Here is how I think server merge and opening should be done.
For each region, a server is opened every 3-4 days, such that exactly 8 server are opened each month for each region, the first of which is on the 1st of the month.
After exactly one month since the 1st, 3rd, 5th and 7th of those server have opened, they are merged with the 2nd, 4th, 6th and 8th pairwise respectively, resulting in 4 servers.
After exactly two month of the opening of the 1st and 5th server, once again they are pairwise merged with their neighbor.
And finally, after three(or four) month of opening of the 1st of those server, they will all be together in one server.
So the end result is one server per month per region (in the long run)
Those merges will be automatic and in addition to existing merge framework (and not take up the "merge slots" so to speak)
Now, I'm sure someone want to ask me, why this complicated scheme when you COULD just have one server per month per region opening to begin with?
Well, from the player point of view, it's like this:
Not everyone is willing to wait a full month before joining the next "new" server. i mean some people might be willing to join the most recent one, learn about the game and THEN join the new server, but without a doubt there would be people who won't play the game unless they can get into a new server either right then and there or very very soon. And there are good reason for this, both in the form of being behind on progress than people who have already been in the server for potentially weeks AND server open events, which are often time limited and/or competitive, people would like to at least have a shot at those... probably.
Having a frequent open but auto merge structure will allow a player to play competitively in a "new server" for at least quite a bit of time before the server is potentially thrust into harsher weather--which they would have to be forced into if there was only one server a month anyway. Furthermore, if a player truly want to be competitive with minimal spending (or non at all), they could always wait for the start of next month for the new "batch" so he/she can have the time advantage on the rest of the batch, it's a known factor that can be planned for. The REAL potential problem of this structure is probably that #1 of each batch will be the most active while #8 might be a ghost town. Of course, #8 will also have less competition so ranking reward might help the players there to catch up with the majority of other server's population. Hopefully, there will be enough population to keep sage up until the auto merges.
From the "company" point of view, well like I said, there are more frequent server opening events which are competition, which makes people spend more money. Frankly, people joining a server after it's 2 weeks old and missing all the opening event are just unlikely to spend, compared to people who joined a new server. So this structure should encourage spending.
And now, for the merge of existing server.
First, I'd like to preclude servers that are less than a month old. Frankly, their population isn't stable nor is their power stable, so it could cause all sort of problems. Besides, a month is not exactly a huge amount of time to wait (chances are the server won't have that long to wait AFTER it become dead, because if a server is dead on week 1 or 2.... abandon ship would be a pretty darn sensible thing to do)
After that, the server with the lowest activity rating (however that is measured) will be selected as the merge "starter". With a "starter" selected, the top ftp power of all servers of that region will be used to find this "starter" a mate whose top ftp is closest in power to each other, and the activity rating of those two server will be added up. If this rating passes a certain threshold, then that's it, the two will merge and it will be finished. If not, they will find the next nearest mate, based on the average power of the top ftp of each server and add the activity rating of the three server together to see if they pass the threshold. This will repeat until the threshold is passed.
After this, this "set" of server is removed from the pool, and the next lowest activity server will be selected, and the process is repeated, up to however many groups of server they are willing to merge together in one batch.
Note that this does potentially have the possibility of merging one or more low activity server into one super highly active server, I think that's perfectly fine. because I don't believe we should exclude active servers from merge, as all that does is making matching servers of similar "strength" more difficult.
The entire process is also highly defined, because I really don't like it when server merge process have "human input". All that ever does is make people complain, because there will always people who feel they are getting the short end of the deal. Fairness can only be achieved by strict guidelines.
Of course, I'm sure there are factors that I haven't quite considered in this plan, so as it occurs, the guideline might be altered to prevent disasters. Still, a guideline should be maintained, infrequently altered(and at that only minorly to avoid disasters) and followed. This post was last edited by 117***@google at 2017-2-16 06:26