Posted 18 March 2009 - 08:54 AM
Assuming unlikely proposition that it would be decided the dot refreshes WILL change, and that it would be YOU (and I mean all of you) who will decide how, which option would you prefer:
1) Tick like corr/EA: every time you refresh a dot, its duration is reset, while ticks happen completly unchanged at 3sec intervals, their coeff recalculated with stats taken at the moment refresh landed. CoA is unchanged
2) Tick like corr/EA with a tick queue: same as above, only remaining (old) ticks will tick with old amount, new coeff is recalculated when refresh hits, but it doesn't come to play untill old duration runs out.
Example CoA, no refresh
1 2 3 4 5 6 7 8 9 0 1 2 3 <- tick count
WWWWM MMM S S SS 0 <- actual ticks, every 2 sec
With refresh 7 seconds before CoA ends:
1 2 3 4 5 6 7 8R9 0 1 2 3 4 5 6 7 8 9 0 1
WWWWM MMM S S S S WWWWMMM M 0
W- weak tick M- medium tick S- strong tick 0-CoA has expired already, no tick R- refresh hits
3) tick more often - for example every 1 or 0,5 or 0,25 sec. This will just lessen impact of accidental fractionally early refreshes (such as because of eradication). The problem is with some "on tick effects".
4) refresh slack - if you refresh dot early by a set amount (ie 0,5 sec or even OPed 1 sec) refresh hit will get delayed by server, and hit precisely at 0,00000 mark - this will of course >require< you to refresh dots within that window (as far optimal dpsing goes), mess with dot timers, but completely remove any negative aspect to eradication procc.
5) tick stack - as long as refresh hits between last tick and one before that, ticks will be stacked. That means dot will be refreshed early but at refresh +3 sec (or 2 with CoA) there will be ~2x tick (or 4x if pandemic proccs). ~ of course refers to variance to tick power betwen different casts. This again will force all to refesh within "2/3 sec to end" window, but should help immensely with eradication and cast queueing.
6) yet someting else - do tell
7) I do not want any change at all