The reason is simple: 1.5 second (or instant) spells scale at 75% of 2 second spells when looking at haste.
0->25% haste for example.
0% Haste: 4LB, 1 LvB
25% haste: 5LB, 1 LvB
With 0% haste, we have 4x2 and 1x1.5, but with 25% it's 5x1.6 and 1x1.2
Both spells see a 25% dps gain when compared against the 0% version, but an overall view gives a slightly lower figure.
Assuming LB = 5k, and LvB = 10k, we see the following:
4x5=20 + 10 = 30k/9.5 = 3158
5x5=25 + 10 = 35k/9.2 = 3804
3804/3158 = 20.455% increase
Same situation, but assuming LvB has a 2 second cast time.
30k/10 = 3000
35k/9.6 = 3645
3645/3000 = 21.5% increase
As you can see, the overall DPS gain increased, but not to 25%. This is because the proportion of lower damage spells increased, while the higher damage spell remained the same.
4*100% + 1*75% = 4.75% effectiveness for 5 spells, or 95% effectiveness. If we assume that LvB has the same damage as LB in the first example, we'd see:
23.9/25 = 95.6% which is fairly close to the 95% mentioned previously (I'm allowing for rounding errors here, as I've just been using a calculator and writing most of the figures down).
So this means that LB scales better than LvB, but LvB haste scaling should not be looked at in terms of it's cast time plus the cooldown. Notice how this situation will be true for any situation involving two spells, irrelevant of whether there is a cooldown involved or not.
The reason is simple: 1.5 second (or instant) spells scale at 75% of 2 second spells when looking at haste.
This is total bullshit. 1.5 second spell will get exactly the same percentagewise increase from haste as a 2 second or a 10 second cast, excluding GCD which is not present here. The only thing important here is that the high damage, more efficient spell, has a cooldown. My last post had a reasonably simple explanation of why. You even say it yourself in your post. Part of the damage increase from haste is offset by the fact that you will spend a higher part of the time casting less efficient spells.
The effect is still quite small, but it's definitely there.
This is total bullshit. 1.5 second spell will get exactly the same percentagewise increase from haste as a 2 second or a 10 second cast, excluding GCD which is not present here. The only thing important here is that the high damage, more efficient spell, has a cooldown. My last post had a reasonably simple explanation of why. You even say it yourself in your post. Part of the damage increase from haste is offset by the fact that you will spend a higher part of the time casting less efficient spells.
The effect is still quite small, but it's definitely there.
2 second spell with 25% haste = 1.6 seconds (0.4 seconds saved)
1.5 second spell with 25% haste = 1.2 seconds (0.3 seconds saved)
0.3/0.4 = 75%
However, if we compare the DPSC of the two spells we see, for an arbitrary 1000 damage spell
1000/2 = 500 dps
1000/1.6 = 625 dps
625/500 = 1.25
The trouble is that with percentages, 25% of one this does not necessarily equal 25% of another. Casting less efficient spells more frequently is part of the reason why we see a % gain that is higher than what we would otherwise expect from the cast times alone, and why I used examples showing how the same damage with different cast times and the same cast times with different damage amounts will affect things.
It was me that you had the previous discussions with concerning the scaling of LvB with haste Bink and we didn't reach consensus other than to determine that you are correct that the DPSC of LvB does indeed scale perfectly with haste. However, my point was that the damage caused by LvB during a fight does not and that this is what scaling should refer to.
Since then I have thought of a hypothetical situation that demonstrates better my point of view. Supposing that Blizz added a talent that enabled the LvB CD to be effected by haste so that with sufficient haste, your CD is reduced from 8s to 6s. Now, the DPSC scaling has not been effected at all (so you would say that the spell scaling is unaffected) but the spell is now clearly effected in a beneficial way by haste. If we can agree on this then the CD of a spell does need to be considered when refering to scaling as saying 'a spell is effected in a beneficial way by haste' is just a long hand way of writing that it 'scales with haste'.
You didn't really. You pulled some numbers that were more or less irrelevant. If the spells have the same dps, the gain will be exactly 25%. If the cooldown/buff spell have more dps, it will be lower. If the cooldown/buff spell has a lower dps (scorch), the increase will actually be higher than 25%. The specific damage or cast times of themselves by themself is irrelevant, what is important is the dps of the spells.
When considering how a spell scales, you should consider spell damage over spell time.
When considering rotation scaling, you should consider rotation total damage over total rotation time.
Considering spell damage over total rotation time is not going to give you accurate information on how a spell scales, irrelevant of whether there is a cooldown or not.
Thusly, LB, CL and LvB all scale well with haste. FS does not because most of the damage comes from the DoT portion.
The rotation on a whole does not scale with haste with 100% effectiveness because of the 1.5 second spells used, as well as the dot being used.
Neither will the rotation scale with crit with 100% effectiveness because of the 100% crit chance for lvb, and the fs dot not being affected by it.
Definitions: Scaling of a spell/rotation relates to the increase in DPSC (damage per second casting) by increasing a select stat. (Saying that it's just damage per second raises the question of damage per second doing what?. Generally this is DPSF: Damage Per Second during Fight)
When considering how a spell scales, you should consider spell damage over spell time.
When considering rotation scaling, you should consider rotation total damage over total rotation time.
Considering spell damage over total rotation time is not going to give you accurate information on how a spell scales, irrelevant of whether there is a cooldown or not.
Thusly, LB, CL and LvB all scale well with haste. FS does not because most of the damage comes from the DoT portion.
The rotation on a whole does not scale with haste with 100% effectiveness because of the 1.5 second spells used, as well as the dot being used.
Neither will the rotation scale with crit with 100% effectiveness because of the 100% crit chance for lvb, and the fs dot not being affected by it.
Definitions: Scaling of a spell/rotation relates to the increase in DPSC (damage per second casting) by increasing a select stat. (Saying that it's just damage per second raises the question of damage per second doing what?. Generally this is DPSF: Damage Per Second during Fight)
I actually agree with most of what you say here except your statement that 'considering damage over total rotation time is not going to give accurate information on how a spell scales'. It will certainly not give accurate information on how the DPSC of a spell scales but it will give accurate information regarding how the overall damage caused by the spell scales.
If we went back over this recent exchange and replaced 'Scaling' with 'DPSC Scaling' or 'Overall Scaling' as appropriate (or some such names) I think we will probably find that everyone is in agreement and that the arguement is over the defination of the word scaling which is dependent upon context and the use to which it will be put.
When considering how a spell scales, you should consider spell damage over spell time.
When considering rotation scaling, you should consider rotation total damage over total rotation time.
Considering spell damage over total rotation time is not going to give you accurate information on how a spell scales, irrelevant of whether there is a cooldown or not.
Thusly, LB, CL and LvB all scale well with haste. FS does not because most of the damage comes from the DoT portion.
The rotation on a whole does not scale with haste with 100% effectiveness because of the 1.5 second spells used, as well as the dot being used.
Neither will the rotation scale with crit with 100% effectiveness because of the 100% crit chance for lvb, and the fs dot not being affected by it.
Definitions: Scaling of a spell/rotation relates to the increase in DPSC (damage per second casting) by increasing a select stat. (Saying that it's just damage per second raises the question of damage per second doing what?. Generally this is DPSF: Damage Per Second during Fight)
As i said, whether you want to say that LvB scales badly with haste because of it's cooldown or that Elemental shamans scale badly with haste because of LvB's cooldown is just semantics.
The fact that some spells are 2 sec and some 1.5 is totally irrelavant. 1.5 spells sec scales just as good with haste as 2 sec spells, even if they have cooldown. Let's look at yet another example.
Another way to look at it: 0 haste gives you 1.5 sec of good dps followed by 8 seconds of lower dps, letting you do good dps 1.5/9.5 of the time. 50% haste means you do good dps 1 second followed by 8 seconds of lower dps, meaning you have good dps 1/9 of the time. Let's say LvB is 6k dps and LB 3k dps unhasted. With 0 haste we would do 1.5/9.5*6k+8/9.5*3k = 3474 dps. With 50% haste we would do 1/9*6k*1,5+8/9*3k*1,5 = 5000 dps. This is a 43.9% increase in dps. If we had just been spamming one spell we would have gotten 50% dps increase.
If LB does same DPS with 0 haste and 50% you have reduced your other stats by an equal amount. If you increase your haste, your spells will do more DPS if they have the same damage per hit.
The short explanation to why you only take the cast time and not the cooldown into account would be that you still cast spells affected by haste during the cooldown. You don't just stand still doing nothing.
Might have gotten something completely wrong here though.
I'm not really following you why the dot doesn't scale with haste (when considering it in the rotation)? It will take less time to "cast" since the GCD will be shorter so you can start with the next spell faster. The duration is unaffected though, but it will be higher DPSC, which in the end will allow you to spend more time casting the more powerfull spells.
In terms of DPSC Lava Burst would easily be our best scaling spell with haste.
However, as has been pointed out, you can't fill that extra time with Lava Burst damage, due to the cooldown - the additional damage comes from our other spells (mostly Lightning Bolt).
In relative terms, if we were to look at Lava Burst in isolation, this would appear to be much worse scaling since Lightning bolt dps is much lower than Lava Burst - however, in absolute gain, it is (to simplify a little) the same as Lightning Bolt scaling.
Lightning Bolt would be our worst scaling spell with haste, except that its the only one that we continually cast.
So over a rotation haste will scale us up at just over our Lightning Bolt's average dps - all spells contribute to this including Lava Burst and Flame Shock.
Haste needs to be considered over a full rotation, as haste will cause that rotation to change, it should also be considered through Bloodlust, where it will most likely cause all our spells except Lightning Bolt to hit the 1 second global cooldown.
Despite these apparent setbacks to its effectiveness it is still a far better modifier than crit, in part because the rating cost of crit is so high, in part because there is so much 'free' crit available to us from raid buffs and in part because crit provides no benefit to our highest damage spell (where we stood to gain the most from it).
With 3x 32 spellpower JC prismatic gems and LW 67 spellpower bracer enchant enchant I got that the Wanton leggings set gives you 2072 spellpower, 536 crit rating and 499 haste rating. And using Benefactor and T7 leggings give you 1998 spellpower, 594 crit rating and 516 haste.
So putting those stats in to Binks spreadsheet I got a 5216 dps with Benefactor/T7 and 5256 dps with Wanton leggings.
So guess if you got every best in slot item like mace/ring from Kel'Thuzad ect, going for the Wanton leggings/JC ring will give you a small dps boost.
But for me at least I'll stick with my original plan to use all mail items with the Benefactor gloves, since I still need to focus on getting mace/ring from Kel'Thuzad and the cloak/trinket from Sartharion.
- My crit rate is obviously a LOT lower and my fight time a little longer, hence a lower % BL uptime. Both will lower my DPS but are effectively outside my control.
- I cast 109 CL & LB, to Syvha's 108 (I haven't tried to separate out LO proccs). Is that because I got less LO proccs, more latency or am just casting slower?
- I don't use CL as much, I don't tend to use it at all during BL.
- I lack a Boomkin or a DK, so I didn't have full group buffs.
I am still doing 3.8k+ DPS which isn't terrible, but if anyone has any ideas for improvement...or is it factors outside my control?
Edit: Had terrible lag on Patchwerk 25 tonight, and slipped to 3.45k with an even worse crit rate. I nearly cried .
It could also be server lag issues - a factor outside your control. During a 10 man naxx yesterday, about 6 server my dps went way down due to lag spikes. Luckily it didn't get really bad until Saph, but it certainly made for an interesting last two bosses.
Edit - I was not 100% clear. If your guild raids on a more crowded server at peak times, you will see less dps than a guild raiding on a less populated server at off times.
On our server the lag issue has gotten bad enough that one of our top two raiding guilds has called off roughly 50% of their raids in the last 2 or 3 weeks.
I am curious if there is any way to push DPS past 4.1k in the current version of the game. The scaling fixes in 3.0.8 seem quite nice, but for now it makes more sense to focus on the present as there is no way to determine when the patch will be released.
WWS Armory -Logged out in my elemental set, should show up shortly.
Base stats if armory does not update:
Spell: 2400 (roughly)
Critical: 24% depending on +hit set being used, obviously raid buffs push this much higher on boss fights, so it's useless to stack
Haste: 637
Hit 290-366
Your gear is pretty much as good as it's going to get--your hit is kinda low but assuming 40 hit food it's fine. I can buy that the low 4ks is about as good as we're going to get in terms of pre-3.0.8 DPS, assuming reasonable crit rate and average kill times. I just wish I had your luck with drops and rolls.
I've read this from page one, and forgive me if I missed it... but I keep seeing 2 red +spell power gems in the Tier7 chest on peoples armory profiles.
How are you all doing this?
Is it a bug in the armory? Did I miss something obvious?
01/13/09, 7:29 PM
.
Similar Threads
