View RSS Feed

Blogenstein

Theorycrafting 101: Haste Buff Stacking

Rate this Entry
One of my long standing buff stacking rules has been to avoid stacking similar secondary stat buffs, ie: donít combine two crit buffs, or two haste buffs. This sort of rule doesnít apply to primary stat (Agi, Int, Str) or their power types (Attack/Spell). Iíve had this rule for a number of years now, so I decided to review the math on it, both to double check it and explain the reasoning behind it.

The methodology was fairly straightforward. Work out the cast time for a spell (in this case, Lightning Bolt) under various haste conditions (base haste 0-25% with Bloodlust, Elemental Mastery and then both Bloodlust & Elemental Mastery at the same time). I also ran the same comparison for the Global Cooldown as this will hit its haste cap at 50% which is fairly easy to hit.

Before we start, I should clarify something about how haste buffs work. The effect of haste lasts longer than the duration of the buff, as spell cast times are calculated at the start of the cast rather than at the end, so any time you want to work out the number of spells affected you divide the buff time by the new cast time and then round up to the nearest whole number.

The cast results were rouhgly as I expected, with a 0.5% increase in the number of casts over 6 minutes when under the ďhardĒ haste cap for Lightning Bolt of 100% during the stacked buffs.

So in this case, at 10% base haste Bloostlust adds 7 extra casts vs the 11 that Elemental Mastery adds (which is 4 per use, less one cast at the end). When stacking buffs, it adds an extra 4 casts to the Bloodlust time, but you lose 4 extra casts from the third Elemental Mastery use. So now weíve got 7 + 11 = 18 and (7 + 4) + (11 - 4) = 18 comparing non-stacked with stacked. The other part is that the non-hasted casts for stacking is 153 vs 142 for non-stacking. This difference is made up for by the third Elemental Mastery cast covering that 11 cast difference.

Things start to change when the stacked haste total passes our hard haste cap of 100%. The slight difference is wiped out, and then the total haste creeps past 110% thereís a slight gain on the non-stacked side.

So, assuming youíre just casting spells like this it doesnít matter which way you go with using your haste buffs, but what happens when you use instant casts? Itís somewhat the reverse, as the Global Cooldown is only 1.5 seconds long so itís already capped when you combine two 30% buffs or one buff with 16% base haste.

Combining buffs results in 3-5 fewer GCD uses in 6 minutes, so given the fairly neutral overall picture for spell casts, itís better to avoid stacking your haste buffs. Of course, in certain situations itís preferable, but thatís usually an exception like a critical burn phase.

If you want to look at the spreadsheet I used for this, itís located here. Just be aware that itís not pretty, so it may be confusing for some.

Submit "Theorycrafting 101: Haste Buff Stacking" to Google

Comments

  1. Taser's Avatar
    You should also add the 5% spellhaste buff as we will always have it (maybe you just included it in base haste?). In your table you are hitting the 1 sec. casttime for LB with 18.3% base haste (7796 rating), but with the buff that would already happen at 12.7% base = 5401 hasterating.
  2. Binkenstein's Avatar
    Quote Originally Posted by Taser
    You should also add the 5% spellhaste buff as we will always have it (maybe you just included it in base haste?). In your table you are hitting the 1 sec. casttime for LB with 18.3% base haste (7796 rating), but with the buff that would already happen at 12.7% base = 5401 hasterating.
    Umm, this is just a pure theory exercise, so there's no reason to include that. The haste values in the table shouldn't be used for anything either.
  3. refjke's Avatar
    sorry for my english

    this spreadsheet means that gcd hard cap = 1 sec =)
    look: http://imageshack.us/f/253/bredu.png/
    all without roundup

    when cast time > 1 sec t=0.46%
    when cast time = 1 sec t change

    so when stack make cast time 1sec -> (Separate > Stack)
    when stack dont make cast time 1 sec -> t =0.46 with any number of Base Haste

    why you roundup?
  4. Binkenstein's Avatar
    Quote Originally Posted by refjke
    sorry for my english

    this spreadsheet means that gcd hard cap = 1 sec =)
    look: http://imageshack.us/f/253/bredu.png/
    all without roundup

    when cast time > 1 sec t=0.46%
    when cast time = 1 sec t change

    so when stack make cast time 1sec -> (Separate > Stack)
    when stack dont make cast time 1 sec -> t =0.46 with any number of Base Haste

    why you roundup?
    The GCD hardcap is 50%, but it's because the game limits it to 1 second rather than anything else.

    I'm not exactly what your spreadsheet claims as you've just posted a screenshot without any column headers.

    As for rounding up, that's with regards to temporary buffs. With most temp buffs when it expires the effect expires, so any spell that is mid-cast when it expires does not get affected by it. For this the effect is rounded down.

    With haste buffs, the benefit is calculated at the start of the cast rather than the end, so that last cast is still affected by the buff even when it expires mid-cast. For this the effect is rounded up.

    eg: if your buff expires mid way through the 10th cast, that cast is not affected by spellpower buffs, but as cast times aren't re-calculated mid cast, it is still affected by a similar haste buff. Thusly 9 spellpower affected casts vs 10 haste affected casts. The actual number calculated is duration/spellcast, so for a 15 second buff with a 1.6 second cast, 15/1.6 = 9.375, which then gets rounded down or up to give the affected spell count.
Leave Comment Leave Comment

Trackbacks

Total Trackbacks 0
Trackback URL:
Powered byvBSocial.com and MMORPG