Instance Scaling: Why Not?
In response to Garrett Fuller’s question about instance scaling on MMORPG.com: “My main question is how hard is it for a developer to scale an instanced dungeon based on the number of players who enter it?” The answer is that it is not difficult to do. The question he should have asked: “How hard is it for a developer to scale an instanced dungeon and make it fun and rewarding no matter how many players enter it?” The answer is that it is impossible without a lot of hand-tuning, to the degree that if you tried to scale it for every combination of players, it would be better to just create completely new content for several new instances.
Why? From the wording you use it seems to have to do with different class combinations, what is it about the different combinations that gets so unbalanced?
The biggest problem, I assume, is with tank & spank. If you design an instance for a group of 6 and assume they have have a tank and a healer, you can make content that is challenging for nearly every group makeup. But if you allow 12 people into the same instance, how do you scale a single tank challenge? In most cases, it’ll remain single tank capable, and the added DPS of the extra 6 people will just make the content trivial.
Just look at EQ1 before instancing and raid limits. Yes, you could fight a dragon with 5 groups of well coordinated people and it was a challenge… but if you brought 10 groups, as long as lag didn’t mess you up, the same dragon became a near guaranteed win. If you brought 15 groups, and again if lag didn’t cause issues, it became trivial. Should the game spawn a second dragon then? a third? How do you scale that encounter?
Plus, you get into too many factors… if 5 groups of well equipped people can kill the dragon, and 15 groups of well equipped people can kill three dragons, but it takes 15 groups of poorly equipped people to kill one dragon, do you now have to scan the players and scale on equipment as well as numbers?
To make it fun, you have to essentially redesign the encounters/abilities/etc. for every number of players. If WoW designs an instance for 5/10/25 men, they’re doing the mob creation, population, and balancing three times. The scripted sequences and such will usually remain intact, but it is non-trivial to even do that one portion of instance design three times. If you add more combinations of players to the mix, you’re adding a significant amount of time.
I don’t know how much time it takes them to make an alternate version of an instance, but I’d guess 1-3 weeks for each version of the instance beyond the first. The two alternate versions probably add about a month to the creation of a new zone. Drop in two more versions and you’ve added two months. And so on.
I also like the social implications associated with finding other players to fill in the gaps of your group. Say you have 3 players to do a 5-man instance. You need to find 2 more. If the instance scaled to 3 players, you’d just go right in. If you needed 5 to complete it, you’d have to meet 2 new people (or find 2 you already know) to complete it, which forms a stronger social bond.
Dungeon instance scaling would be great, but I want to see “World Zone” scaling. Zone overcrowding is annoying and has been a problem with every MMO for the greater part of a decade. When new MMO’s launch thousands of players are thrust into the same starting zone as everyone else, all of us competing to those 5 boar tusks, grain bags, etc. to complete our quest.
Some MMOs, EQ2, AoC and Aion have tried to solve this problem by creating different instances of the world zones to compensate for overcrowding. In my opinion, this merely dilutes the immersion factor of the game and makes it a pain in the ass to group up with your real life friends who are all scattered across 20 different version of the same area.
I want to see world zones scale based on their current population. The overall scale of the zone increases based on the amount of players around as do the number of quest mob spawn points to accommodate the # of people in the zone all doing the same quest. The only game I have played that does something like this is Battlefield 2. The same map exists in 16, 32 and 64 player mode, all scaled accordingly. Of course, those map sizes are determined before matches begin. The challenge would be to create dynamically scaling zones that shift automatically.
Dunno if it would work, but it would be neato.
Cheers
What about designing instances such that there are a mix of challenge levels to the encounters – such that early trash and bosses can be taken down by a non-optimal or sub-sized group, but later trash and bosses require a full group. Perhaps there could be different pathways through the instance, one with lots of weak trash and one or two weak bosses, and another which is the shorter but more deadly route to the big boss. Provide appropriately scaled loot.
In some way this is already done, what with later bosses being marginally harder than the earlier bosses. Or so I hear – often the real difference is more due to different boss mechanics/gimmicks, where the group that wiped on Grobullus over and over totally pwns Kel Thuzad.
Some games have scaled based on population by increasing respawn rates. It works to some extent. While scaling a zone’s actual size would definitely be great for launch times, it would be wasted effort within a couple weeks of launch. Unfortunately, I don’t think it is practical to spend the amount of effort required to really hand tune overland zones to get bigger/smaller/have more quest target locations when they become virtual ghost towns shortly after launch anyway.
“Say you have 3 players to do a 5-man instance. You need to find 2 more. If the instance scaled to 3 players, you’d just go right in. If you needed 5 to complete it, you’d have to meet 2 new people (or find 2 you already know) to complete it, which forms a stronger social bond.”
That’s one possible outcome. Another outcome is thinking to yourself ‘why am I standing around spamming LFM in this game when I could actually be playing in another game’.
I do agree that it would take a lot of hand tuning to make and instance scale in both number and challenge. Most of the reasons for this is the impossibility that is character balancing. If you made all the classes with the same abilities, people would call the game boring and it would only last a couple months. So we have games where every class has their own nitch to fill. If you group with the wrong “nitch” you end up with a gaping weak spot . If the zone you are trying group up for has elements that capitalize on that weakness, it can make it a no go for the group. Where a group of the same number, with strengths where the other group had weakness could blow right through the zone. Sure you can always say “well wait for the right group make up to do it,” but if the games been out a year, the cap is 50, and your level 20 doing a level 20 instance, you may not have the luxury of waiting around for the “right” group. The more realistic outcome is like JuJutsu said, I’ll go play a different game.
“That’s one possible outcome. Another outcome is thinking to yourself ‘why am I standing around spamming LFM in this game when I could actually be playing in another game’.”
It all depends on how important random social interaction is to you. For me, I pretty much hate every MMO to come out since WoW (and including WoW) because they are all so solo friendly to the point that no one needs anyone, until they raid, and so the only people I socialize with are the people I bring with me, my real life friends and people I met in older games that were not so solo friendly.
“It all depends on how important random social interaction is to you.” Not important enough to stand around spamming LFM. Luckily, most modern era games don’t hold me hostage to the willingness of strangers to allow me to play the game.
*shrug* To each their own… if I play a game and can get to level 10 (or the game’s equivalent in skill training, etc) without ever grouping with another player, I quit. In my opinion, there is absolutely no point at all in paying to play an MMO if you are going to do it alone.
Don’t bait and switch random socialization with pure solo play. If you refuse to play a game that someone can solo that’s your choice, but not playing with strangers is not the same as not playing with others.
Some really good discussion here. I think JuJutsu and Jason point out the two major play styles in mmo’s today. To me if your going to design a sucessful mmorpg you need to allow both kinds of play to occur. Reward players with better gear and loot for grouping in lower level instances but don’t make it the only way in the game to level or get gear.
WoW did this but it could have been far better by doing two things. One was not setting a level limit on lower level instances and the second was not introducing mentoring. So instead of grouping and experiencing the dungeon the way it was ment to you get a high level toon to run you through and by pass the challenge while still reaping the rewards.
With level restriction and mentoring in place it allows a far greater possibility to put a group together to experience the instance as it was intended. The mentored player levels far quicker as a % of exp from the mentor goes to the apprentice. The one thing that needs to happen that I have not seen is to put some kind of rewards on boss mobs that the higher level players who are mentoring can use towards gear at their level. (Ie insentive for them to particiapate.) An example would be putting the same badges ( currently conquest badges ) that high level player would get from a doing heroic dungeons at level 80. Of course this would also mean putting a lock out on the dungeons so that gear and badges could not be farmed.
As far as instancing goes I think the way SOE did it with EQ2 was the best way to address high populations in any one zone. Does this break immersion or makes it a pain to group? To some extent but I think people figured it out pretty quickly and while it might break immersion no better system has been created to address over population. How is your immession any better when 20 people are standing around waiting for a mob to spawn so they can kill it to complete their quest?
Are you only teaming up if you are forced to do that? I see nothing wrong with being able to play solo, teaming up should be an encouraged activity, not enforced. Offer a somewhat different and possibly better experience with a team than solo but allow both.
City of Heroes/Villains has done an excellent job doing that encouragement IMHO. Of course the ability to scale has in that case go hand in hand with allowing people to set their own difficulty setting; without that it would certainly end up with the excessive work in tuning the encounteras as mentioned. And that also fits in with how flexible character development and team set-up work out in the game, XP gain and many other pieces. No single feature works in isolation.
CoX is the MMO I have teamed up the most in by far, even if other MMOs has done more to provide the “need” to team up.
@JuJusu Its not a bait and switch. I’m talking about random socialization, which means meeting people I didn’t plan on meeting. By definition, if I am playing with people I already know, it is in all likelihood not random.
The problem I’m talking about is that when a game is too solo friendly and doesn’t do much at all to encourage grouping, people in general tend to not group. So the newbie zones are largely full of people running around not talking to each other and soloing the content. If a game is introduced that way, people will tend to continue that trend: not talking and playing solo. Yes, I can invite my real life friends and online friends I’ve met in other games to join me on the same server in the same game I am playing, but that isn’t random socialization, that isn’t meeting new people.
Yes, yes, yes… I understand that there are plenty of people who like to play “alone together”, to solo content while existing in a world with other players so they can make Chuck Norris jokes in global chat channels and laugh together without ever actually grouping up and doing things together. But I have no interest in that… I’m in an MMO to play with people, not to be a lone wolf running around the other people.
We disagree. That’s awesome! But I think the original EQ was one of the best designed games ever because I have a few dozen friends now that I wouldn’t have without that game. My attempts to play in a similar fashion and make new friends in WoW largely failed because the game was too solo friendly at the bottom and too guild-centric at the top. Its too easy with too much instancing with too few reasons for people to just random NEED each other.
The thing is, even in EQ I never stood around spamming LFG or LFM, held hostage to the willingness of strangers to allow me to play the game. I traveled around, talked to people, played in sub-optimal groups in shared zones, making connections that more often than not payed off later in invitations to group and people to invite to my groups. Sure, there were times I couldn’t put together the “right” group to do something I wanted to do, but I could always find something to do. Whereas in WoW I pretty much could always do exactly what I wanted to do because most of it didn’t require a group at all, and 99% of the time I couldn’t find anyone to group with anyway (most people actually respond with “Haha! You don’t need a group for that!!” or some variation), and it was f-ing boring.
“I traveled around, talked to people, played in sub-optimal groups in shared zones, making connections that more often than not payed off later in invitations to group and people to invite to my groups. Sure, there were times I couldn’t put together the “right” group to do something I wanted to do, but I could always find something to do.”
Generally great fun, sounds like what I enjoy the most. I think EQ is a really bad example though. I think EQ has to be the most ‘unforgiving’ game ever made. In my experience a ‘sub-optimal’ group meant someone was going to die. And lord knows we mustn’t have a weak death penalty, death has to hurt and hurt badly.
I like exploring, EQ was not the game for it even with feign death to smooth the way. I think EQ was far and away the worst designed game of any that I’ve ever played. But you’re right, tastes vary
Sub-optimal only means that some people insisted on a cleric and a warrior (and an enchanter). The most fun I had in that game were in groups with Paladins or Druids playing the healing role, and ranger tanks, or even monk tanks, other forms of crowd control, etc…
And personally, I like steep death penalties. I don’t need perma-death or item loss (PvP with full loot is more annoying than fun). But, in my opinion, when you take the sting out of losing, you diminish the joy of success.
Its funny though… I love exploring, and to me EQ was the best game for it. So yes, tastes vary.
Maybe doing it for ANY possible number of people is impossible but I’ve already seen it done plenty of times pretty well for 1-18 people. (See City of Heroes) Perhaps when the numebr get too large it gets ridiculous so just cap the scaling at 18 or some large number.
In a strange coincidence I just penned an article entitled Why Scaling Challenge Should Be the Future of MMO Content. I’ve looked at the advantages and disadvantages of the current status quo and instancing methodology that scales to the number of players that you bring inside.
Sure it’s not going to be easy to implement but the rewards of social cohesion and player inclusion outweigh the cost in my opinion.
Frankly, having a lot of other people wandering around a zone breaks my immersion. The only time I want to see other players is in town or just before heading into some kind of instanced “heroic encounter.” Even then, I’d much rather do an instance alone or with a small group of people I already know. Pick up groups have generally been more frustrating than fun in any of the half dozen MMOs I’ve played.
Here’s the perfect example of what I think ruins a game. I beta-tested Vanguard (no instancing at all). The coolest newbie area was, IMO, the orc area. You start as a galley slave, escape, make your way through the front lines to a friendly camp, and get enlisted as a sort of “special forces” soldier. One of the earlier missions involves sneaking into an enemy camp and assassinating the enemy leader. All very fun and heroic until you get to his tent and find that there’s a line of people waiting to kill him. I can’t put into words what a lame feeling that generates.
I understand there are people who want to play in a group. I don’t understand why they would want to submit themselves to that sort of frustration anymore than I understand why some people enjoy skateboarding or skydiving. But hey, as long as I can do what I want to do in the game, more power to them for playing it their way….
I enjoy playing in groups. But not all of the time. And, usually, not of the size that designers deem acceptable. That’s why the topic of scaling is of special interest to me [note the crude attempt to veer back on topic
]. The very best group size for me is 2. In games that my wife is interested in playing the two of us adventuring is the best gaming has to offer. She won’t play EVE with me now but I have a RL friend that fleets with me.
That’s what I’d like to see more of, content that scales down to 2 or 3.
Of all the games out there, EQ2 has done the best job of scaling instances, IMO. At the very least, SOE has made an effort to scale for 1, 2-3, and 5-6 in many places. Turbine has also at least acknowledged this dynamic by tagging their quests as “solo,” “small fellowship,” “fellowship,” and “raid.” Granted, they haven’t been very good at judging this (most small fellowship stuff is easily soloable, even when the quest is orange, while most fellowship stuff is either completely soloable or utterly impossible, depending primarily on whether it’s overland or instanced). I guess that just goes to prove Ryan’s underlying point, that scaling encounters is difficult ;^)
Even so, given the amount of effort that goes into level layout and art assets, I’d still have to argue that it requires less effort to scale a single instance to multiple group sizes and compositions than it is to create an entirely new instance from scratch. I’m sure it’s mind-numbingly dull for the developer involved, though….