Thursday, February 10, 2011

Gevlon estimates that there are 1 to 2 million morons playing World of Warcraft, and concludes with a postscriptum: "PS: of course there are not just good players and M&S, there are casuals too, who just play for the content. But they are a minority and I doubt if Blizzard would be happy if they would only get the \$15 of good players and casuals." Wait a minute, a player is either a good player, a moron, or a casual? What about the average player?

If you sort a population by any measurable criteria, like their height, their IQ, or how talented they are in a video game, you will almost always get a curve which is so universal that it is called a normal distribution, also known as Gaussian distribution or bell curve. It basically tells you that most people are of average height, IQ, or video game talent, and that both extremely (small, dumb, bad player) and extremely (tall, clever, good player) populations are very small. The curve is also symmetrical, meaning the number of people at the both extremes are the same, e.g. exactly as many people are morons with an IQ lower than 70 than there are brilliant people with an IQ higher than 130.

Now the bell curve centers around a natural average. For IQs the curve is normalized, so that the average IQ is by definition 100. Other natural distributions don't have such a normalization, and people making subjective judgement cutoff points between good, average, and bad might not place them symmetrically. If you have an IQ of 130 and believe that anybody less clever than you is "a moron", you end up living in a world where 97.5% of people are "morons" in your eyes. If you not only have an IQ of 130, but also understand what a normal distribution is, and use a scientifically accurate standard definition of "moron" being less than IQ of 70, suddenly there are only 2.5% of "moron" in the world, and 95% of average people between 70 and 130.

If you say that 1 million "morons" are playing World of Warcraft, and you fairly put your cutoff points symmetrically, saying that a "moron" is as far away from the average as the "good player", you also get 1 million good players. And you get 10 million average players. You can put the cutoff points somewhere else, but in any reasonable definition of "good", "bad", and "average", you will always get a much bigger population of average players than of good or bad ones.

Ignoring the existence of the average is what Gevlon would call an "ape subroutine", a social mechanism where people tend to think in terms of "us and them", or "black and white". It is rather typical of certain WoW players, who tend to define "good player" as "playing as good as me", and then use some sort of insult as term for "everybody not playing quite as good as me". It is an extremely social reflex of status thinking to pretend that there are "good players" and "bad players", with nothing in between. Very few people want to admit they are average (For the record: I'm a pretty average WoW player, as far as I can measure that.). Thus you get that endless stream of comments on various blogs and forums on how Blizzard is catering towards "the dumb", or "the lowest common denominator", or "the morons and slackers".

Blizzard, who have much better data than we have, and better business sense, probably realize that there are very few really dumb players in World of Warcraft, and very few exceptionally good players. The huge majority of WoW players is average. And the only design decision which makes business sense is to make World of Warcraft reasonably challenging but doable for the average player. Depending on how large or narrow you define average, between 50% and 95% of players are average. That is where the bulk of their income is, and World of Warcraft *has to* be fun, that is challenging but doable, for that bulk of players.

And that is exactly what they do. For example Gevlon cites the "cloth geared, ungemmed warrior tank". Now how exactly is the recently introduced 15% buff for random pickup groups helping that warrior in cloth armor to tank? It appears pretty obvious to me that even after the buff that "tank" will still be exactly as unable to run a heroic as before. On the other hand there is the average tank, who is wearing a not perfectly matched mix of blue gear from quests and normal dungeons; who has gems and enchantments, but is maybe missing a few here or there because he counts on replacing those pieces, and who didn't buy the most expensive version yet either; who has a reasonable understanding of what all of his buttons do, but might not be aware of the latest spell rotation or priority list from the theorycrafters at Elitist Jerks; in short: The average Joe. That is exactly the guy who will profit from a 15% buff, because it is just enough to change the game from "too hard for the average guy" to "doable and reasonably challenging for the average guy". It is exactly that guy that Blizzard is catering for, that Blizzard is designing their game for. Because that guy, and the average healer, and the average damage dealer, are making up the huge majority of Blizzard's playerbase.
Let me extract some quotes from your post and prove some inconsistencies:

1) "For the record: I'm a pretty average WoW player, as far as I can measure that."

2) "the theorycrafters at Elitist Jerks"

3) "who has a reasonable understanding of what all of his buttons do [...] The average Joe"

So in the same text you define yourself as an average player, you mention EJ and then you talk about an average Joe who (by definition is average - like yourself) cannot know about Elitist Jerks.

If I am reading Gevlon right "good" players are actually average and above ones. The fact that so many morons exist proves that there is a skew from the normal distribution when looking at wow players which should be due to the types of people attracted to this game.

My argument in favour of Gevlon's three types rests on the distribution being skewed away from normal.

Bad players are his "morons". They exist, for a variety of reasons. They may be the green fire lover, or the melee so entranced by the boss's arse that they can't see they're standing in fire and ragequit in response to the "terrible healer", or the 75-year-old hunter who invariably breaks his own freezing trapped mob out with melee and has no idea he's doing it.

"Average" players are his "casuals". They are not average, though, but below mean. Casuals may even have above average aptitude, but they usually don't enhance that with optimizations from theorycrafters (ie: Elitist Jerks), watching raid boss videos, addons (many will debate this, but everyone should have SnowfallKeyPress and a boss mod), et cetera.

Good players may even have below average aptitude, but do what they can to be informed and prepared. One doesn't need stellar reaction times if one's watched the video a few times and knows exactly what the warning signs are for the raid AoE's.

The latter two categories are defined more by their behaviour than by aptitude, and the best numerical measures are often not the ones players tend to focus on. In a raid environment, damage taken, for example, is probably a better indicator of whether a DD is a good player than DPS.

Tomorrow there will be a clarification post about Morons and Slackers.

As a preface: the term has nothing to do with how good he is. One can be a terrible player and still not an M&S and one can be a better player than me, playing more than me and still a terrible M&S.

Tobold, if I wasn't already married I'd be asking if you wanted to gay-marry me.

Gevlon's social behavior and lack of recognition of that for what it is remains the major reason I keep reading his blog, I enjoy seeing an intelligent person routinely prove themselves wrong without realizing it. It's a thing of beauty.

you talk about an average Joe who (by definition is average - like yourself) cannot know about Elitist Jerks

Not true, I never said that. For example I myself "know about" Elitist Jerks, and sometimes I even look stuff up there. But are all my characters completely up to date based on the latest EJ theorycrafting? Certainly not!

Furthermore while my *performance* as a WoW player is pretty average, I wouldn't say that my knowledge about the game is average. Due to my blog I read a lot more about MMORPGs and WoW than the average player does.

This comment has been removed by the author.

Statistically speaking, I don't believe that the average person raids. Going by wowprogress kills, I would say that less than 20% of players raid. Its easy to believe that more people raid as these players are most prevalent on the blogs and forums that we read. Purely on that fact, and the game knowledge you clearly have, you are well above average.

+15% DPS can do a lot.
+15% HPS & HP the same.

You can of course wear 1-2 cloth items and survive a Hero using CC or having a good performing group (HPS & DPS what got buffed) So it is more likely that the 2 piece cloth tank will survive.

Now think about low-lvl. Yes cloth tanking was never a big problem. And adding 15%... ...how will new player adopt this?

@IQ-Thing: Below an IQ of 70 people will have serious problems with complex relationships and abstract systems. So we could say they are not likely to play WoW or other MMOs. I personally think that the real bad M&S have an average IQ.

I'm pretty sure about this since I have seen some test results in trivial knowlege at average IQ-pupil. It was a test in Germany and even some of them did not recognize Merkel or Hitler on any pictures.

BTW: EJ-Stuff is basic math. If you graduated anything in technology it should not be cloudy to you. It is far below Gauss algorithm, Laplace transformation or Furier. It only goes up to analysis.

In low level dungeons the 15% will make them even more of a joke than they are now.

I think the key issue is:

- 50% of players are better than average.
- Level 1 to 85/iLvl 333 is about 90% of the game (by development effort, and so on. For example heroic dungeons have few if any unique art assets).

If you change that balance so that 95% of the game (i.e. up to iLvl 346) is solely of interest to the average or below player, that's unbalanced. The remaining 5% is highly time and commitment-gated (i.e. you have to join an arena or raid team, arrange a time to play and and show up, so it is more a hobby or sport than a game).

So it does seem to me that if the game holds nothing of casual interest for the above average player, then those players will jump ship as soon as a viable alternative comes along.

Only 1-2 million? That's a strange contradiction of Gevlon's past claims that 95% of players being bad, or as he attempted to shuffle back fruitlessly, 95% of players in randoms, which due to their popularity are still going to account for a huge portion of the total population. Could it be that he's ever so slowly adjusting to the concept of reality?

"the term has nothing to do with how good he is. One can be a terrible player and still not an M&S and one can be a better player than me, playing more than me and still a terrible M&S."

Nope. When confronted he resorts to the classic "There is an unmeasurable problem, but I know exactly how much there is."

The 'average' WoW players is a terrible MMO player. That's the general issue when others talk about WoW and how it 'dumbs down' every three years.

It's a non-issue for WoW itself. If 95% of the 12m love to faceroll PUG dungeons, hey, knock yourselves out. But when those same players go on vacation, and now demand to 1-2-3-2-mash all of the content in MMO X (and regardless if the devs cave or not, they jump ship after 3 months anyway), that's where MMO players start to build up a hate for WoW players.

An hour of Rift beta general chat was the latest bit of evidence here, but overall its something that has been building for a few years now, and honestly, it's a well-earned reputation.

The 'average' WoW players is a terrible MMO player.

The average human being is a terrible MMO player. The average human being actually even sucks at Farmville. But even if 6 billion is too large for you, and you limit your target audience to "People in North America and Europe owning a computer able to run a game, having a decent internet connection, and being potentially willing to spend money on a computer game", you *still* get a population whose average is a terrible MMO player. But that would be a profitable target market towards which you could tailor a MMORPG.

The alternative is to make a game specifically targeted towards "The 1 million best MMO players out there". Unfortunately there is a high chance that the majority of those 1 million players aren't all that interested, and you end up with a game that is not or just barely profitable. Even if you would consider Darkfall to be such a game for the best MMO players (which I wouldn't necessarily agree with), you would have to admit that not all the 1 million best MMO players are playing Darkfall. And then you get into a vicious circle of low number of players meaning low revenue, resulting in low budget games, which result in less players willing to overlook lack of polish etc.

I tried the rift beta, and while i liked it, i found it even easier and casual friendly than wow. And what is a "good" mmo player? Not to start beating on a dead horse, but good is not an adjective i use outside of work (and even then)

And i can't think of a MMO general chat that i didn't want to turn off (yet), that wasn't created by wow.

I think there is a confusion of terminology. Some people are taking 'average' to refer to ability (a medium between good and bad), rather than a 'most common' player.

If the 'mode' player struggles with your content, you will lose subscribers.

While I generally support the thrust of your post, there isn't actually a mathematical basis on which to say that most people are average.

We know that heights are normally distributed, and we know from the law of large numbers that many things will necessarily end up being normally distributed with a sufficiency sample size. We can even combine what we know about nature and biology with math to make a reasonable guess that natural traits (like height) will end up being normally distributed, and then check this against the data.

But when it comes to intelligence, the distribution is totally artificial. IQ 100 is defined as the mean result on IQ tests and 15 is defined as the standard deviation. If you have an IQ of 130 that means precisely that you are (approximately) in the 95th percentile.

The problem is that there is no objective measure of intelligence. There is no linear scale to plot it on. We decided it was normally distributed to have a way to talk about it. We could just as easily have decided it was bimodally distributed in order to talk about he "smart" and the "dumb" people more clearly.

Play skill is the same. There is no linear measure of it that can be objective obtained, so any distribution we put on it is us imposing the distribution.

Of course one could argue that there being no possible objective basis for the division of good vs. bad players just furthers the main point of this post. That being that who you decide is good and bad at the game is a social or political decision, not an scientific one. If you think that the majority of people are bad, it is a conceit on your part, not a fact about them.

We could just as easily have decided it was bimodally distributed in order to talk about he "smart" and the "dumb" people more clearly.

No, that statement of yours is not true. While you are right that the IQ is normalized arbitrarily, that only changes the numbers. You can use a wide range of very different tests of intelligence, and you will always end up with a bell curve, and NEVER with a bimodal curve. Whether you then call the average of that bell curve "100" or "0", and what increment you give to a standard deviation, is arbitrary. But the shape of the curve never changes. Whatever measure you use, an average intelligence is ALWAYS the most common one.

@sthenno: In the case on LFD DPSers, recount gives me a definite comparison point. If most hunters I am grouped with do X dps, and this hunter is doing X/3 dps, can I not say he is a bad player (assuming all other variables are equal)?

It should also be noted there is a difference between an average player and average subscriber. This is why a bell curve doesn’t work.

I define player rating as your ability to interact successfully in game.

I define subscriber rating as your ability and desire to pay for one or more accounts and your utilization of you paid for time.

I propose good players are online less than poor players, and that average (skilled) players are online less than both good and poor players. You can only a “player” while you are in game. So when I say “player” I am not talking about how much you play but how well you play.

It is about presence. Bad players simply have more presence than other players.
The main reason for this is Time to succeed and focus.

Time to Succeed Goal – daily heroic. A Good player will complete this goal faster than an average player, whom will complete it faster than a bad player. The question then becomes, what does a player do once they complete their goal?

Focus – Good and Average players perform better than poor players because of gear, skill, and knowlege. They have better gear and skill because they focus more on improvement aka:moving goal-to-goal. Where-as bad players are more likely to play without focus. A focus could be getting a certain piece of gear, or having 100% uptime of some buff, etc.

Many goal-to-goal players log off when they hit their goal. “Just logged on to run my heroic daily, see ya.” Since a good and average player will finish faster than a bad player, they log off sooner. So in the case of goal-to-goal players, more bad players stay online longer.

So to mix in some numbers

Low subscribers plays average <1hour a day
Average subscriber plays average 1-2 hours a day
High subscriber plays average >2 hours a day.

You are much much more likely to run into a high subscriber in game than either an average or low subscriber. I propose mid to late content that there is a larger proportion of bad players within the high subscriber category. So even though 75% of wow subscribers may be average players many of these players don’t spend much time in game. 15% of the total subscriber base probably accounts for 90% of the players you see in game and of that 15% I would say the breakout is something like 60% bad, 15% good, 15% average. But that is just me

It should also be noted there is a difference between an average player and average subscriber. This is why a bell curve doesn’t work.

I define player rating as your ability to interact successfully in game.

I define subscriber rating as your ability and desire to pay for one or more accounts and your utilization of you paid for time.

I propose good players are online less than poor players, and that average (skilled) players are online less than both good and poor players. You can only a “player” while you are in game. So when I say “player” I am not talking about how much you play but how well you play.

It is about presence. Bad players simply have more presence than other players.
The main reason for this is Time to succeed and focus.

Time to Succeed Goal – daily heroic. A Good player will complete this goal faster than an average player, whom will complete it faster than a bad player. The question then becomes, what does a player do once they complete their goal?

Focus – Good and Average players perform better than poor players because of gear, skill, and knowlege. They have better gear and skill because they focus more on improvement aka:moving goal-to-goal. Where-as bad players are more likely to play without focus. A focus could be getting a certain piece of gear, or having 100% uptime of some buff, etc.

Many goal-to-goal players log off when they hit their goal. “Just logged on to run my heroic daily, see ya.” Since a good and average player will finish faster than a bad player, they log off sooner. So in the case of goal-to-goal players, more bad players stay online longer.

So to mix in some numbers

Low subscribers plays average <1hour a day
Average subscriber plays average 1-2 hours a day
High subscriber plays average >2 hours a day.

You are much much more likely to run into a high subscriber in game than either an average or low subscriber. I propose mid to late content that there is a larger proportion of bad players within the high subscriber category. So even though 75% of wow subscribers may be average players many of these players don’t spend much time in game. 15% of the total subscriber base probably accounts for 90% of the players you see in game and of that 15% I would say the breakout is something like 60% bad, 15% good, 15% average. But that is just me

Part of the problem I see is the inference by Tobold that intelligence is strongly correlated with being a WoW Moron. I think that correlation is weak at best.

Yes, intelligence has the standard distribution. That doesn't necessarily translate to morons, normals and elites having a standard distribution.

It is about presence. Bad players simply have more presence than other players.

I'm getting the same feeling. The reason elites focus on morons and not normals is that normals know that they are not elites. They have enough competence to recognize that. Morons are so incompetent, they don't have the basic skills to recognize that they don't have basic skills.

Morons insist on putting themselves right beside the elites. Normals don't try (or recognize that they are punching up when they do.) That is why the morons get all the attention.

The alternative is to make a game specifically targeted towards "The 1 million best MMO players out there".

But that's not the alternative proposed. The question is whether you make any attempt to cater for the 30-50% of players who are above the median of play-skill, but not ultra-hardcore in terms of commitment.

For example, in another Blizzard game, Starcraft II, you can get into a game requiring a pretty high skill level (100 APM, knowledge of this month's tactics) by simply pressing the 'join game' button and winning. You don't start joining clans, making appointments to play matches and so on until about the point the some of the podcasters know your screen name and play style, and you maybe get an offer to turn semi-pro.

In comparison, WoW is just astonishingly badly designed. Partly it is just that non-arena PvP is, by near-universal consensus, unfixable broken. But why have so few, and so small, dungeons? Why put them all at a flat level of progression, with indistinguishable gear and zero storyline?

Why make everything outside that require fixed, exact number and pre-arranged teams?

Scrap arenas, put raids and instances on a tiered, non-random dungeon finder system, and you will have a game that will leave WoW struggling to decide exactly when to announce the server merges.

@Tobold: Need to separate business from the player.If we are talking business, yea, clone WoW and pray. That's kinda obvious, and has been going on since 2005. It's also just another reason to dislike WoW though, as I'm guessing those devs making WoW clones could also be making some interesting stuff if given the chance.

I'm talking players thought, and since non-EVE servers cap out at around 5-15k players, what does having more than that get me? More servers to split friends between? You say a small dev team and budget, but WoW has 12m, a massive dev team, and still they lag in content delivery (or at best, don't smash everyone else like they should if more subs = more content). The team behind WAR spent more than DF will ever spend, and look what that brought. Minecraft was made by two people, etc.

Hell, the only 'good' thing about WoW having 12m subs is that when it's mentioned on a blog, a lot of people can relate or comment about it. How's that working out for ya?

I'm talking players thought, and since non-EVE servers cap out at around 5-15k players, what does having more than that get me?

So you honestly don't think that let's say Darkfall couldn't be improved by putting more money into it?

And what does having only 1 server as opposed to 100 get you? The population per server is the same, so it's hard to blame bad behavior on player numbers. How does a player even notice that he is playing on server #1 of 100 instead of #1 of 1?

Thought experiment: assuming 13 million WoW players what if the worst 12 million went to play Rift/Titan/FarmvilleAzeroth. Now someone who was better than 12.1m and in the 93rd percentile, top 7%, would now be in the bottom 10% being better than only 100,000 of the remaining million. So with no change in that player's abilities, they went from "very good" to "very bad." I.e., it is a normative not objective measurement.

1) With a dramatically better player base, do you think that eventually the % of heroic/raid/hard modes would increase dramatically? I say no; that Blizzard would not want a 99% success rate and would tune everything up until the difficulty/failure were what they wanted. Similarly, 7 million joined and increased the subscribers to 20m, I would expect that over time, the content would be adjusted so the normal distribution returned. WoW is not a static world. If everyone in WoW went to a month long class for WoW, then they by 4.3, the success raids on heroics and raids would be back to the curve Blizzard desires.

tl:dr; The success rate of raids/instances in WoW has very little to do with player skill but is determined by a Blizzard Activision employee looking at a spreadsheet.

2) My layperson understanding is that when polled a very sizeable majority of men consider themselves above average in looks and at least above average lovers. ( In defiance of both mathematics and their partners' reviews. ) For WoW, we can call this the "Gevlon number." This is the percentage of their fellow players that a typical forum poster thinks are "morons and slackers." This number is little influenced by the actual skills of the playerbase but is driven by human nature. My theory is that if the skills of the WoW playerbase were dramatically improved, the % being regarded as M&S would not change much.

Glad to see this post. I was starting to worry that you had been completely lost into the mmogs as sport meme that worships hardness. Your post shows you havent been.

I've come to the view that the only overiding concept a videogame should place as its primary goal in Fun. Everything else is only a means to acheave that Fun. Most concepts in video game design if pushed to far will eventualy stop being fun a certain point. Once a design concept passes the point of no longer being fun it stops being of benefit to the game and can easly turn into a negative. So everything but fun should be used only as a guide for game design up to the point where they stop being fun.

Everything but fun in moderation.

I believe Blizzard has the numbers.

They just have to look at achievement scores, or gearscores, and then take the top10% or bottom10% or whatever they want, and just look at what these guys are doing most - I imagine the top 10% will have tons of raid kills and the lowest will have... whatever it is they do.

"So you honestly don't think that let's say Darkfall couldn't be improved by putting more money into it?"

One would like to thing so, sure, but I've not seen enough to say "Look at this game,it got huge and the content really took off". 'Bad' things happen to niche games when they go mainstream when it comes to MMOs, and really, so long as the dev team is not a skeleton crew, things are generally ok for most games. Factor in community aspect (look at LoTRO post-F2P), and yea, I'm not really in the bigger=better camp.

The bell curve argument is wrong because in assuming a normalized spread of WoW ability it _presumes_ a normalized degree of difficulty.

People who argue that there are "M&S" in WoW believe that it does not take much effort to play the game well. Read some info on EJ, apply it, and move during the important moments. They don't believe that these tasks are too hard, or at the least easily learned/practiced, which leads to incredible frustration on their part when witnessing those who for whatever reason don't improve themselves.

As far as they see the game, there's a certain "difficulty" floor to break through, that the difference between being bad and good is a binary difference, not a continuum.

The bell curve argument is wrong because in assuming a normalized spread of WoW ability it _presumes_ a normalized degree of difficulty.

You don't understand. The bell curve is there, it is a natural distribution of skills. The degree of difficulty doesn't change the shape of the curve, it is just a vertical line *somewhere* on the X-Axis, telling you which part of the bell curve is "good enough" to pass. You can put that difficulty so that only a small percentage of players pass, or you can move the line to the left past the maximum in the curve and have the majority of the players be "good enough" to play. Difficulty is arbitrary.