2013-12-25, 07:32 AM
Actually, I need to add something on.
Before any 30ms-rounding function, the DPS increase increases with each successive speed step increase.
Look at it this way: skill delay function has (10 + x) / 16 in it, so DPS increase simply reduces to (10 + x) / (10 + x') where x' is the new speed stage. So 13/12 --> 12/11 --> 11/10 --> 10/9 as DPS increments are extremely significant (8.3% --> 9.1% --> 10% --> 11.1%).
The 30ms ceiling function screws things up, obviously, but that also implies that for some steps the DPS increase is smaller, and then for some steps the DPS increase to the next is tremendous. It also implies that for skill delays below a base of 16*30 = 480ms (480ms gives a steady increase like that just described because the ceiling function always returns the same number), there may be some steps that offer no DPS increase.
Hadriel
Before any 30ms-rounding function, the DPS increase increases with each successive speed step increase.
Look at it this way: skill delay function has (10 + x) / 16 in it, so DPS increase simply reduces to (10 + x) / (10 + x') where x' is the new speed stage. So 13/12 --> 12/11 --> 11/10 --> 10/9 as DPS increments are extremely significant (8.3% --> 9.1% --> 10% --> 11.1%).
The 30ms ceiling function screws things up, obviously, but that also implies that for some steps the DPS increase is smaller, and then for some steps the DPS increase to the next is tremendous. It also implies that for skill delays below a base of 16*30 = 480ms (480ms gives a steady increase like that just described because the ceiling function always returns the same number), there may be some steps that offer no DPS increase.
Hadriel