Kemi

Common sense tells one many things that are not true. I've already had my common sense overturned by my simulator results more than once, and I'm not interested in arguments based on common sense. I'm interested in observed numbers and mathematical proofs.

But perhaps I've expressed my concern badly. Consider the following, please:

1) No matter how much haste I have, 15.7 haste rating will always increase my DPS by 1% (setting aside, temporarily, the question of mana issues).

2) If I have zero shadow damage, adding 100 shadow damage increases my DPS by something like 20%. If I have 1500 shadow damage, adding 100 shadow damage increases my DPS by something like 5%. That's a pretty drastic drop-off in the effectiveness of shadow damage at increasing my DPS.

3) If I have zero crit, adding 1% crit increases my damage by more than 1% (because of ISB). If I have 99% crit, adding 1% crit increases my damage by 0.5%. Again, 50% less effective at 99% crit than at 0% crit.

So, just from looking at this, it would seem that highly geared people would want to invest more in spell haste, because it continues to offer the same percent increase in DPS (and applied to a higher base DPS, even!) per rating point, while both spell damage and crit offer smaller percent increases in DPS per rating point as gear improves.

But that's not what the simulator is showing, and that's because, as far as I can tell, the synergistic effects of each stat on the others are non-linear.

It's that precise non-linearity that I'm having trouble modeling (mostly because my practical math knowledge stops at algebra, in spite of a painful year spent in Calc 1).

Here's the thing--as far as I can tell, when evaluating the effect of 1 point of crit rating or 1 point of hit rating, it matters not at all whether I have 0 haste rating or 1000 haste rating. But when evaluating the effect of 1 point of shadow damage, it matters a great deal.

If you have a lot of spell haste, you will see a bigger percent increase in DPS per point of shadow damage. But the percent increase in DPS per point of crit is not affected by how much spell haste you have.

And that's why it matters whether "spell haste gets worse the more you have" or "spell damage gets better the more spell haste you have". Because if the former is true, then spell damage and spell crit remain in the same relationship with each other, and spell haste falls behind them. But if the latter is true, then spell crit and spell haste remain in the same relationship with each other, and spell haste pulls ahead of them.

This seems to be what my simulations are showing: that as you get more spell haste, spell crit and spell hit start to lag behind spell damage.

Thoughts?

 

Vlar

Check the spreadsheet. I am mid-way through Tier 6 content (5/5 MH and 5/9 BT) and have not had immolate in my rotation for a while.

 

Sylvannae

My guild currently doesn't use curse of recklessness, though granted we do not have warriors with improved demoralizing shout. Is this wrong? I understand that CoR is a 5-6% increase in physical DPS, but just how does the increased AP evaluate? I am under the assumption that only one +AP and only one -AP can work on any single target, though this may be wrong.

Curse priority as far as I understand is CoS, then CoE, then CoR. Should we be using CoR regardless of the fact that we do not have improved DS?

 

Vlar

You would need to answer a few questions before your final could be answered.

Note: Most, if not all of this is in the first post.

First, what is your raids average physical dps? (Hunters, Enh. Shaman, Rogues, Warriors, Feral Druids.) Remember that would include the tanks dps, since even though it should be miniscule compared to the others, it would increase as well.

Would that increase out damage your dps from Curse of Doom?

Do your healers have issues keeping the tank alive?

 

Arelenda

For crit I could accept that. ISB is very elusive and hard to model.

But for hit rating? Lagging behind? If we ignore dots, hit is equivalent to haste: it will allow you to land more spells in a given time frame. One stat will get you less resists, while the other lets you cast more spells.

If you increase hit chance from 90% to 99%, you should get a 10% increase in damage, regardless of your crit, haste, or spelldamage. If your simulator tells you otherwise, I'd go look for programming errors.

 

galzohar

Your simulator seems to ignore a lot of factors that the spreadsheet is handling well. Unless there's a serious flaw in the spreadsheet that you found that makes you not want to use it (in which case, point it out) - you're probably better off using the spreadsheet since it will take into account things that you ignore.

When you generlize things and ignore small corrections like ISB effects and lifetaps (and yes, they're small but far from neglicible which is why we use the spreadsheet), DPS = (base damage + dmg X coefficient) * (1 + crit chance) * hit chance * (1 + haste increase). Understanding that basis tells you everything you need to know about how stats scale in general. Not the most accurate since it ignores some small corrections, but it should get the idea through of how stats scale. If you look at the relative DPS you just calculate new dps/old dps - 1, which means any stat you left constant will not matter since it appears both in the dividor and the dividend with same value (excuse my english if I used an improper term )

 

Trickykid

The spreadsheet, and simple intuitions based on a relatively small number of factors in a spell-spamming scenario, show every other DPS-contributing stat increasing in DPS/point if you raise haste. Haste drops by a small amount in DPS/point (due to nonzero MP5). The simple intuition is that haste speeds up the entire rotation, which means all of the other stats will be used more per unit of time. If I cast 20 spells a minute and haste myself to 21: crit, hit and +dmg all affect one more spell per minute.

 

PSGarak

Kemi: X% compared to what? Percent changes aren't necessarily the best thing to be looking at anyways. For a fixed crit/hit/haste, 100 damage adds a flat amount of damage. That flat amount of damage is, like you say, 20% of the damage you would do at 0 spell damage and 5% of the damage you would do at 1500 spell damage... that would seem to indicate that the percentages are illusory.

The percent benefit by adding crit drops off dramatically, but the flat damage increase dropss off rather slowly. If it weren't for ISB, it wouldn't drop off at all, it would just be a static damage increase at fixed damage, just like +dmg is a static damage increase for fixed crit. ISB is, as we have noted, hard to model, but it is known to be monotonically decreasing.

Haste scaling works the same way. 100% haste gives you twice the damage of 0% haste; 200% haste gives you three times the damage. The first and second 100% haste have given you the same amount of damage. Hit and crit are capped at 100% hit/crit, haste is capped at 1.0s global cooldown.

How do these scalings interact? They multiply simply. The complicated part you're seeing isn't very far off from linear, except for crit. It's called affine linear: There are additive constants on all the terms, which get caught by all the other terms. It makes them look like they're behaving funky. Basically, when you try to change two variables at the same time (say hit and crit), there's the expected second-order term (hit*crit), but also two extra terms (crit*constant and dmg*constant) from the base damage and base crit chance. Asymptotically those extra terms become irrelevant, but the point where that's true is impossible so we have to muck around with them.

http://sig.gamerdna.com/quizzes/INFL...erks/Garak.png

 

Carnate

Quote from the Compendium

Curse of Recklessness - Stacks with every other armor reducing debuff, and will typically increase melee and hunter damage by around 6% if FF and 5xSunder Armor are present, according to [RAID] Boss armor values. You typically want as many armor reducing debuffs on a boss as possible, since each one adds more (Increasing Returns, as opposed to Diminishing). It will make the enemy hit slightly harder, though. To be avoided when the boss has an ability that is based on his melee attack power, such as Mortal Strike or Cleave. Useful on enemies that don't hit hard enough for a tank to keep aggro, it will help with their threat generation on those by a small amount.

CoD vs CoR: Which one is more profitable depends on the raid setup. More physical dps means CoR becomes better, obviously. Physical Raid Dps needs to exceed (CoD damage / 3.6) for CoR to be the better choice. (CoD dps = CoD damage/60, and CoR dps = Physical dps *0.06). Note however that you need to refresh CoD twice as often.


Simply put if you have more than 3333 physical raid DPS then CoR is better than CoD (assuming 12k CoD which is T6 lock range). So CoR should be up every fight. In most raid setups it's better than CoS or CoE.

Exceptions are those bosses that have special attacks affected by the attack power or who already have 0 armor. Illiadan & Gurtogg during their enrages are two times that we don't use CoR.

 

Raithlin

Can anyone confirm for me that a nether protection proc while tanking Illidan would in fact cause mayhem and carnage?

i.e. does a proc cause aggro to reset or would a lock spamming SP still be able to keep hold of him?

Thanks

The Horror! The Horror!
Profile

 

◊ Caffeine

It doesn't reset threat, but it makes you immune to his attacks for the duration of the proc and he will turn for 2nd on the aggro list and Shadow Blast him instead. If you are going to tank Illidan, either spec out of it or have a macro with /cancelaura Nether Protection, which you need to spam (but I don't recommend this, it's how I know for sure he will ignore the warlock tank and turn for 2nd on threat :P )

 

Flamingcloud

Heh.. One of our locks tried to tank illidian while spamming a macro to remove nether protection.. Didn't work, we wiped, everyone got mad at him =p

 

Kalle

This is wrong. Your DPS is
dps(haste_rating) = unhasted_dps * (1 + haste_rating/1577)

Increasing your haste rating from 0 to 15.77 (0% -> 1%) yields an increase of
dps(15.77) / dps(0) - 1 = 1.01 / 1 - 1 = 1%

Increasing your haste rating from 157.7 to 173.47 (10% -> 11%) yields an increase of
dps(173.47) / dps(157.7) - 1 = 1.11 / 1.1 - 1 = 0.91%

 

galzohar

I think the compendium needs to be fixed then, since CoD DPS is not just the damage divided by 60s, as you have to substract the shadowbolt DPS lost by spending that GCD as well, so the DPS increase is actually smaller.

 

Arelenda

The compendium already mentions this.. It's one GCD per two minutes.

 

galzohar

My bad then, I thought this was quoted from the compendium

 

Arelenda

It is. and it's right there.

 

Telyz

have you discusses if immolate and Incinerate is better then SB and what lvl gear would it be better?

 

galzohar

I suppose you mentioned double GCD usage, although it's probably more appropriate to also mention that the damage lost due to that extra GCD is simply 3/5 shadowbolt (plus some neglicible difference in mana consumption).

 

Eph

I mentioned this in another thread, but this is the Brutallus specific spec I used once we decided to use a SE Affliction lock. Talent Calculator - World of Warcraft

In the spreadsheet I found Ruin and Imp attacking (since he will live here) to be the highest dps setup. With an improved curse and Imp dps the rdps loss wasn't too bad. Though if your healers can handle it Destruction is obviously the still best choice for Brut.

 

Tacitus

Wouldn't 50% haste mean your GCD would be 1,5s/2=0,75s? Meaning that the haste cap for GCD is slightly over 500 or 33,333... %?

Brotherhood, Peace, Unity

Welcome back, Comrade

 

Trickykid

dividing by 2 would be 100% haste.

50% haste -> 1.5s / (1+0.5) = 1.0s, so the original post is correct.

 

Arelenda

Haste speed modifier = 1 / (1 + haste rating / 1577)

We're talking about 738 haste rating, resulting in 1/(1 + 0.5) for a reduction of 33%.

GCD is then 1s. Which is the cap.

 

galzohar

And that's only the cap for GCD spells, shadowbolts would gain from haste until you have 150% haste which is even more impossible to get. Not like the old GCD was a real limitation as back in 2.3 you still needed 67% haste to stop benefiting from it.

 

Arelenda

There is a post somewhere called "Warlock PVE raiding compendium", I suggest you check it out.