by the rules.
 
Keen over at Keen and Graev's Gaming Blog postulated recently that slower combat had more depth. While I agree with the premise that slower combat can have more depth, it doesn't necessarily make it true, and Keen's sub-arguments rapidly show that it's too easy to conflate complexity and depth. More complexity does not automatically ensure more depth.
 
Complexity of Decisions 
"Today there are very few decisions to be made. One simply walks up to a mob and executes abilities in any order. The real decision is which order to use the abilities to kill the monster fastest–everything is about actively attacking."
First of all, I agree with Balkoth that Keen is conflating lack of difficulty with lack of depth. Leveling in WoW, Guild Wars 2, Final Fantasy XIV, and Wildstar all has to be sufficiently easy that the grand majority of the populace can perform it, or you lose your subscribers. In all of those games there are more difficult content where you do need to perform more difficult maneuvers to defeat the enemy.
 
Frost mages in WoW facing something higher level than they are will root, slow, stun, blink, and so on, using the tools at their disposal. Kiting is still used for difficult mobs all the time. Heck, just the other day I kited a bunch of mobs with my Enhancement Shaman while my friend DPS'd them down slowly, all the while avoiding damage.
 

 
I wish I had a video of me soloing Chimaeron on my Enhancement Shaman, because that was one hell of a puzzle trying to figure out what combinations of talents, abilities, timings, etc. I needed to win, but above is a video of the next fight, soloing Nefarian. While not as difficult, it still required the use of extant abilities such as Earth Elemental, Rockbiter Weapon, as well as all of my self-healing and defensive cooldowns to win.
 
Things like Root Rot, where you literally just root them and DoT them, still works to an effect. It's just that in most games you can't root them indefinitely. Being able to lock down a monster forever would be bad design, as there's no depth or thought there. Mind you, the depth comes from the players discovering this technique from chaining together different abilities, and that is a good thing.
 
But to say that using player-created techniques via a combination of abilities is dead just tells me that you're not attempting difficult content to begin with. The content is there, it's just not in your face for subscriber-maintaining reasons. You have to seek it.
 
One final note here: rotational complexity is an interesting beast. At higher levels of difficulty, it's basically muscle memory. The more complex classes, the better your ability to memorize chains of abilities needs to be to manage it. Once you've "solved" the intellectual challenge of "What is my rotation?", which for most games you can look up online these days, it becomes the mental/physical endurance challenge of maintaining it for the duration of a fight. A good blog post I think later will be delving into physical versus intellectual difficulties.
 
Aggro
 
First of all, most games still have aggro. Final Fantasy XIV and Wildstar it plays an extremely important part of multiplayer content. Guild Wars 2 doesn't even have the Holy Trinity or anything like it, so aggro as a concept is less about one person tanking and more about, "I'm dying, I need to drop aggro." It's really only WoW that's vastly simplified the concept.
 
The concept of aggro also has very little to do with the speed of combat. Wildstar combat is quite frantic, and yet aggro as a concept persists and works. Final Fantasy XIV combat is very slow in comparison to WoW, GW2, Wildstar, etc. aggro as a concept works identically to Wildstar's.
"Tanking took time, monsters took time to taunt and build up a safe aggro, and players respected that or died."
Tweaking aggro to Everquest levels such as Keen mentions above is just a knob on aggro as a whole. But the question here is does aggro add depth? Aggro is certainly another set of rules; taunts, waiting to DPS/heal, not ripping threat off the tanks, threat reducing abilities all add a layer of complexity for sure.
 
But I argue that in the Holy Trinity model, aggro does nothing to add depth. Your choices are let the tank get aggro, or you die. That's the possibility space added from the concept of aggro. I think a better way to look at this is to look at a system that uses aggro, but doesn't have tanks, per se, like Guild Wars 2. I'd argue it has a bigger possibility space granted because anyone can "tank". So if you're running low on health, you shouldn't be tanking, someone else can tank. Or if you can pin-pong an enemy between 3 or 4 different "tanks" to keep it running rather than attacking. Kiting, if you will.
 
As long as the Holy Trinity model only allows for one or two tanks, and anybody else gets dead if the boss so much as looks at them, there's no decision to be made. It's a trap. It definitely requires skill and knowledge of the system to handle, so is complex, but it doesn't add depth.
 
Class Specialization
"This could also be called the 'characters do one thing well' category. Having certain classes in your group would actually slow down the rate at which you could kill a single mob, thus slowing combat, but might improve your abilities to survive, pull multiple mobs at once and take a tougher spawn, or recover from battle quicker and move on to the next kill...These days everyone is a DPS."
Keen's complaint here stems from a lack of coordination required. Rather than having the requirements for a fight to be distributed across multiple players and each player having a single job, each player has more tools at their disposal to perform at fights and may have multiple tasks to perform in a single fight.

Instead of having debuffers, buffers, damagers, tanks, and healers, buffing and debuffing got rolled into everyone's classes in most modern MMOs, leaving you with damagers, tanks, and healers all capable of also buffing/debuffing. The tasks all still mostly exist (just ask any Warlock who has to put up Curse of Elements in WoW, or any healer who's tossing external cooldowns on tanks), it's just that each player can do more. That's not a reduction of depth, that's just a redistribution of complexity.

Combat speed once again has nothing to do with this point; it clearly does not support Keen's hypothesis whatsoever.

Managing Resources
"Managing mana consumption was often the difference between a great player and a good one. Healers who knew which heals to use and when, Wizards who knew how many times they should nuke to add the most efficient DPS to a group (the key being “efficient”), etc. Consume your resources and combat was slower. Have to worry about them at all and combat naturally becomes much, much slower."
Consume your resources, and combat would last longer, but not necessarily be slower. Run out of mana, and you'd need to wait for it to regenerate, but the cadence at which enemies use abilities doesn't change. Mind you, where his idea is correct is if resources are designed to run out quickly, it would necessitate not using your abilities at their maximum cadence.

Basically, if you go full throttle all the time, you run out of mana/focus/etc. That idea is quite sound, and you'll note it's in effect in FFXIV, where everyone has to watch their resources carefully. The difference is that you have abilities you can use that are effectively resource-neutral, or resource-gaining, and higher damage/healing abilities that eat away at your resources, so the game becomes a choice between when to go all-out, and when to conserve (also see: Arcane Mages in WoW).

Combat speed is independent of this variable, unless the game has no resource-neutral or better abilities. Then you're forced to hit fewer buttons: the downtime between each ability becomes greater than the GCD. If the designers have made combat a frantic, fast-paced affair, clearly waiting between abilities won't really work. Human nature would be to hit buttons as fast as possible in a panic-response to fast combat.

Leisurely combat, where enemy abilities are far and few between, you have lots of time to choose actions, or interrupt other abilities, to get out of the bad, and so on, would naturally allow for the player to pick and choose their abilities more carefully, and having that extra time means you can allow for more thought for each individual action, meaning you can allow for more depth. But that being said, individual abilities would naturally need to have more impact to make those choices meaningful. The fewer choices you're making, the more impact each choice needs to have to make it worthwhile.

I think this is the closest Keen gets to actually supporting his hypothesis, but again, having to manage resources can be independent of combat speed with the one exception I laid out. Then you may as well have a turn-based game--which is okay! I love turn-based games--but in the current "real-time" model of MMOs I'd argue doesn't really work. It's another trap for the players. An unnecessary, unfun skill cap.

Auto Attack
"Remember our old friend “white damage?” I love auto attack. I remember the days when it comprised of a massive portion of overall damage done by melee characters.  The entire concept is all but completely done away with in favor of rotations and constant ability usage. Older MMOs had fewer abilities (most of the time)."
Okay, I'll be brutally honest here. When I got to this part of the blog post, my first thought was, "Wait, is he trolling us?" Because auto-attack is the opposite of complexity and depth. Having your character do damage automatically means even if you as a player do nothing isn't a decision, isn't a rule, isn't... anything.

Oh, don't get me wrong, with a game that can have latency spikes, I think auto-attack is a useful tool to allow players to continue being mildly effective while they're having a lag attack (see the Nefarian solo kill video, near the end where I spike for a good 3 seconds), but if you were to take auto-attack away entirely, the complexity of the game and the depth of the game wouldn't change, meaning it adds literally nothing to either. And it's also independent of combat speed.

La Jeu en Rose

While Keen was clearly conflating complexity and depth in a couple cases, I think what he's really arguing for games as they were in Everquest. He views that as his ideal game, and that's totally fine. I don't care for it myself, but enjoyment is subjective. However, none of the arguments he posited support his initial hypothesis that slower combat means more depth. In fact, almost none of them have anything to do with slower combat at all.

#GameDesign

16 comments:

  1. I'm going to have to add on one comment about auto-attack.

    Player decision (or choice or agency) about autoattack comes in when you have to decide whether to let the game autoattack for you, or interrupt it in favor of doing something else.

    In some MMOs where white damage is numerically meaningless in the face of exponentially huge skill attacks, autoattack may not mean much.

    But let's take GW2 for an example, where autoattack is really skill 1 on autocast and the amount of damage it can do is substantial compared to other abilities. In a case like this, it can actually become significant to consider when you're autoattacking and how.

    Just off the top of my head, a warrior greatsword autoattacks slower than an axe, but may do a little more damage. I use very fast attacking sword and scepter autos (scepter is interesting because your range to a target also increases how fast each projectile hits) on my guardian a lot and find it very tough to get used to anything slower. Theorycrafters calculate those things to a precision that I can't be bothered with, but it's there.

    In DOTA 2, which I'm barely dipping a toe in, managing auto-attacking there is also part of the complexity and depth. Apparently knowing when to animation cancel the later part of your autoattack after a projectile is fired can speed up your movement and help you catch up to someone you're chasing. Certain gear and stats alter how fast some heroes can do autoattack damage - some of them up to very lethal proportions - and so on.

    ReplyDelete
    Replies
    1. Good points! The decisions there largely rest in the setup before combat, but decisions exist nonetheless. For WoW, auto-attack is largely negligible for reasons you've mentioned--to the point where Enhancement Shaman actually interrupt their auto-attack to hard cast Lightning Bolt all the time, which is still a decision technically.

      Both cases demonstrate a very high knowledge requirement/skill cap, and for GW2 different weapon types having different auto-attack characteristics does provide additional depth. But that's not based on the auto-attack itself. The depth is coming from the different weapons, which you could/should still get from your abilities interacting with said weapons (especially for GW2, where your weapon dictates what abilities you have available). The auto-attack itself could be changed to you manually spamming the '1' button with a faster/slower cooldown based on weapon speed and nothing would change (other than you having a very annoying button to spam). It's convenience, not complexity or depth.

      Animation canceling in DOTA 2 is interesting, and I think a better example of potential depth here. You have an event occurring on a cadence (auto-attack), and depending on when/how you interrupt that cadence, different results occur. Definitely a high skill cap/complexity, and I think it does add a little depth there.

      Delete
    2. I guess I over-summarized my GW2 example trying to get it in one paragraph. It's sometimes hard to articulate every last nuance that's in my head but another MMO player may not get.

      Eg. autoattacking in GW2 tends to face you in one direction to hit one enemy. Choosing to untarget the enemy and manually press 1 means you can position yourself at an angle that hits 2-3 enemies. This is situational as sometimes you can get the same effect with autoattack on, and sometimes not.

      Autoattacking in GW2 in certain situations, like having a bunch of confusion stacks on you (automatically does damage on each attack you do) or when the enemy has retaliation (automatically does a percentage of the damage you do back to you) or projectile reflection on, can be a recipe for committing suicide in short order.

      I suppose this situation would be closer to what you describe as complexity/depth as a decision would quickly have to be made to stop attacking and/or turn off autoattack to choose moments to manually attack.

      Not that I'm really disagreeing with anything, just coming up with more examples and ideas how combat can differ from game to game, and be complex or deep without necessarily being fast or slow paced.

      Delete
    3. Your case where you can rearrange your self to hit more targets makes me think auto-attack shouldn't exist at all for that ability. That's just bad UI.

      I think the mechanics you're talking about are interesting ones (i.e.: the confusion stacks, projectile reflection, etc.), but nothing there changes really if you had to spam '1' over and over instead.

      Your examples have been great to further convince me that auto-attack itself is a convenience, rather than a mechanic. Not necessarily a bad one, but still doesn't add anything to combat unto itself. I appreciate the discussion!

      Delete
  2. "I wish I had a video of me soloing Chimaeron on my Enhancement Shaman, because that was one hell of a puzzle trying to figure out what combinations of talents, abilities, timings, etc. I needed to win, but above is a video of the next fight, soloing Nefarian. While not as difficult, it still required the use of extant abilities such as Earth Elemental, Rockbiter Weapon, as well as all of my self-healing and defensive cooldowns to win."

    I'm sorry, but since you didn't have to solo Nefarian or Chimaeron in that manner while leveling and it was a completely optional challenge, they don't count. Game is too easy.

    "It's a trap. It definitely requires skill and knowledge of the system to handle, so is complex, but it doesn't add depth."

    I'm sorry, but threat is awesome...because reasons...so you're wrong.

    "That's not a reduction of depth, that's just a redistribution of complexity."

    I'm sorry, but you're clearly incorrect since I don't get a raid spot solely for being an Arms warrior and giving 4% physical damage.

    "It's another trap for the players. An unnecessary, unfun skill cap."

    Still wrong. Because...shut up, you're wrong.

    "Because auto-attack is the opposite of complexity and depth."

    Once you're reached the appropriate level of zen all will be clear. Then you will regret horrid statements like that.

    "In fact, almost none of them have anything to do with slower combat at all."

    /serioushaton

    Indeed, which is why I was so baffled. It's one thing to say "I found X, Y, and Z the most fun in Everquest and think it's the best" and quite another to say "I think A because of reasons B, C, and D which have basically nothing to do with A."

    ReplyDelete
    Replies
    1. Wow, that post really got under your skin :)

      But yeah, the disconnect between the hypothesis/title and the reasoning given really made the entire post not make a whole lot of sense to me, either.

      Delete
    2. Like I said, I really hate nostalgia and people who go "OMG MOLTEN CORE SO HARD HEROIC GARROSH SO EASY" or who confuse ridiculous complexity like five different resist sets with actual mechanical difficulty.

      Delete
  3. Good post. I still think, however - and I said this on Balkoth's post too - that the fundamental difference between your perspective and Keen's is in the baseline difficulty.

    "Leveling in WoW, Guild Wars 2, Final Fantasy XIV, and Wildstar all has to be sufficiently easy that the grand majority of the populace can perform it, or you lose your subscribers."
    "...you're not attempting difficult content to begin with. The content is there, it's just not in your face for subscriber-maintaining reasons. You have to seek it."

    This is what I mean. I think Keen is wanting to make that challenge in your face, for everyone. I don't think he agrees that it *has* to be sufficiently easy for the lowest common denominator, nor that it *should* be. I think he's aware that it would mean losing/failing to gain as many subscribers, and I think he's fine with that.

    ReplyDelete
    Replies
    1. That's a fair point, I think. His preferences are his opinion, and he's quite welcome to that.

      Interesting to note that Dark Souls, the canonical hardcore game, had 2.3 million sales after 2 years, which is pretty damn good for a console game. Not the best, mind you, however 2.3 million units is plenty profitable for an A-game, but would barely break even for a AAA game (see http://kotaku.com/how-much-does-it-cost-to-make-a-big-video-game-1501413649 for a good ballpark on costs).

      So okay, he'd prefer that. I don't see it happening for non-niche MMOs ever, given they tend to run $50 million+ for development costs (and that's the low end). They absolutely need to have the subscribers, or the game will fold.

      It's the unfortunate (neutral?) aspect of art and games ultimately being a business. You still need to make money. So I don't think his preference is realistic in an MMO.

      In a single-player game? Absolutely, especially if you can cater to the niche market you want to, like the aforementioned Dark Souls. Just unfortunate that MMOs as they exist today can't afford to be niche.

      Delete
    2. "I don't think he agrees that it *has* to be sufficiently easy for the lowest common denominator, nor that it *should* be."

      Which...has nothing to do with depth at all. Or slowness of combat.

      Delete
  4. I think the Trinity is a consequence of the level of our developers to produce responsive AI. We're just not good at that yet so the trinity continues to be the primary combat model. But as someone who's been playing the trinity for years, it's quite stale and boring. This is why games like GW2 and Wildstar are trying to put fresh paint on it, which is admirable. And tbh it works for those games, albeit limitedly.

    I'll be glad for the day when AI isn't the center of encounter design. If devs ever decide that AI responsiveness isn't the cornerstone of combat, we'll get more variety that doens't rely on the trinity. It's just hard to imagine what those games will look like, but probably some mix of God of War with Demon's Souls with WoW.

    ReplyDelete
    Replies
    1. Agreed entirely on the trinity as an AI replacement.

      As someone who has to write AI, writing AI is hard. It's relatively easy to write super dumb AI, and for some games it's actually relatively easy to write "perfect" AI depending on the rules--see Perfect Dark's Darkbot; assuming you weren't playing with explosives, at which point the AI broke down significantly. So even "perfect" AI is limited to a subset of cases.

      But anywhere in between is pretty difficult to manage. Your AI needs to understand the rules of the game, and how to use/exploit those rules. And when you have a system as complex as an MMO with so many abilities that even players discover certain combos months after the individual abilities were introduced, how can the programmer get AI to do interesting things?

      Basically, with the Trinity, or even a game like God of War and Demon's Souls, AI is prescriptive. The programmer tells it to do specific actions based on specific criteria. This style tends to make emergent behavior difficult to achieve. But it's easy to program! Things like threat in the Trinity make it look like enemies are reacting to players, rather than just performing their actions in a vacuum, so it's a useful (if staid) tool, as you point out.

      Perhaps one day someone will develop a neural net-type AI usable in a game that actually can exhibit "thinking" on its own. That'd be cool.

      Delete
    2. "see Perfect Dark's Darkbot"

      I loved that game. Need to see if I can play it on a PC.

      "Things like threat in the Trinity make it look like enemies are reacting to players, rather than just performing their actions in a vacuum, so it's a useful (if staid) tool, as you point out."

      Threat also acts as a substitute for blocking movement with bodies -- the fact that an NPC or player can rush right past you, leave their back open with no defense, and suffer no real penalty means threat is necessary to protect weaker party members. Theoretically the monks don't try to rush past the tank because they cannot or because it would leave them too vulnerable...but there's no collision in WoW and there's no such as thing "Oh, he left his back to me and is completely defenseless...*instakill.*"

      Delete
    3. Perfect Dark was soooo good. It really took the best of Goldeneye and just made it even more awesome.

      Not having collision with models is definitely a downside in terms of developing emergent behavior. I never player the Warhammer MMO, but I heard they had collision for player models to allow for tanks to physically block things?

      Makes me wonder how that'd work in a potentially high latency environment (think Aussies playing on NA servers with 250ms+ ping), or avoiding the ability to grief people by pushing them into a corner...also the additional server calculations would be astronomical. I really wonder how Warhammer implemented it, because the idea is sound (see any Tactical RPG or D&D 4.0 RAW).

      Delete
    4. Man, all this Perfect Dark talk...really, really need to see if there's a way to play it reasonably on a PC.

      I heard WO had collision detection, yeah, but I don't know the details. Latency definitely seems problematic and unless it's really, really good then it's worse than not having it at all.

      Delete
  5. This may be a little off topic, but there's another interesting element within the complexity vs depth topic that I haven't seen discussed. If you consider player focus as a resource, you can look at fights in terms of how, where and when that resource is spent. WoW has massively ramped up the complexity of rotations compared to the BC era, leaving relatively less focus available for things like movement and coordination, whereas TSW and Wildstar went the other way - very simple rotations and more hectic, mobile encounters. I really like the second method, as your rotation is mostly the same fight to fight, so shifting focus off the players' characters allows more variety and complexity in encounters. IMO TSW really got that balance right.

    ReplyDelete