Reply to thread

Message: <blockquote data-quote="Dungeoneer" data-source="post: 5392397" data-attributes="member: 91777">My guess is that this kind of data is going to be most useful at the margins.  In other words, they should look for character options that EVERYbody picks and those that virtually NEVER get taken.If every Barbarian is a Rageblood Barbarian then that says that Rageblood is too good and that other barb builds need beefing up.  If nobody ever plays a Runepriest then that class isn't working and needs a major update.  But the numbers they base these on should be pretty extreme.  We're talking 90/10 splits here, not 60/40.  A 60/40 split might indicate people gravitate towards a certain archetype.  But if 95% of players make a dual-wield Ranger, you know there's factors other than Drizzt at play here. They should also be able to screen out the 'new and shiny' effect.  Obviously, most character options will see a spike in popularity just after release.  They should be able to get a handle on what this spike looks like and filter it out once they start to get some good historical data.</blockquote>