Monday, October 05, 2009

The Gini coefficient is a measure of equality or inequality, used in economics. It can be used for example to describe how income is distributed in a population. If one person had all the income, and everybody else had nothing, the Gini coefficient would be 1, perfect inequality. If everybody earned exactly the same, the Gini coefficient would be 0, perfect equality. It is easy to see that neither of these extremes would be stable, if one guy had all the money, the others would hang him up the next tree and distribute his wealth, or he would have to pay others to protect him, also leading to wealth distribution. If everybody earned exactly the same, regardless of effort, there would be no incentice to excert any effort, which is why communism failed, although the Soviet communism was still far from a zero Gini coefficient.

So the Gini coefficient is interesting insofar as there is no easily visible optimum. Very poor countries usually have a high Gini index, around 0.6. Europe has a very low Gini index, 0.31 in 2005, while at the same time the USA had a Gini index of 0.47. Any measures that change the Gini coefficient of a country usually give rise to strong emotions and discussions. Just look at issues like Obama's health care plans (which would lower the Gini index of the USA) or the excesses of banker's pay (which raise the Gini index).

Applying that concept to MMORPGs isn't all that easy. In the real world your wealth and income is measured in monetary terms. In a MMORPG it is measured in the power level of your character. For example somebody in World of Warcraft at the gold cap, geared up in the best equipment money can buy, would still be "poorer" than a top raider. So as a first approximation we could define the "wealth" of a WoW character as the sum of the iLevels of his equipment. But that would still leave us unable to compare the Gini coefficient of World of Warcraft with that of another game, because while the principle would be the same, the exact way on how to calculate power and wealth of a character would be different.

Nevertheless, while we might not be able to actually calculate the Gini coefficient of MMORPGs, we sure can state some general trends. For example we can say that over the years the Gini index for World of Warcraft went down, that is WoW became more equal. The difference in power level between an average character at the level cap and a top raider diminished, as nowadays it is a lot easier to get epic gear even for a non-raider, which wasn't true in WoW 1.0. We can also say that in other MMORPGs the power difference between the most successful and the average is wider, thus they have a higher Gini coefficient.

Looking at equality has important design consequences for a game. Games that are not MMORPGs, but nevertheless have "PvP", whether that is chess or Counterstrike, have a Gini coefficient of or near zero. Right now a bunch of developers from Blizzard is extremely busy trying to get the Gini coefficient of Starcraft II as close to zero as they can. Because if the result of PvP depends on a mix of power of your character(s)/units and player skill, by balancing the power you get a PvP that depends only on player skill. It is easy to see that the Gini coefficient of MMORPGs never is zero, and even with measures like battleground brackets and pairing based on arena ratings, it is impossible to get MMORPG PvP to be solely based on player skill.

Even for PvE the Gini coefficient is important. A game with a high Gini coefficient is like a third-world country with a ruling elite and lots of poor. That suits the elite just fine, but is obviously not so pleasant for the poor. And unlike the real world poor, the virtual world poor can always just leave and stop playing, which isn't good for the earnings of the game company. World of Warcraft lowering its Gini index and making players more equal is a direct consequence of such a consideration, of pleasing the average player to the detriment of a small elite. Nevertheless you wouldn't want to go too far, because if everybody is perfectly equal, then there is no game any more in gathering better equipment and raising your power level. Just like in the real world it isn't obvious where exactly the optimum lies, and any changes lead to heated discussion.
Interesting discussion topic. Something else to consider is that with extremely high inequality, the ultra-poor also lose their motivation. When someone is too far down, the middle (we're not even going to think about the top) is unimaginably out of reach. Throw in the tiers of progression and a low player might look at end-game and just go "well fuck that" and go look for another game. This is sort of what happened when I tried EVE. I looked at how long it would take to become competitive, let alone have any sort of influence, and it just wasn't worth it. The smaller gap in WoW, in addition to the various forms of alternative progression mean that people actually can pull themselves up by their bootstraps.

That's what I'd consider the ideal: the range in which a person is motivated and able to move up and doesn't need the permission of anyone to do it. It's not equal or fair, but it is close enough without falling into the realm of science fiction.

There is also one other important factor to consider: economic mobility. The U.S. has been marketed for quite a while as a "land of opportunity": Even if you immigrated there with nothing but clothes on your back, you could work your way from rags to riches. Even though the immigration from Europe tapered out when economic mobility was improved there, the image is still vivid in Latin America, especially Mexico.

If one tried to map economic political opinions, it would be useful to use a dual-axis system not unlike D&D's Lawful/Chaotic Good/Evil system: High/Low mobility and High/Low gini. For example, feudalism would be high gini, low mobility, and communism would be low gini, low mobility. Laissez-faire capitalism would be high gini, high mobility, but I don't know what would be the equivalent of low gini, high mobility.

That's a very interesting way of looking at things.

When you look at Blizzard's actions through this lens, it becomes clear that not only did the adjust the Gini coefficient between TBC and WotLK, but they're also balancing around it. The introduction of HTotC, a 5-man with 10-man second-tier loot, and the promotion of Emblems of Heroism and Valor to Emblems of Conquest seem designed to keep this power difference at a given absolute level, and keep 5 man dungeon-runners a tier below 10-man raiders (and thus two tiers below 25-man raiders). The introduction of new dungeons dropping ToC10-level loot in 3.3 will further this trend.

Overall, I think it's a good thing. A lower Gini coefficient allows for more social mobility. In the MMORPG case, it makes it easier for someone to go from a newbie to a hardcore player.

There are several ways to measure equality and none of them is the right one. Actually, this post could do entirely without the gini coefficient.

"Flow tends to occur when a person's skills are fully involved in overcoming a challenge that is just about manageable."

This is what MMOs try to supply to their customers. More important than the equality is the individually perceived equility.

I think that gamers will put up with a big inequality in wealth as long as the "have nots" see a mechanism that will allow them to one day reach the current wealth level of the "haves. No-one complains about the huge imbalance between a level 1 and a level 80 becaue the level 1 will one day get to 80 if they keep playing. The fact that the present day 80 may have reached level 100 by that time doesn't really matter either - as long as there is no barrier to any individual player progressing as far as they wish.

I don't think you can apply it to money in WoW but it would perhaps work better if you applied it to time. As MMORPGs reward time more than anything else, if someone has a lot of time to invest in the game then they have an advantage over someone who doesn't. Their wealth will eventually cap and become meaningless but what they can achieve with time will not.

"it is impossible to get MMORPG PvP to be solely based on player skill."

Um, why not? Is it really IMPOSSIBLE to make an MMORPG without character advancement?

Is it really IMPOSSIBLE to make an MMORPG without character advancement?

Yes, I would say so. I think that character development defines the RPG genre as much as persistent worlds do.

It's hard to compare, because most MMO's are crap economies at best.
There is always money/gold pumped in by the higher authority without any consequence except perhaps some artificial money-sinks.

also there are infinite resources to grab, so even though it's rare it is always an infinite amount to gather.

imho, they should work on that part.

A lower Gini coefficient allows for more social mobility.
While a low gini coefficient makes it easy to become "rich", it's only because the bar for "rich" is set so low. While the relative mobility might be high, absolute mobility would be minimal. And like Tobold pointed out, that creates an incentive problem: If more effort results in only minimal benefit, why make that effort in the first place? While that can be disastrous in real life, it doesn't matter much in a subscription-based game. Even if the player logged in once per month, the company would still make the same amount of money. With a pay-as-you-go system, the dynamics would be quite different.

It's hard to compare, because most MMO's are crap economies at best.
There is always money/gold pumped in by the higher authority without any consequence except perhaps some artificial money-sinks.

But as We Fly Spitfires noted, perhaps the currency in MMORPGs is time, which is always finite.

The Gini coefficient is a great tool for many purposes, but I just dont see how it could apply to an MMO setting. Because one thing that the Gini coefficient doesnt consider well is how "efficient" people are with their time/money...and for an MMO, that would be a high point of contention I feel, because each game has different economies, levelling curves and playtypes(PvP-PvE-RP).

"If everybody earned exactly the same, regardless of effort, there would be no incentice to excert any effort, which is why communism failed"

What on Earth are you talking about? Two things:

1) Why is the wealth distribution fixed in your perfect equality example but not in your previous perfect inequality example, where you had wealth redistribution occurring? If it's not fixed, then there is a *huge* incentive to try to make slightly more in the perfect equality example, because then you're the richest person in the country, with all the social benefits that carries. Also, everyone starting at the same position is the same state as the beginning of a... *race*. Race to be the richest.

2) Communism failed in the USSR because it tried to control the entire economy from a central bureaucracy, and was an oppressive regime. An example of the bureaucratic approach: bean counters in Moscow would decide what and how many shoes a shoe store in Vladivostok would carry, regardless of demand or what the people in Vladivostok thought. It had nothing to do with everyone having even close to the same wealth, which is untrue in any case: in an oligarchy like the USSR, wealth and power was concentrated at the top, just like any other screwed up state.

Note, however, that totally free markets are not the solution either; you need a good set of regulatory laws in order to keep everyone honest and encourage competition. (Free market proponents tend to ignore the fact that people will lie if they can get away with it, like Enron, and also that monopolies can eventually arise which can destroy economic development.)

Good post. I don't completely agree with it, but it's good for thought.

I don't fully agree with people assuming that WoW didn't profit from timing. Alot of people have said if LoTR had come out before WoW that would be the king. While I'm not sure if I agree with that, I do believe if WoW had waited longer it wouldn't be the powerhouse it is today. That's not to say someone else would be instead.

The biggest issue I have now with WoW is that it has jaded everyone. If any MMO that comes out is harder than WoW, it's to "grindy". If it comes out to much like WoW, it's a "WoW Clone". The market refuses to remember what MMO's where like before The Burning Crusade. I think that expansion is what started us down this slippery slope.

I believe WoW has reached it's peak. When MMO's first come out the early adopters are typically nerdier customers. We research games and seek out the new. Early versions of MMO's cater to us. As the age of the MMO ticks by the dev's slowly make the game easier and easier, thus enticing the more general public.

The general public never played WoW at launch. They will never be able to suffer through the hard times of a newborn MMO coming into it's maturity.

Blizzard has only added to this problem by making every aspect of WoW so easy. When a new game launches it's "to grindy" because it hasn't had all the difficulty removed to entice the general public yet, and thus gets negative reviews.

"It's hard to compare, because most MMO's are crap economies at best.
There is always money/gold pumped in by the higher authority without any consequence except perhaps some artificial money-sinks.

also there are infinite resources to grab, so even though it's rare it is always an infinite amount to gather.

imho, they should work on that part.
"

Working on that might make a more realistic simulation of a real world economy, but it would not necessarily result in a more FUN game. Which is what they really need to stay focused on (well really they need to focus on keeping subscriber numbers up, but the primary tool for that is keeping the fun up, which is good for me because I care very little about a games subscription numbers, but a whole lot about how fun it is).

Ultima Online originally had an economy with a fixed amount of goods per world, and ditched it in favor of an unlimited creation of new goods. The fixed amount of goods tended to make things un-fun. People would hoard all the resources that constituted a monster (the loot, or even the meat) and then the game wouldn't make a new one, and things got boring.

There is a bunch of details in: http://www.mine-control.com/zack/uoecon/uoecon.html -- warning it is long, and unless you are really interested in MMORPG economics it isn't very exciting. It was also written pre-WoW and talks about a state of the art that has long since become history. I found it interesting, but I'm not sure others will.

I expect there would have been ways to improve the original closed system (what they call recycle bins, and half of what WoW venders do for example...plus making the in-flow proportional to the number of players...and maybe some other things). However I expect the end result of that would be a lot of work tuning things to get a system that is not MORE fun then the "just create new stuff whenever it is needed!" system.

So MORE work, and NOT more fun. It is a better investment of programmer time to design something that _is_ fun and just apply a crude hack to the economy.

(even if you just look at the economy I would rather have the artificial open economy with auction houses then a well designed closed one with no auction houses...or sticking to WoW I would rather the econ team add "buy orders" to the AH then redesign WoW to be a closed economy system)

"No-one complains about the huge imbalance between a level 1 and a level 80 becaue the level 1 will one day get to 80 if they keep playing."

Yep. The complaints mostly start when folks hit the level cap and want to do whatever end-game raids exist. Some of this is because nothing in the game really explains to you that once you hit the level cap you effectively level up by getting better gear, and for most folks that means the next levels are:

* farm mats and gold to get the mid-price crafted items

* farm rep to get some faction only items and enchants

* do the easier 5-man heroics to get some gear upgrade and get badges

* do harder 5-man heroics (for the same reasons)

* do 10-man plus the hardest 5-mans

* do hard modes on 10-man, or do 25-man...or something.

In fact you can argue about the order to do these in, and about which heroics are doable by who with what gear. Which is part of what makes things confusing for new level capped folks. It is also part of what keeps things from being too boring for folks that hit the level cap a while ago.

That there is no in-game easy way to look at someone's total "gear score" or get an in-game recommended list of dungeons to attempt is also a big difference between the "leveling game" and the "end game".

It would probably cut end-game attrition if there were an easy way for someone to find out "if you and a group of similarly equipped people get together you should do ok at X, Y, and Z...you will find A, B, and C really hard" Or better yet "the party you have formed should be able to do A, B, C...". Or even better "the part so far plus the guy you have selected in LFG should be able to do A, B, or C...".

It doesn't have to be all that accurate. I mean sure there are level 80s in good gear that pump out 800DPS while someone else in the same gear hits 2500DPS. However at level 20 there are groups that can handle deadmines and some that will wipe, but 20 is what the quests claim (more or less). So anything approximately as accurate as the levels assigned to quests would do fine.

The downside of a system like that? Well it would make it easier for folks who are "under geared" figure out how much "leveling" is left and that might discourage them. It also might make folks equip the WRONG items just because they have a higher ilevel... a +AP +Str trinket is not an upgrade for a holy pally, but if a higher ilevel item gives a higher score on the "in game gear level" system then some folks will blindly follow it...more so then the 3rd party systems that do the same thing. So I think a in-game gear score would have to be smart (more like be.imba.hu, less like wow heroes).

Personally I think it would help more then it would hurt. It would focus new players on the simpler instances without having them run TOC or OK and ending up with wipe after wipe and having no idea where they can go to "level up"! Right now people hit the end game and if they don't know about the "gear game" they stall and are stuck either learning something the game doesn't really teach, or deciding to level an alt... or quitting and finding a new game.

...to be continued

...continued (blogger hates Wall O' Text...but I'll show it!)

The mount drops from brewfest are similar. New(ish) 80s can pick up trinkets (and rareish weapons) that are really upgrades. They can find better geared folks pretty easily if they don't care about the better geared having a summons (or offer not to roll on any mount drops).

To bring things somewhat back onto topic... I think the Gini coefficient measured at each level would be more useful then the Gini score across levels. With the Gini score at the level cap being the most interesting (at least to me).

For PvP the Gini score in each battleground level range, but also separated between those that have turned XP gains off and those that still have XP gains turned on.

I don't know if you ever watch the videos on TED.com, but one of the best ones is on the economics of the crack industry. The gang leaders don't hook in new dealers with great pay and benefits. In fact, the researcher found that drug dealers make \$3.50 per hour, at a job where you are significantly more likely to be killed than someone who is actually on death row. The gang leaders hook new dealers in by selling the promise of something better, that one day they will become a big rich gang leader.

Players in an MMORPG play for the prospect of better gear, not because of how good the gear they have been handed is. No one would play an MMORPG where you log in and are handed the best gear immediately. This is where WoW is starting to fail, and why you can't have so much equality. What does it mean to have the best gear in WoW right now? Everyone else will have that gear a month after the next patch.

"This is sort of what happened when I tried EVE. I looked at how long it would take to become competitive, let alone have any sort of influence, and it just wasn't worth it. The smaller gap in WoW, in addition to the various forms of alternative progression mean that people actually can pull themselves up by their bootstraps."

Frankly you were mislead in what being "competitive" in Eve entails. Expecting to be influential and make huge decisions is both foolish and overly ambitious. If however, you wanted to be competitive in combat, trade, manufacturing, or any other basic area, this is easily accomplished in a relatively short period of time.

The large PvP battles are populated with not just capital ships, battlecruisers, and carriers, but smaller cruisers and frigates that are accessed with skills taking less than a month to acquire. Joining an experienced corporation speeds up the process as they often provides free ships and materials to outfit you for any given role.

Combat in eve is 75% tactics based, 25% skill (as in, the skills your character has) and equipment based. Because even a cheaply outfitted frigate can be of vital use* to a fleet, even a nearly brand new toon will be able to fight in PvP competitively.

*Smaller ships are used to pin down larger capital ships and cruisers to prevent them from warping, enabling larger vessels to blow them away

Note: Many people have the misconception that in Eve you are forever at the mercy of griefers and Pvpers. This is frankly not true. You can choose to enter 0.0 space and get destroyed. You can also choose to stay in safer 0.9 or even 1.0 space until you are ready to begin Pvping. Its up to you the risks you take, and the rewards scale along with them.