MMO Development Lesson #6
Resist the desire to make frequent, minor tweaks. Any small tweak you make to a spell, an item, or anything will be perceived as a nerf if that tweak is downward, and will be perceived as negligible if that tweak is upward. Things don’t have to be perfectly balanced as long as they’re nearly balanced. For those minor tweaks upward, bundle them together as part of your major patch and the effect will be felt more significantly, and that’s also the time to make tweaks down if you have to. Note the word “minor.” 
You’re being a bit imprecise with your language here. For patches in general, you are right: it’s better to bundle a bunch of changes together. This is easier to hand off to a licensee in another market, and it’s generally easier to test. However, for changes to a specific system, it’s better to make frequent small changes in order to approach the mystical “balance” you speak of.
The truth of the matter is that we developers aren’t omniscient. This means that we have to make multiple tweaks in order to get something “balanced”. (And, even then, someone somewhere is going to be pissed off that you ruined the game with your stupid changes.) In my experience, it’s better to make smaller nudges in the right direction rather than making huge changes and hoping for the best.
You should be more precise in your language, especially if you consider this to be a “lesson”.
Ahh, perspective.
Any small differences between spells, items or anything else being compared between will be perceived as unbalanced, and will be accompanied by calls of nerf or boost.
Damned if you do, damned if you don’t.
There’s more than one side to this lesson.
MMO Development Lesson #6
Yep, it is a damned if you do, damned if you don’t scenario. But I am actually advising against what Psychochild seems to be implying in his second paragraph: In my experience, it’s actually the opposite of your statement.
If you make frequent minor nerfs, people perceive the nerfs as greater than if you make those necessary nerfs as part of a major patch.
This is coming from EverQuest II, where we’d make hotfixes several times a week, and they’d sometimes contain minor tweaks in either direction. Interestingly, if you nerfed a particular class 5 times over the course of a month by 0.5% each time, or nerfed them once by 2.5%, they’d perceive the former as a greater sleight against them.
Strangely, they’d also view one “big” buff to their class of 2.5% as better than 5 separate 0.5% tweaks. It doesn’t make a lot of sense, but that’s how it always played out from the community’s perspective.
Note: I’m basing these opinions (which all of my lessons truly are) on my personal experience. This one in particular is from the perspective of a couple years of Community Relations experience, watching fan reactions and the like.
Point of Clarification: We may not be arguing two different perspectives on the same point: While I was writing this, I was considering minor tweaks to multiple spells, items, etc. individually. I agree that you should really make small nudges in particular directions, but should bundle those small nudges into one patch instead of several. However, the optimal method would be to get a very active Test server community and have them help test your tweaks in a live environment, making the frequent small nudges there, then the final live ready nudges all at once on Live servers.
Screw “balance” - I’m tired of it. Jedis are SUPPOSED to be more powerful than correllian pirates and wookies. You just need to kill the Jedis young, before they have a chance to grow
Balanced gameplay on MMOs are a myth, and the publishers and designers should make an effort to explain that to the players, up front.
With the exception of bug fixes, especially client-side fixes that correct problems like crashing, falling through the world, and so on, I prefer getting big patches once a week or twice a month or whatever than hotfixes every other day.
Not so much because of the nerfage, but more because of the interruption it causes. It’s a lot easier to deal with it all at once on “Patch Day” when you know it’s coming ahead of time. When every other day is “Patch Day”, you start getting a bit more cynical.
On balance, I’ll say this. Every class or profession or whatever should have a specialty, something they do that other classes or professions can’t really do nearly as well. Whether that’s crowd control, or taking damage, or dealing damage, or healing, or whatever (to use a purely combat-related, and thus very limited example). The goal of “balancing” should be to keep the abilities of each class/profession unique and desireable, without allowing one or more of them to become overpowered to the point where it negatively impacts the others.
For PvP purposes, the best you can hope for is a rock-paper-scissors style of balance, where a mage will fry a warrior, but a rogue will kill a mage, and a warrior will send the rogue packing. If you try to balance PvP combat between all of your classes you end up with everyone basically being the same and having the same bag of tricks to use.
That would be Machiavelli, yes: hit them once, hard, because minor pains over time just keep them constantly angry. He does recommend a stream of minor benefits, which is how our modern politicians prefer to do it. Of course, balancing that guy’s class upwards can be declaimed as a nerf against [competing class], so it might make sense to wrap it, too.
Of course, if you change everything at once and the balance no longer works, how do you diagnose which tweak did it?
As Ryan touched on, the best approach is to figure out a way to encourage a segment of your population to actively participate on a Test Server. Your test server should be patched and tweaked on a near daily basis… then at a certain point, you let it sit stable a week or two, then roll up the changes and push them live. In an ideal world, this is how it would work, with a patch going live every 2 months (6 weeks of tweak, 2 weeks of stable and polish, or something like that).
The question becomes: how do you get a consistant active “good” test server population?
For new games, you could cultivate them from Beta… invite the most active and helpful testers to a free account that has access to both the release version of the game and a private test server, for as long as they continue to play on test… if they go a month without notification of playing but not playing on test, their account can go back into regular pay mode. You could also get groups attached to fan sites, getting a group of testers and simultaneously working with them through a limited NDA to release info about upcoming changes. Those two ideas among other things might help…
For me, what I dislike is the lack of communication that often comes with nerfs. Rarely is there any reason given for why a change is being made, and in many cases, it seems like changes are made without any regard to actual gameplay. If a class, ability, item or whatever is legitimately overpowered, most players recognize it as such and don’t mind when the nerf comes.
Another thng that annoys me is that abilities, even ones that might be legitimately broken, are nerf months or even years after being released. If an ability is nerfed a day or two after beng introduced, most will understand, but delays on taking action are seen as tacit approval of the ability in its current state. When the nerf finally does come, it is seen as a major change in how the class works. However, by this point, these classes with their abilities have already been incorporated into groups and raids, so losing that ability is a major blow to the functionality of those classes.
*All of these are based on my EQ2 experience.
Lets not forget, that even if your “tweak” is correcting a clearly evident error with something, some portion of players will perceive the lack of corresponding bonus to their focii to be a form of nerf or an imbalance.
The whole concept of ‘immersion’ means that players will rationalize the flaws in your product, and sometimes the urge to set things straight needs to be weighed against how an error has been absorbed into the games’ culture. Like feigning in EQ1.
And I have to wonder if balance is even attainable in an MMO because nobody (generalisation, obviously) is happy unless they have the *larger* half of the cake. What’s the average PvE K:D ratio in EQ2 or WoW?
Actually, I see a convergance between what the original post and what Brian was saying, Ryan; if you take your philosophy of balance-needn’t-be-perfect and his of err-towards-caution then you will make smaller tweaks [first].
Obviously you should be fully dilligent in trying to make the changes as informedly and accurately as possible to avoid their being a second tweak, but where there is a margin for error, you should err in favor of the players. Cleric’s heal is somwhere between 3-5% too powerful? Reduce it by 3 or 4% rather than going 5% and then having to backtrack.
As a Bard in EQ1, my class was continually being indirectly tweaked (bard spells being reduced-combinations of two or more spells from a wide variety of other classes) and directly tweaked (when those changes made the bard too powerful or completely broken).
I spent months doing plane of hate 2-4 nights a week when an adjustment had over-zealously limited the rate of bard armor drops to effective-0 (it was fixed 3 weeks after I quit).
At one point, one of our guild members was under GM/Guide scrutiny and during a hate raid several GMs/guides turned up to watch. We thought we were being given a great honor until remarks were made implying we weren’t doing the zone “as designed” (I’m sure our min/maxers had found some loophole that was allowing us to avoid frequent wipes *shrug*). And then, somehow, “Inny” agrod on us under peculiar circumstances while we were fighting for our lives.
On impulse, I charmed him, and made him kill the person who’d “trained” us, before setting him to mopping up some of the adds. *That* pissed someone off enough to reset the zone, booting us out, and it seemed only a few weeks before in went the bards-cant-charm-squat patch.
These are, of course, my personal perspectives on events, as a Bard. I saw the Taketh Away in harsh black and white. As a player, I could only see hardships. If the approach taken had been a combination of your balance-imperfect and Brian’s softly-softly, I would certainly wouldn’t have burned out so badly on EQ1.
I think I would have been more ameniable to a gradual toning down of those higher-end abilities, allowing me to see how little I was actually losing, rather than having lots taken away and then being “told” that some of it was being given back.
It didn’t happen that way, so I’ve been able to retain my noble self-image
In my time at WWIIOL I’ve definitely learned that tweaking is to be avoided at all costs because stability turns out to be more precious than fixes sometimes, and so I have to question whether my noble aspirations to being an acceptor of babysteps is genuine or just unproven.
Those might not be the types of balance issues you were thinking of, but are you proposing that devs should stick with their tweaks regardless of which side of the mark they landed over accepting a tweak-too-far and backing down a little?
PS - I would probably have forgotten the Inny incident, but the same GM had commented on my fearing Naggy one night, done out of pure survival instinct rather than malice, saying “you can’t do that” and when I remarked I’d done it to Vox the other night he said, “yes, I noticed that” and shortly afterwards - next week - a no-fearing-dragons patch went into test.
I have no concept how prevalent that was, and since what it got me was Vox falling into a pit and then summoning me for a private audience and a VERY confused guild wondering where the dragon had gone and nobody believing I had done it, it didn’t seem like a “well known” or “widely used” caveat. After the patch, I only ever heard people say ‘could bards do that?’ or bards say ‘it wasnt possible anyway’. All in all, it might just be my naievete, but you can see how I might just possibly have an inkling that I was somehow participant if not precipative to those adjustments.
Undoubtedly it was coincidental, and my story is annecdotal and the experience of having innocently triggered a reprimand common to other players who had similarly done the ‘impossible’. Certainly the outcomes of the engagements weren’t changed any more drastically than one of our enchanters rolling a success chance to land a mez uninterrupted or a cleric managing to get a heal out. But the resulting ‘tweak’ effectively removed a line of functionality - from the class for the best part of a year.
Always tweak the interactive model towards the better, are the controls confusing, is the camera feeling wobbly? Maybe even your fotstep sound dosnt provide a satisfying feeling of walking or running. These kinds of tweaks will improve the feeling of the game and positive tweaks are always good.
Most mmorpg’s also make a shoddy job of these aspects of the design and I dont know why as the interactive model is the thing people will judge your game from. Abstract things like stats and skills are important but not as important as a good feeling from using the parts of the game directly.
This brings up a Question, then. I’m curious — are most of these tweaks & their results monitored more through a “feel” or hunch… (e.g. - it takes a few weeks of gameplay to check it out) or are the developers able to create automated metrics/measurements to determine the validity of a tweak? Or various combinations of the two?
Mmmm. I think the lesson is pretty well taken. And yet I have a very good reaction when I see constant patch notes, correcting one thing or another.
Maybe the trick is not so much to patch design changes often, but bug fixes. Nerf in groups, but fix in bits and pieces.
I think it’s partly because you know the numbers and we don’t. How many patch notes actually say “Monk’s evasion reduced by 0.5%”? For people like me [that aren't going to do hours of combat log parsing] without numbers all we see is 5 nerfs versus 1 nerf.
In some cases it might be visible in the GUI someplace but then we get used to display bugs as well…
kfsone: “where there is a margin for error, you should err in favor of the players.”
Yes, please.
Paladyne: A combination, really. The problem is that a lot of developers base their tweaks almost entirely on numbers rather than feel, and they really need to put a greater emphasis on the feel in addition to the numbers.
Bob Kelly wrote:
Balanced gameplay on MMOs are a myth….
Not really. You just have to realize that “balance” doesn’t mean that everyone’s equal, rather it means that all choices are equally appealing in one situation or another. To use your Jedi example, things are balanced if playing a smuggler is as appealing for some people as playing a Jedi. This doesn’t mean you have equal numbers of smugglers or Jedi, but that you have enough smugglers to fill out the world. If 90% of your players play Jedi and the other 10% are spread out over a dozen other classes, then you probably have an imbalance in the game that needs to be addressed.
The trick is that this is still very subjective. Your design could call for having 90% Jedi; in this case, there’s no balance issue. But, if you originally planned on having 30% Jedi then your game probably isn’t working in a way that’s fun for most people.
Paladyne wrote:
This brings up a Question, then. I’m curious — are most of these tweaks & their results monitored more through a “feel” or hunch… (e.g. - it takes a few weeks of gameplay to check it out) or are the developers able to create automated metrics/measurements to determine the validity of a tweak? Or various combinations of the two?
Deciding to fix something usually comes from feedback (read: bitching on the forum) or observation. The extreme version of observation is Oliver’s examples above where the developers get upset that someone is doing something too easily and jump to change the game to prevent it.
When it comes to actually making the change, it’s almost all hunch from the stories I’ve heard. This is one reason why I advocate multiple small changes instead of a big one. I know I need to buff/nerf a power, but how much? It’s interesting to hear Ryan talk about people using numbers, since data mining seems to be a relatively unknown thing in developer circles.
Ryan Shwayder wrote:
Interestingly, if you nerfed a particular class 5 times over the course of a month by 0.5% each time, or nerfed them once by 2.5%, they’d perceive the former as a greater sleight against them.
Strangely, they’d also view one “big” buff to their class of 2.5% as better than 5 separate 0.5% tweaks. It doesn’t make a lot of sense, but that’s how it always played out from the community’s perspective.
The problem is that even if the power “needs” to be nerf 2.5%, few developers can hit those numbers with such unerring accuracy. Here’s what happens in most games:
Patch 1: Power is nerfed 5%. Players bitch that the power is now worthless.
Patch 2: Power is buffed 4% in response to player complaints.
Patch 3: Power is deemed still too powerful, nerfed 3%.
Patch 4: Power nerfed 3% again, someone was using outdated information.
Patch 5: Power buffed 6%, players complain that power gets too much luv.
Patch 6: Power nerfed 2% because a dev was PKed using the power.
Patch 7: Power buffed 3% because another class power was nerfed.
Now, I’m not saying that a slow, steady increase will eliminate all these problems, but you’ll avoid situations like this that happened all too often in the “bad old days”.
The further problem is that you’re measuring success by how much people complain. This is a foolish measurement for a designer (but a very nice dream for a community manager ;), because players are going to complain no matter what. The goal is to let the most number of people have fun, and making small changes tends to discourage negative behavior while not punishing people that don’t engage in such negative behavior.
Besides, there’s only one way to get players to not complain (most of the time): don’t make a game in the first place.
Psychochild brings up a good point: Since balance is really a matter of opinion and relative appeal, you should almost strive to “balance” the game such that certain classes are appealing to certain playstyles. Which means that some players who think they want to play class X based on its name or general capabilities may not actually like that class because you made it for a slightly different playstyle.
On the numbers thing: I’ve only worked on two MMOs so far, and EverQuest II used a lot of numbers (metrics from actual play as well as internal formulas) when balancing mechanics. I’m not sure what industry standard is because of the limit scope of games I’ve worked on.
The concept of balance has so many different meanings that statements for or against this comment hinge greatly on the title and its mechanics that you are appling it to. In MMOs that contain PvP, balance relates very specifically to how a given class or class ability can impact its opponants. If I possess a CC mechanic that dominates other players to the exclusion of their participation in the combat, imbalance ocurs and players will complain. In a non-pvp situation, players will not mind that one class can keep npcs CCd for infinate periods of time. In fact, nerfing the ability across both PvP and PvE applications will serve as a nerf for both the individual class in the PvP space, and to the playerbase a whole in the PvE space as it effects their ability to tackle group or raid content.
Thus it is critical to identify the specific application or manipulation of a game mechanic that causes an imbalance and address it in a fashion that mitigates the unintended mechanic without wholly altering the original intenstion of the mechanic.
Example: Warlock Fear in World of Warcraft. This was designed with PvP and PvE implications in mind, but it was quickly seen to have very drastic negative effects on PvP situations, while serving its intended function quite well in PvE situations. The resulting ‘nerf’ was to afjust how it was resisted or applied to other players while allowing it to preserve its initial design for non-player charachters. This was a well applied nerf. People still bitched but its the MMO space, and everyone knows MMO players are neer going to be happy unless their class is the only overpowered one in the game.
I am Scissors, Paper is fine… NERF ROCK!