Wednesday, November 28, 2007

Guesstimating numbers

Changed asked in a recent comment: "I am curious about your estimates for these percentages of WoW players (say the 2M in North America and Europe):

% of people with level 70's
% of people who have entered Karazhan
% of people who have finished Karazhan (killed Nightbane)
% of people who have entered 25-man TBC raids

Wowjutsu provides some numbers but doesn't indicate how many of those toons are on the same account / same player. Any thoughts?"


I posted a reply, but before I copy that here, I'd like to discuss the subject a bit. Estimating numbers based on fragments of data we have and personal experience is extremely difficult. For example Changed mentions 2 million WoW players in North America and Europe, while I thought that of the 9.3 million total players about 4 million are from North America and Europe. But those are just numbers I remember from the last time Blizzard published numbers by region, and this could easily have gone down to 3 million or less since then.

The best source of raid population data is WoWjutsu. They use data from the WoW Armory, which is much more reliable than relying on user input. But the best they can do is count "raid guilds", and declare every member of a raid guild to be a raider. This is only true for the top raiding guilds. There are lots of casual guilds with large memberships, of which only a fraction ever went raiding. WoWjutsu only counts raiders, their total number of "players" listed is the total number of people in guilds with Karazhan loot. Which is why they have 100% of "players" listed as having been to Karazhan.

One principal problem of all sorts of counting population in online games is whether you count players, characters, or characters actually online. For example if we wanted data on how many people play Alliance and how many play Horde, the sum of the two as expressed as a percentage of all subscribers is over 100%, because some players have both. If you count online players, you give more weight to the players who are online more hours a week, and less weight to those who rarely play. My estimates are based on the number of suscribers, not on the number of characters, or online characters. The number of raiders as expressed as a percentage of online characters would be higher, because raiders usually play more.

So here are my half guessed, half calculated estimates:

% of people with level 70's

I'd say about 70%. I've used CensusPlus and saw over 60% of people online on my server being level 70. But of course many of the lower levels online were alts, while on the other hand all census software undercounts very casual players who just aren't online.

% of people who have entered Karazhan

WoWjutsu counts 1.9 million raiders out of a population of roughly 4 million US and Euro players. But this counts every guild member of a raiding guild as a raider, even if half of them never went to Karazhan, thus I'd say around 30%.

% of people who have finished Karazhan (killed Nightbane)
% of people who have entered 25-man TBC raids

Those two numbers are probably close to each other. According to WoWjutsu 1/3 of all players entering Karazhan also entered Serpentshrine, so the number is probably around 10%.

Feel free to give your own estimates in the comments. But don't just throw around numbers, give some arguments with them and tell us why you think there are more / less raiders than I'm guessing. And remember, this is percentages out of total number of European and North American subscribers.

No comments:

Post a Comment