Ghostcrawler: The Exchange Rate

Quote from: Ghostcrawler (Source)

How do developers decide what to change and when?

My last two blog posts explained various upcoming changes. This is not going to be one of those entries. If what interests you the most is the news about World of Warcraft, and not so much the design process behind the game, you can skip this entry without problem.

A lot of game design is about finding a balance, and by this I mean not only making sure that all the different classes and other aspects of the game are fair enough, but also how easy it is to go to one extreme or the other. . You also have to find a balance in the amount of changes you make. At one extreme, it would be not to change anything, with what seems that the game is obsolete, and it is understandable that players are frustrated when they verify that no one solves those errors or problems in the game that arose long ago. At the other extreme would be excess change, which can cause what we often call the 'roller coaster effect' because the game design seems unstable and players, especially those who play more sporadically, cannot keep up. Today I would like to explain a small part of our philosophy when it comes to change, how much is too much, and when we consider that change is necessary.

First, some basic technical concepts

World of Warcraft is a client-server game. The servers (which are our teams) manage important regulatory elements, such as combat calculations and the loot system. There are a couple of reasons for this. First of all, it makes it much easier to share information between groups. When a rogue stabs your priest, it is important that both your computer and the rogue's computer agree on when or where the hit happened and how much damage it has dealt (and what procs have been triggered as a result, etc.). Second, the server is much more reliable than a personal or public computer can be.

Over time, as our team of developers has gained experience and recruited more experienced engineers, we have been able to do larger, and sometimes bolder, server upgrades without having to upgrade your client as well. Updating the client (the game on your computer) requires a patch. It could be a large patch, like 4.2, which introduced the Magma Front quest area and the Firelands gangs, or it could be a small patch, like 4.2.2, which fixed some bugs. Client patches are a bit tricky. They take a long time to build and test, and they carry some risk, since if we screw something up, we have to release another client patch to fix it. Now it is much easier for us to modify the game code on the server. It also carries some risk, but it's much easier for us to fix any bugs. We call these changes on the server live fixes, because we can often apply them even while you are playing. If we were to introduce a live fix for Mortal Strike damage, suddenly, in the middle of a fight, it could change to more or less damage. Sometimes if we haven't announced the fix live yet (or in the rare case that we choose not to announce it at all), players will call these types of changes "silent buffs or buffs." We generally can't modify things like images, sound, and text with a live fix (at least for now), so for example we won't add a new boss or change the appearance of a weapon without a client patch (although yes, we could use the live fixes to activate a boss that was previously added with a client patch).

I comment on all this to explain that one of the reasons why you see so many corrections live today is that we have the technical capacity to do them. That's not to say that the game contains more bugs, more clumsy design decisions, or more class balance issues than before. It just means that we can now fix those issues in the moment, when in the past we would have (and would have) had to wait months for the big day of the next patch release. In general, it doesn't seem fair to players to keep everyone waiting for things that are easy for us to correct. Whether players like the change or not depends largely on the nature of the change. If we fix a class ability that was buggy, players playing that class will often appreciate the change ... except if the correction decreased damage taken, or if they had to swap gems or enchantments to benefit from the newly repaired ability.

Great power implies ...

This is the challenge here. If your hunter is nearly maxing out on the counters you might ask: what's the rush? And many players do. But keep in mind that other players are annoyed when their raid leader dumps a warlock to bring in a third hunter (because their damage is formidable) or it could frustrate their high probability of losing to your hunter in PVP . The degree of "need" for a change depends on one's perspective.

We try to gather a lot of voluntary information from players, for example when they cancel their subscriptions, about the reasons that led them to feel this way. As time goes by, we see that concern regarding class balance has diminished and concern about frequent changes in the game has increased. There is certainly the risk that we change things too much and in doing so we scare the players. The roller coaster effect of too much change can tire a community, even if each specific change has a noble goal. We have to weigh our goal of providing corrections when they seem justified to us and the fatigue players can feel from having to learn the mechanics of the game over and over again. We constantly have discussions about whether or not a certain change is necessary immediately, or if it is a problem that we can live with for a long period of time.

There is no set or quick rule to help us resolve these conflicts, so I found it easier to give you several examples of the kinds of things we would or might not be tempted to change in a live fix, patch, or a expansion.

First example: The parity between specializations

After looking at a lot of raid analysis, we concluded that Arcane Mage's damage routinely outpaces Fire Mage. (Right now I am leaving aside many of the elements related to this discussion, in order to try to focus the scope of the decision on something that can be explained in a simple way). For example, if Fire is better than Arcane in area of ​​effect fights, this is a factor to take into account. If the Fire spec is more difficult to play, or inherently more random, that's another factor to consider as well. Even leaving aside all the confusing aspects, this is still a very complicated task. Ideally, players who like Fire should be able to play this spec without feeling like they are a hindrance to their friends.

The extent to which Fire can lag behind Arcane while still being "viable" depends on many factors. Some players find a difference between the damage dealt by both specs below 10% acceptable. Others will change their specialization for a theoretical gain (that is, that has not even been demonstrated empirically) of 1%. If we saw the possibility of making some numerical adjustments to Fire with the confidence that they would place it at the level of Arcano, we would consider that we owe this modification to the players.

But this decision carries certain risks. If the Fire buffs could make this spec more dangerous in PvP, we'd have to be very careful with the change. If more mages switching to Fire meant that some raid buff or utility provided by Arcane mages became harder to come by, we would have to be careful about the switch. But the worst outcome, in our opinion, would be to exceed our goal. If this were to happen, players interested in Arcane would feel compelled to switch to Fire, potentially needing to swap gems and enchantments or reforge, and might be angered by having missed a certain item that fell the previous week. This puts players in an awkward position.

When players mention that they feel like they are on a game mechanics roller coaster they are often referring to this. Last week Arcane was the most interesting spec to play with. And maybe, before that, Frost was. Next week, who knows what it will be. In the past we've screwed up all the way in this regard, at times when we thought we were improving parity between, say, Hunter, Warrior, or Death Knight specs, but we actually brought players to feel that they had to change specialization. Given enough time, we can get pretty good results on balance adjustments, but live fixes, and often even patch changes, don't always provide enough empirical experience.

Remember, it's not about seeing how much damage the Fire mage and Arcane mage inflict on a target. What the players (and us) care about is seeing how they manage in specific encounters, taking into account the experience of a wide selection of players with varying degrees of skill, different raid formations, and continual changes. on gear, other players in PvP, etc. We often take bigger risks when there are major differences in play style. It is more difficult to ask an Enhancement shaman to change to Elemental than to ask a Demonology Warlock to change to Destruction. Players who are very fond of Demonology may not find this fair, but we must take into account the risks that even the smallest and in principle safe changes represent for the game and the entire community of players.

Second example: The creative use of game mechanics

There are a lot of smart people working on World of Warcraft, but we have no way to compete intellectually or creatively with the combined efforts of millions of players. Despite our best efforts, players have terrifying ingenuity capable of coming up with creative solutions that we had never thought of before. Here we find a wide variety of examples: a player finds a trinket, a weapon based on the proc, or a very old set bonus that does very well in new content. A gang hatches a strategy that makes a boss much easier to beat than we expected. An arena team finds a way to manage their crowd control or burst damage that is nearly impossible to counter.

Much of the fun in World of Warcraft is about problem solving. In general, our philosophy is not to punish players for being creative. Whenever we can, we try to give groups the benefit of the doubt. If a boss turns out to be easier than expected because players band together when we expected them to spread out, or manage to crowd control aggregates with higher power than we thought they would, we simply congratulate the players silently for your ingenuity. If a boss turns out to be easier than we expected, we may act on it. (Although we generally add a lot more debuffs than buffs to encounters in live fixes and patches.)

We are more likely to act in situations that force players to behave strangely, especially if it is something they are not going to enjoy. If to complete a raid they consider that they must go find a specific trinket from among very old content, or leave six players on the bench to enter a specific specialization that has an ability that simplifies the fight, then it is likely that we will have to act . These kinds of changes are very subjective and involve a lot of internal discussion. Remember that our litmus test is usually: "Are the players having a good time?" And not "Are they doing something we didn't have?"

Third example: The difficulty of an encounter

When it comes to encounters, the decisions almost always come down to whether or not to make a live correction. If we wait for patch 4.3 to make major changes to encounters from 4.2, when the player's attention has already shifted to 4.3, that is probably not a well-spent development time. When launching new dungeons or raids, our initial philosophy is just to keep all the nails at the same height, which means pulling out the short ones a bit and hitting the ones that stick out too much with the hammer. After a week or so, it is rare that we improve a match to make it more difficult. We tend to bundle several of these mods, often at the start of a new week, so that it looks like a kind of mini-patch rather than a constant stream of boss debuffs.

For the bands, we look at the graphs that indicate the number of new players who have made it through a specific match each week. The slope tends to be steep at first, when the more experienced gangs grapple with content, and then softens as other players advance. When the line turns horizontal and no new player gets past the content, it's time for us to jump into action. It's a bit easier for 5-player dungeons, because we want players to win most of the time. Nobody wants to return week after week to the Throne of the Tides until they finally defeat Lady Naz'jar.

The stats we look at the most are the number of attempts it takes to beat a dungeon boss, the number of kills the boss causes, and the time it takes to complete the dungeon. At the launch of Cataclysm there were bosses like Ozruk of the Stone Core who stood out for their strength. Sometimes we can manage these changes with a few simple adjustments (such as reducing boss damage) and other times it is necessary to modify the mechanics of the encounter as much as we can through live corrections, which actually give us a good margin of control. maneuver, since almost all information about the creature is on the server.

Fourth Example: Changes in Class Rotation

Here we find a couple of subcategories: voluntary and involuntary changes. We often make corrections to make a certain class more fun to play with. Allowing Weapons warriors to reset Rend without having to constantly reapply debuff was a change that improved their quality of life, since it made the rotation a little less overwhelming. It also ended up assuming a slight DPS improvement. It forced Armas players to learn the slight modification to their rotation, but it was an overall improvement, and there weren't many players protesting.

Fifth Example: Too Powerful Specializations

This might seem like an easy enough case to solve, but it is one of the most contentious, because the community will never agree if someone has too much power, or if their excess power reaches such a level that developers need to act. That they apply demeanors to us is disgusting. And period.

In general, players would rather we benefit all other specs than downgrade theirs; even if the result was the same. It is within human nature to want the improvements to be applied to the other specializations to be immediate, but when it is your character who is in question, you ask yourself: why in such a hurry? Again, it doesn't depend on whether the developers are heartless bastards (which we are), but on whether the players are having fun or not. You have fun being a one man army. You are not amused when a one-man army overwhelms you. It amuses you to hit the maximum on the counters. It's no fun feeling like you have no hope of competing with the guy who tops the counters.

Also, keep in mind that when we make changes to classes using live fixes, we try to find the simplest possible fix that will solve the problem, to minimize the risk of breaking something else, and the amount of testing required before we can apply the change. This is the main reason we are more likely to use live correction to downgrade one class than to benefit all the others, because it requires fewer changes. (Remember that if we were to benefit everyone's DPS to match that of the too powerful, it is likely that we would have to benefit creatures as well, to prevent certain content from losing all your interest. This would further complicate the change).

I also want to make it clear that today we hardly ever apply silent upgrades to classes, at least not on purpose. Players become very paranoid thinking that their damage can suddenly change. At most, it could happen that our programmers applied a change before the community team could report on it in the last live fixes blog entry, but generally, such a situation should not last more than a few hours. .

Sixth example: exploitation

It's not easy to tell the difference between when players know they are doing something they shouldn't, and when they are unsure whether developers might consider their action to be out of line. Like I said before, we generally give players the benefit of the doubt. If they figure out something clever they can do and it doesn't give them an unfair advantage or make other players feel disadvantaged, we usually don't do anything about it, at least not in the short term.

Unfortunately, there are a lot of bad guys out there trying to break the game for personal gain, or to satisfy their mere malicious nature. We consider it our duty to other players to stop these abuses when they happen. Understandably, we don't want to over-advertise these changes either. If someone has figured out how to take down a solo boss to get a lot of gold gains, we don't want to give thousands of players ideas by explaining the rift we found in the game and how we fixed it. Nor are they changes that we can meditate on for a long time. We have to get them out right away.

The reason I bring this up is that sometimes some players are surprised that we released a patch developed to prevent or discourage players from using exploitative behavior. A common reaction is to ask, "Was someone really doing that?" Remember that by their very nature, these types of changes are silent, and should remain so.

Seventh example: The expansions

We usually reserve a lot of design changes for expansions. We know some players find this too much to do, because they don't want to have to re-learn their character rotation, much less how glyphs work or the new PvE difficulty philosophy. However, we believe that if we want players to continue playing, we have to solve at some point the problems discovered in the game design. In this case, we find it interesting to add a reasonable amount of changes for the mere fact of changing.

Some players say, "My character hasn't had any major changes in years," and they want something, whatever it is, that allows them to look at their character with new eyes. Of course, we do not intend to fix things that are not broken, but we do try to make sure that each new expansion has a fresh air. Expansions give us the opportunity to strengthen the player base and the game itself. Therefore, you should not assume that every time there are changes to a class, this is because it was seriously flawed and drifted on the sea of ​​designer ignorance and apathy. We will likely never reach a point where a particular class reaches perfection and not one more touch-up is necessary. Change, in moderation, is healthy.

It is aspects like this that lead me to say that game design is an art, and not a science. Given the opportunity, there is no doubt that several of you would make different design decisions than ours, and in some cases, I have no doubt that your decision would be even better. We would love to see there is discussion on this topic. How much change is good? When can we let a problem sit for several months and when does it require immediate attention? What level of risk should we take when applying small changes that improve the quality of life? We are on the good way? We are crazy? Do you think this is nothing more than additional propaganda launched from the Ghostcrawler throne of lies?

Greg "Ghostcrawler" Street is the lead systems designer for World of Warcraft. He feels an unnatural rejection of the night elf's set of shoulders.