Greedy Goblin

Monday, June 27, 2011

Size doesn't matter, 4.0 summary

I wrote about it in December and January. The result was that server progression is not affected by server size. The conclusion was: "All servers have almost the same amount of raiders and casuals. What turns the system upside down are the M&S. Morons and slackers are not casuals. They are not just "bad players". They are bad players who refuse to learn to play, yet they feel entitled to rewards. Raiders learned to play, casuals ignore epics and just play. M&S suck and want to be carried. In a small faction they soon visit every single guild. The raider guilds ignore their applications at best, send it to me for "morons of the day" at worst. The casual guilds accept them, but the M&S is not happy there, as casuals can't boost them to epics. There are few obnoxious M&S guilds (/2 xXxDethkillazxXx need moar active geared tankz and hilz 2 start raiding whisper ipwnudie 4 inv) but after a few attempts they tend to dissolve or transform into an alt-boosting lolguild. When the M&S runs out of options, he reaches the conclusion that "this server suxxs cokk hard", and moves to another one. Since larger population servers have more guilds, it takes much longer for him to get rejected from all the raider guilds and to figure out that the countless casual ones can't boost him. Also, on such servers the M&S guilds spawn and dissolve fast enough for making the impression that there is always another guild to try out, despite they are just permutations of the same M&S base. So these servers act as black holes, sucking up all the M&S."

Now 4.0 is closing so I collect the data once more to see if there is any difference in casual raiding. I mean, if you had any intention of raiding, no matter how casual you are, you could at least get in a Halfus pug.

The chart shows the wowprogress score of the server vs population:
As you can see the results are totally uncorrelated (0.03 R2 is very low). The amount of warm bodies gives zero contribution to the progression.

Since commenters claimed that HC guilds are overrepresented in the score (so a realm with 1 13/13 and 5 1-5 bosses guilds gets similar score as the realm with 1 13/13 + 50 1-5 bosses guilds), I manually compared the 5 largest realms on the US realm list and 5 next to each other on the same list in the 3500 players region. The results are the following (sorry for the lame format, did it in a hurry)
The "segment" field is the corrected average, as the HC guilds were also counted as 9/12 and 1/12 in the data. The red line says that if we increase the realm size by 910%!!!, we get 20-30% more guilds. Not just HC guilds, even 1/12 guilds are only increasing a bit.

The high population servers are nothing but ghettos full of M&S. Avoid them like the plague.


Anonymous said...

I'm fairly sure that server score is already an average of some sort. Since the wowprogress site doesn't actually say how realm-rankings are created I'm not sure what the use of putting that score into relation to population is. There's a good chance only the top 10-20 guilds are going into this score.

Assuming player skill follows the same or a similar curve across all realms - even if its not a bell curve - it should be fairly obvious that the total amount of good players goes up with the total amount of players. Additionally, people that go realm shopping don't just look at population size, but the actual score as well.

Anonymous said...

On a high population realm with top guilds you can know quite a number of good players from various guilds, since there are many of those playing on that realm. If they know you are a good player/class/role (yet a casual) you have a peer to play together with. He'd rather invite you than trade PuG. If they need your class/role, you can get that unique opportunity to kill an otherwise impossible boss, or at the very least have some good tries at the fight. The bad players on a high population realm you simply /ignore and forget about. This creates a sortof cliquish/elitist atmosphere, not easy to enter for new comers.

A bad realm does not exist according to your argument. However, on a low population realm, there are less good players you know. It is therefore harder to clear content which requires a lot of players. Before Cata, it was required to raid 25m in order to get the good gear in order to progress further & easier (e.g. grab ICC25 gear to clear ICC10N, or ICC25HC 4/12+ gear to clear ICC10HC LK). In Cata, almost all 25m boss fights (normal and heroic) are easier than their 10m counterparts. I don't see any BH25 trade PuG runs on Agamaggan alliance. The result: less (quick) progress, less loot drops, less well geared; fights are harder.

As you said in another post: it is not about who you know, it is about who knows you.

Anonymous said...

This post is interesting. Well alot of this data is dramatically affected by the presence of top guilds on certain realms it seems to me that it shows pretty clearly that population doesn't effect progress.

This might mean that the proper realm to choose is a realm with a low score with a high population since they are lacking any decent guilds to form around. Either that or they are truely filled with M&S.

Camiel said...

Even if there were a correlation between server size and progress, there would be one possible way to explain it:

If a server is home to a top-guild (world-first kills kind of stuff), it attracts fanboys like bees to honey. Obviously these fanboys are clueless M&S, because raiders don't transfer to another server unless they have an accepted invitation to a raiding guild and casuals couldn't care less on what server they are.

Anonymous said...

It's a myth that 10man bosses are easier created by the fact that some of the easy hm bosses like halfus were way overtuned on 10 man. If you look at the mechanics for the end bosses like Al'akir, Nefarian, Sinestra, and Cho'gall, you'll find they favour 10man, as you dont have to set up interrupt rotations on raidwipe spells and much more space to spread out and kite orbs and the like around.

Steel H. said...

Not sure if my post made it through - had errors, here's a rewrite

One thing you may not be aware of, since you created your own guild, with your own rules and schedule, is that it's difficult to find a guild with the progression level you want, with raid hours you want, and that is also looking for your particular class spec. On my initial low pop 'new players' server, where I first started playing, there were like two guilds doing ICC 25 HM, a couple of ICC 4/12s, pug raids that were the end of all hope, and so forth. And the two serious guilds all raided at bad hours, and didn't need hunter dps.
So I transfered to a high pop, more progressed server, with scores of guilds on each level. I was able to do stuff like: hmm, this 8/12HM guild is recruiting hunter but raids at 6:30, that's like 30 minutess too early for me, I'll try this other 8/12HM guild that is also recruiting hunter but raids at 7:30. That's just not possible on a low pop server.
As for how many M&S are on this new server, and how your average PuG goes, I wouldn't know. I haven't been joined to /2 since way before Cata, and never run PuG raids - it's why I transferred to this server in the first place.

Squishalot said...

I said that your conclusion was invalid 6 months ago, and I still maintain it. You're trying to prove an ideology that has no identifiable relationship to the numbers that you're presenting as evidence. You could use the same butchered argument to state that the bigger the server -> the faster the fishing respawn rate -> the reason there is no correlation is because people are fishing rather than raiding.

Of course, that being said, here's your biggest flaw. The Wowprogress Realm score is based on the straight line average of the top 15 guilds on the server (as rated by Wowprogress). As a result, the Wowprogress score is completely and utterly irrelevant to the proportion / number of raiders, casuals and M&S on a particular realm, assuming that the top 15 guilds on each server are all serious guilds (15th ranked guild on the lowest ranked US server is still 9/12 progression).

Squishalot said...

Let's also be clear - you could transfer every hardcore guild in the world with between 5000 and 10000 wowprogress points onto US-Illidan (top ranked US realm with world rank 8 guild Blood Legion) and you'll still get exactly the same wowprogress score. You could pull everybody off the realm other than the hardcore raiders in those top 15 guilds, and you'll still get exactly the same wowprogress score.

This, by definition, invalidates your argument.

Gevlon said...

@Squishalot: just for you a new chart

Squishalot said...

Thanks Gevlon. It's just a shame that it doesn't prove anything though. The number of guilds with raid progression is still irrelevant to the number of M&S on the realm, especially since Wowprogress ignores PuGs (has to be a significant number of guild members in the kill for Wowprogress to acknowledge them). It doesn't just look at the individual player achievements, it looks at guild kills.

Summary: You still can't ell the difference between a non-raiding casual and a M&S in your table.

Anonymous said...

Are you trying to intentionally misguide people or do you just not know where your data comes from and how it's generated? You're being ridiculous if you think data that counts less players than the wowprogress data (which only counts raiders) to provide any sensible data point. All your handpicked low pop realms have less than 50 snapshots taken.

Think about the simple fact that any decent player that transfers will a) transfer to a server where he'll join a top guild or b) transfer on a server high up on

Gevlon said...

@Last anonymous: do you know any better data than warcraftrealms? Also don't you find it funny that you talk about "decent players moving to top guilds" which is irrelevant to the post as top guilds exist on every realm?

@Squishalot: Wowprogess does NOT ignore /trade pugs. It checks for players, not guilds. If 10 members of a guild kill Halfus in 10 different trade pugs, WoWprogress consider their guild Halfus killer.

Also, while we really don't know who is a non-raider casual and who is an M&S, to disprove me, you have to somehow prove that casuals flock to certain servers. Please note that for casual play every server is equal and low pop realms have "recommended" green sign, while high pop are red. So why would casuals flock high pop servers?

Anonymous said...

It is not relevant if I have better data or even if better data exists. The data you used is irrelevant and misleading. I'm not saying WarcraftRealms data can't give you decent information either, just that you have to be careful which realms you look at, as the site clearly states that realms with red ratio numbers are not to be trusted simply because there's not enough data available.

And with regards to topguilds existing on every realm, the guild I'm in (we're 13/13HC for some time now), is constantly recruiting new players just to keep up with the people leaving for a top100 guilds. They also transferred away from a low pop realm quite a long time ago, precisely because there weren't enough solid players to recruit from. This doesn't have to be true for all low pop realms or all guilds.

Lisa Treadway said...

Roleplayers flock to Wyrmrest Accord in droves, and it's the server I most hear spoken when people ask, "What's a good server to RP on?" Now, I'm not saying that roleplayers don't ever raid... but a lot of them are extremely casual, in my experience.

Armond said...

Your recommendation is sound for mid-level players. Unfortunately, high-level players looking for world class guilds or world firsts have little choice in the matter.

Squishalot said...

@ Gevlon: From what I can deduce on comments in the FAQ page, it looks at date of kills as well, to determine if a kill was made as part of a PuG, or if it was a guild kill. Only guild kills are recognised (look for comments relating to guild unions). So unless everyone PuG'd down Halion on a single night, it's not going to count.

As for your theory about flocking, that relies on the assumption of perfect movement and zero transaction costs between servers. My current realm is a high population one. Due to cost, I'm not going to move to a low population one just because my realm is more crowded. I would argue that most 'casuals' would remain on the server that they initially rolled on, because they're not 'hardcore' enough to pay any more than the monthly fee.

They're also going to play on servers that their friends play on. All things equal, they are more likely to have friends on a high population server than on a low population server. Therefore, there is more incentive for them to join a high population server, resulting in similar proportions of casuals on any given server.

Gevlon said...


It says we killed Professor Putricide 25-men on Jun 25. We never killed PP25. We did not even raid on 25 men. Simply different people on different trade pugs killed it.

While "casuals go to friends" make sense, what makes the initial disproportion of friends? I mean the first players started on various servers.

IrOn TuRtLe said...

Out of curiosity, I looked up my guild and server (US-Kael'thas) info on both these sites. WowProgress stated about 4000 thousand raiders, nearly evenly split between Horde and Alliance, and a score of 4939, which would put it somewhere in the middle. Which sounds credible. But it showed my guild as being 0/12 (we are currently 7/12) and my raiding toon as still being level 80. So it clearly the site can't be relied on to have up-to-date overall information.

WarcraftRealms puts the total population of the server at 9000, again somewhere in the middle of the graph. But it shows an Alliance:Horde ratio of 1:8, which is preposterous. It does admit the Alliance data is unreliable, but given that it claims there are only 3500 level 85 Horde characters (which would mean that more than half of them raid) I find that data pretty dodgy too.

There are clearly huge problems with the data, and not limited to small servers. When the data is so unreliable that it might as well be random noise, then it is not surprising that there aren't any correlations.

Saying that since there aren't any better sources for realm population is not a helpful response. The lack of reliable data does not make this bad data somehow become good.

Coralina said...

My old guild moved realms because of population issues.

We were not an under populated realm by any means we were probably just above medium but of course with the factions split pretty much evenly on our realm we kind of had a lower raiding population to choose from than a less populated server that had 98% on one faction.

We ran intro trouble doing 25 man HC modes. We could clear 25 normal and 10 HC but the problem in 25 man was that we had a few weak DPS and Healers that we had to carry.

We simply couldn't recruit players of that calibre. I'd say we were top 3 for our faction but there just weren't enough quality players, that could commit to schedule to support that many 25 man guilds.

We moved to Silvermoon which had a much high population plus due to the imbalance towards alliance we had a ridiculous pool of people to choose from.

Here is the weird thing though...

We still found we were carrying 3 or 4 weak DPS and 2 or 3 weak healers in our raids and had trouble getting people of the calibre required for those 25 man HC modes. Whilst there were more good players on the server we were competing with more guilds aiming to clear content at that level in the recruitment market.

If we had merged with someone and kicked out the weaklings I guess we could have done better. I guess at the end of the day regardless of realm size, the talent is always spread slightly too thinly over X number of guilds.

Anonymous said...

Wrong unit of analysis. Server size doesn't matter; faction size does. Players on the same server but different factions can't raid together.

Wilson said...

All five of the servers in your last table are categorized by WarcraftRealms as having too few data entries on both Horde and Alliance side to be reliable. The populations are probably much larger, which would invalidate the argument.

Squishalot said...

@ Gevlon: If you look at the life cycle of a server, it starts off as a "New Players" server, where all new players will inevitably go to start off. Then, when it gets busy, it falls off the "New Players" list, and another one takes its place.

From here, a new player has two options. If they are a fresh new player with no contacts in-game, they will go to the new (less populated) server. If they have existing friends in game, then chances are, they will go to the old (more populated) server to be with their friends, thus making it even more populated.

If you follow this logic through across all servers, you will naturally find similar proportions of 'friends' across all the servers. The reason for this is because not all servers were created at the same time - they were implemented slowly over time, allowing such a demographic to build up.

Anonymous said...

"Please note that for casual play every server is equal and low pop realms have "recommended" green sign, while high pop are red. So why would casuals flock high pop servers?"

Equal? When I started to play WoW I started to play on a realm which the realm list suggested for "new players"! Later, I found out this realm existed since TBC, so it wasn't a new realm. I had no idea what realms were. I simply went for the first option. I wanted to try the game. Turned out the realm was full with scrubs and morons although it was a PvE realm allowing me a safe leveling experience (in hindsight, IMO not a good thing; world PvP is good exercise for skill). In dungeons, people often laughed at my realm, especially when I was playing bad. Only in the very start did I play bad, or when I did not know tactic. After that I played well. Remember, without heirlooms, it was not faceroll. I did not know the quality of the player base until I was max level, going to raid.

While not having BiS gear from heroic dungeons I was lucky enough to get in a PuG with a RL from a top guild who did not care about GS. Most PuGs were bad though. After a few "raiding guilds" (full with M&S, or not enough active raidings), I was doomed to PuGing. It was horrible. People weren't even inspected. IF they were, they often only looked at GS, or achi (can be faked). The two hardcore raiding guilds were not within my reach, and the casual raiding guild a friend of me was in did not require the class/role combination of my main character.

Unable to find a guild, leaving my woW peers behind (some joined me later), I peaked at WoWprogress and made a character on a top realm from the very same battlegroup. It was a realm where, according to what I observed, people were good players (met them before in dungeons). I started leveling up a character on this realm and investigated raiding guilds and PuGs on their trade chat. Rather sooner than later, I found the realm was active, people raided aiming for progress, the trade chat was much more mature. Many top guilds actually killed end boss on 10/25m HC. There was less childish behaviour, more business-like. Only thing was that PuGs required certain ilvl (GS back then), achievement, AND inspect. I realm transfered my main (and later alts) to this realm. In the guild there, I got more progress than on my previous newbie realm both in guild as well as on alts in trade PuGs. Now, my sample size is only 1. But for me it worked.

Anonymous said...

Is there anyone who's datamined Armory achievement data by realm? That would give pretty good results.

Anonymous said...

Or, alternately, use the number of valor boots generated as a proxy for number of boss kills. There should be some sort of relationship and this is accessible through the Undermine Journal. (Of course, heroics do cloud the issue.)

Anonymous said...

This conclusion is completely unsurprising; with every server hosting precisely the same game, what possible reason would there be for a concentration of skilled players to crop up in any one geographic or demographic region?

However, I'm not sure that avoiding high population servers is the best thing to do for a greedy goblin - if a fool and his money are easily parted, would a high population server be the best place to make a killing off the M&S?