Thursday, October 2nd, 2008
One of the biggest Black Swan events in history has been the meltdown in the financial sector. Triggered by risky lending practices, unemployment, real estate inflation, and overextended insurance policies (the “Credit Default Swap ” insurance market), the United States is currently undergoing its worst financial crisis since the Great Depression of 1929-1933.
Games have their own forms of economic disasters. As with combat systems, game developers need to walk a narrow tightrope between giving players too much and not giving them enough. Inflation in gold, equipment, and character abilities is inevitable in a long-running game (either a game series or a massively multiplayer game). Careful designers put a lot of time and effort into balancing game economics so that the game is as fun as possible for most players for as long as possible. This can be done by restricting aspects of the economy, reducing restrictions in other areas to keep things balanced, or by providing bread and circuses so the players are distracted from economic issues.
How to Kill a D&D Campaign
Two of my early Dungeons & Dragons campaigns come to mind when I think about game imbalance. In one game, the beginning dungeon master tried to foreshadow the ultimate battle of the first few months of the game. Unfortunately, players tend to focus on the here and now, so we immediately went after an opponent that should have easily destroyed us. Unwilling to have the campaign end in its first week, the DM had an NPC lend us some powerful magic weapons. Using them, we easily won, nabbed the treasure, and found ourselves equipped with magic that should have come much later in the game. After that immediate gratification, the rest of the game was an anticlimax and soon ended.
In a game that I ran, I had a “Deck of Many Things” as a treasure at the end of the first major dungeon. One player pulled a card that jumped her character 5 experience levels, so we had a party of 2nd and 3rd level adventures with a 6th level Cleric. I could handle it these days, but at the time, I found myself completely unable to balance the fights. Anything that would be a challenge for the Cleric would inevitably kill the rest of the party. The campaign went on for a few more sessions, but then died out because I couldn’t keep it balanced.
Whoever Said That Making Sequels Is Easy?
The Wizardry computer game series ran into this problem. Following the typical RPG trope, players started out as very weak beginning adventurers. They gained spells, abilities, and magic items as the game progressed until they were demigod level by the end of the game. Wizardry 2: The Knight of Diamonds allowed players to import their Wizardry 1 characters and continue the game. While KoD was still a very fun game, there was much less room for variety in the game because the player characters started out at a very high level of power. Unable to keep that going, the authors set Wizardry 3: The Legacy of Llylgamyn a generation later, so that players created new characters and began again as beginning adventurers. Wizardry 4 had a nice twist – You play as the evil Werdna, trapped by the goody-goody adventures from Wizardry 1, and have to escape from your own dungeon … starting again as a weak character with few powers.
Dungeon Master had similar problems when they created a sequel that continued where the first game left off. Lori and I loved Dungeonmaster on the Atari ST and spent many hours exploring it. We barely got past the first section of Dungeon Master 2 – It was just too difficult and stressful to be fun.
We kept these lessons in mind when we planned the Quest for Glory series. We intentionally masked the limits of the skill system by putting skills on a 0-100 scale. We figured that players would assume it was a percentile system and that 100 was the highest possible. However, we knew that we would be setting the limit to 200 in the second game, 300 in the 3rd game, and so on. Instead of letting the player wield godlike power by the end of the first game, we gave him a few abilities that would be equivalent to 1st, 2nd, or perhaps 3rd level spells in D&D. That left plenty of room for improvement in the later games.
Despite that, we still had issues with inflation. We had to balance the combat and puzzles so that a player who started in the second game had a character comparable to that of a player who imported her character from the first game. We tried to keep the game fun and challenging for all character skill levels, encouraging the player to practice skills, but not absolutely requiring it. There were flaws in this process. I think that by the 3rd game in the series, many players found it boring to repeatedly throw rocks until they had a high enough throwing skill to win a spear throwing contest, and so on. We relaxed the “practice makes perfect” requirement in the last two games to try to cut down on the tedium. Of course, in doing so, we also reduced the challenge for players who really wanted tough fights.
Oh Yeah, About That Gold
You might find it strange that I’m deeply into a post on economics and so far haven’t even mentioned money. While gold and silver are the most obvious economic systems in games, there are really multiple economies at work. I’ve mentioned character skills. There is also learning curve inflation – where a player gets so good at the game it stops being challenging – character abilities, equipment, pun tolerance, combat and experience points, and several other reward and challenge systems. If any reward comes to easily, it loses its value to the player. If it becomes too difficult, players become frustrated and might stop trying to get the reward.
In the original Bard’s Tale, no character class had a healing spell until (if I remember correctly) level 3. It was also possible to meet very dangerous enemies with your level 1 party, and it was difficult to earn silver. As a result, almost all of the characters’ money was spent on healing and resurrection spells at the temple. Most players ended up “cheating” by creating mule characters, adding them to the party, transferring all their money, then deleting the characters. Players had to work around the game system because the economy was too stingy.
World of Warcraft and similar massively-multiplayer games probably have the hardest time with inflation and balancing the budget. WoW has been running for over 4 years now, and many of the players have become jaded; it’s difficult to find any reward system that will keep them playing the game. Blizzard has relaxed many of the rules that originally created a tight economy because they know that long-time players don’t want to put in as much work for rewards that will soon be superseded. (Oh, and I learned how to spell “supersede” a long time ago in a Superboy comic book. Lana Lang became suspicious of a Clark Kent impersonator because he mistakenly spelled it “supercede”. Never let it be said that reading comic books is a waste of time – at least not all the time!) Dungeons that used to require long and arduous “attunements” before characters could enter them were opened to all players. This has allowed a lot of players in “second tier” guilds to experience content that they would otherwise never see (see The Burning Crusade paragraph below).
Blizzard initially made a number of good decisions to keep their economies balanced. Rewards for slaying monsters and completing quests were scaled throughout the game, so that each level of player tended to have just a little less money than they needed to buy everything they wanted. Blizzard also restricted magic items by giving them a minimum level for use and “soul binding” them – Once a character has equipped a magic item, it can no longer be traded to other characters. More powerful items are “bind on pickup”, which means that only the character who first picks up the item can ever equip it. These decisions meant that a player with a rich level 60 character couldn’t create a new character and have it instantly become much more powerful than its level.
WoW also keeps players involved by holding special events, often tied to seasonal or “real world” events. During the Olympic Games in China, participants in the battlegrounds became “competitors” and earned special tabards. Winning a battleground gave players a chance to win a “Spirit of Competition”, a Chinese Dragon pet that had no game purpose except to look cool. Other “critter pets” include the Invisible Wolpertinger, a jackalope-like creature that you can supposedly only see when you’re drunk, the Baby Murloc that was only made available to attendees of the first live BlizzCon convention, and many others. The rarer ones are highly prized; some are sold for hundreds of real dollars on eBay.
Black Swans Invade a World of Orcs
There was at least one “Black Swan” phenomenon in WoW that might have been obvious to Blizzard, but caught me totally by surprise. Prior to the release of The Burning Crusade expansion, characters were “capped” at level 60. Unlike in Everquest, where the maximum level was a goal attained by few, a high percentage of World of Warcraft players got to level 60 with their characters. In order to keep them active and interested, Blizzard provided a large amount of content designed specifically for level 60 characters. This included the three major “raid dungeons” – The Molten Core, Blackwing Lair, and Naxxramas – that required guilds to put together well-balanced teams of 40 players at a time to have any chance of success. Most of the big World of Warcraft guilds were structured entirely around the raid dungeons.
The Burning Crusade changed that dynamic entirely. Players who had spent a year or more at level 60 raiding the dungeons or fighting against other Level 60 players in battlegrounds now found themselves doing quests and exploring 5-player dungeons again. In order to encourage players to buy the expansion and try out the new Outlands content, Blizzard greatly increased the availability of gold and the power level of magical weapons and armor in the new environment. Instead of the level 60 raid dungeons gradually becoming phased out, they were all abandoned the day TBC launched. With 20-20 hindsight, we can see that much the same thing will happen with all of the level 70 dungeons after Wrath of the Lich King launches. A few guilds may visit one or two of them for nostalgia, but they will soon be entirely abandoned as players focus on reaching level 80, exploring the new dungeons, and find more powerful items than they could get from the level 70 dungeons.
Overnight, the big raid guilds splintered. Nobody had any interest in struggling for hours in a level 60 raid dungeon to have a small chance of winning an item only slightly (if at all) better than what they could find in a 5-man dungeon in Outland. Some guilds set up teams of 5 players to work together to “beat” the dungeon content and get “attunements” to the new level 70 raid dungeons and heroic-mode dungeons. Many of the guilds collapsed, because the 40-man raid dynamic was what had provided the social environment that held them together. Others survived by emphasizing role-playing or by working together to make sure that everyone in the guild could find fellow adventurers when they needed them. The entire gameplay for most of the players altered.
The Bird’s-Eye View
How do you balance a multiplayer game? Step 1 is to make good initial decisions. Always be aware of the dangers of a runaway economy and build balances and restrictions into the game. Step 2 is to create simulators. Set up test cases for characters of various levels (or skill development for a non-level-based game) and earned or purchased equipment. Run simulations of combats and other quest activities with them to see if the better-equipped characters are overwhelmingly powerful. If so, make adjustments. Step 3 is to put the systems into the hands of beta test players and play testers. Separate them into groups and present different rules to each group. Then keep the ones that are most successful. Finally, pay attention to your players once the game goes live. Gamemasters and community managers can observe the players to see if they’re having fun and where they’re getting frustrated. If there’s too much money in the economy, find things they can spend it on that give them marginal improvements. If there isn’t enough, introduce new quests or other activities that make it easier to earn money. If all else fails, release an expansion.
More than anything else, game balance is what makes or breaks a computer game. Creating and maintaining well-balanced economic systems for gold, equipment, character skills, tension vs. reward, and other aspects of games is essential to creating games that last. Keeping a game balanced while half the players are trying to find ways to unbalance it is one the most challenging tasks faced by game designers and developers. As Julia Ecklar wrote in her song Crane Dance, “Balance is to understand the crane while it’s in flight.” Cranes and black swans – Look at game economies from a bird’s eye view. You’d be bird brained not to.
The Black Swan, by Nassim Nicholas Taleb (2007) discusses rare, unpredictable, and catastrophic events that – after the fact – people say, “Oh, it was obvious that was going to happen.” The current government intervention into the U.S. banking system definitely qualifies as a Black Swan.
Unemployment statistics understate the real issue, since they do not generally include “discouraged workers” who have given up on trying to find work. As unemployment has grown, people have been unable to make their mortgage payments. That’s what is really meant by “high-risk” mortgages – Lending to people who lose their jobs.
The CDS, or Credit Default Swap, is a new type of financial instrument within the last 10-15 years. Companies insure lenders against customers who default on loans. The problem with CDS’s is that they are based on averages. That leaves the insurers (such as AIG) unprepared for catastrophic Black Swan levels of default. The current CDS market is estimated at $55,000,000,000,000 – that’s $55 TRILLION. CDS’s allowed a lot of lenders to overextend themselves with risky loans because they were insured. That didn’t help when the insurers couldn’t cover the losses.
This article was written by Corey and illustrated by Lori. Mojo is Lori’s Troll Beastmaster in the World of Warcraft game who is the proud owner of a baby Murloc.