SimulationCraft for Warlocks (4.0.3a numbers at 80 and 85) - Page 71

supplicium · 03/29/09, 5:14 AM

Originally Posted by turturin

His helm, chestpiece, cloak, sundial trinket, ring, boots, and belt could all be upgraded and would provide significant boosts to avg spelldmg (by which i mean to include trinket procs as well as damage from gear). His gear is very good no doubt, but to assert that he couldn't make much improvement in the spell dmg department is wrong. There are warlocks with over 200 more spell dmg from gear.

With current BiS gear or lacking only 1 or 2 pieces, it is certainly possible for your demonic pact to be better than ToW even when taking the uptime into account according to your chart, you don't need any ulduar upgrades to make it viable as a replacement.

His helm, chest are for the 4pc bonus, as last I looked they only come from t7. The spelldmg increase you gain from changing this gear is negligible it's other stats you gain the most. The difference is some what significant but not as much as you may think.

While you're right that there is room for improvement, I think we can use his gear as a good "into ulduar" set. What we can expect most locks to be bringing to the Big U on 3.1 release day.

On the note of just leave ToW down. It doesn't work like that. Do you use CoE just in case earth and moon is down? You are talking about improving the elementals shaman by a good 2-300 dps min by allowing them to swap totems.

Also you are forgetting that 86.7% was done on a target dummy (not from simcraft) and as such, is a reference for fights where your doomgaurd never moves from the boss. In situations like Hodir, Freya, Thorim, Thorims gauntlet, razer scale, x32 deconstructor, Auriya, General Vayzex, Mimron, your target is going to be moving, or you will be target swapping, or you will have to micromanage your pet. IE that 86.7% will plummet for a great deal of the fight.

1 More thing, Many of the warlocks you mention have more spelldmg on their gear, but most aren't alchemists and factor in mixology, which is equivalent to ring enchants (or ~40 spellpower).

To compare the results from above, to what would come off a best in slot warlock; Maalakai from premonition (sorry for calling you out I just happen to know your gear is rather good). He has 2142 spelldmg but negates the 4pc t7 bonus also hes not an alchemist so we loose 154 damage not factoring in glyph of life tap which is additional 60 spelldamage on top of that with the t7 4pc bonus. which leaves malakai 10 spelldmg above what Motif is sporting and that's not including his trinket which can be safely assumed to be 98 spelldamage if it only procs once per 60 seconds.

Oh also ToW provides 3% increased crit chance to all enemy units with in 40 yeards, DP doesn't do that, Ret pallies do it to 1 target only, and that 3% crit is for all attacks not just us spell casters.

Warlocomotif · 03/29/09, 5:23 AM

Originally Posted by turturin

His helm, chestpiece, cloak, sundial trinket, ring, boots, and belt could all be upgraded and would provide significant boosts to avg spelldmg (by which i mean to include trinket procs as well as damage from gear). His gear is very good no doubt, but to assert that he couldn't make much improvement in the spell dmg department is wrong. There are warlocks with over 200 more spell dmg from gear.

With current BiS gear or lacking only 1 or 2 pieces, it is certainly possible for your demonic pact to be better than ToW even when taking the uptime into account according to your chart, you don't need any ulduar upgrades to make it viable as a replacement.

The boots are BiS if you can use the hit rating, and I can use the hit rating since 0/56/15 doesnt have hit talents.
The helm and chest would definitely be a downgrade for demonic pact purposes. Even if those pieces were to prove a DPS increase over 4xT7, they'd be a raid dps decrease.

Cloak, Ring, Trinket and Belt do need upgrades though. However probably more importantly, on PTR I don't have the turning tide or the surplus limb.

[edit]

On the note of just leave ToW down. It doesn't work like that. Do you use CoE just in case earth and moon is down? You are talking about improving the elementals shaman by a good 2-300 dps min by allowing them to swap totems.

It's more than that, even the worst totem (Searing totem) does 400 DPS, fire elemental totem does 600+ dps or more (900) when combined with heroism through it's uptime and Magma totem can do ~900 if there is something to be aoe'd.

• Zakalwe · 03/29/09, 10:40 AM

I made a few fixes to how pet special attacks were modeled - in particular, I made sure the sim doesn't let any of them be affected by glancing blows. I also changed lash of pain from being modeled as a melee special to a spell. The OP has been updated to reflect these changes.

Warlocomotif · 03/29/09, 5:31 PM

The difference after taking away 4xT7 and Alchemy is around 100, however that is disregarding a lack of 3% hit (79 hit rating), if you really want to argue that its better to be 3% below hitcap- I honestly expect that to turn out a raid DPS loss.

However if you take my gear, add Signet of the Malevolent and Pennant Cloak- you gain 60 spellpower with no loss of hit rating. So there's a 40 spellpower loss there, but a +3% hit chance gain.

Snidelyw · 03/29/09, 6:59 PM

Can anyone explain how the "Scaling Factors" in the OP are determined please? They seem to have changed quite a bit recently

Warlocomotif · 03/29/09, 7:44 PM

Originally Posted by Snidelyw

Also do these calculations take into account partial resists from level-based boss resistance? And is Chaos Bolt not affected by level-based resistance?

Yes Chaos Bolt is affected by this, and as far as I know theyre being accounted for.

Arnath · 03/29/09, 9:24 PM

Originally Posted by Snidelyw

Can anyone explain how the "Scaling Factors" in the OP are determined please? They seem to have changed quite a bit recently

As far as I know the scale factors are determined by running the sim with the base stats then adding some number (I think 100) of that stat and then re-running the sim. Difference in average DPS is divided by the amount of the stat that was added and that gives you the scale factor for that stat. They've changed a lot recently because the DPS values for each spec have been changing a lot.

Darkmantle · 03/29/09, 9:54 PM

Arnath is correct except about the interval of the change in the various stats that is used for determining scale factors.

From Dedmonwakeen's post here they are:

scaling_t::scaling_t( sim_t* s ) : sim(s), calculate_scale_factors(0) 
{
  for( int i=ATTRIBUTE_NONE+1; i < ATTRIBUTE_MAX; i++ )
  {
    gear.attribute[ i ] = 250;
  }
  gear.armor_penetration_rating = 250;
  gear.attack_power             = 250;
  gear.crit_rating              = 250;
  gear.expertise_rating         = -150;
  gear.haste_rating             = 250;
  gear.hit_rating               = -200;
  gear.spell_power              = 250;
}

So 250 is the standard change to the stats that is used.

• Zakalwe · 03/29/09, 10:22 PM

Actually, there's a new option to generate better scaling factors, which I've just started using for the OP. That may be why you've noticed them changing recently.

The new option runs two simulations for each stat, one with half of the delta added, the other with half of the delta subtracted. In other words, it "centers" the delta on the current value of the stat, meaning you get a better picture of scaling factors at the current gear level rather than at a future gear level.

Althor · 03/29/09, 10:26 PM

I recently added a new way to calculate scaling coefficients to SimCraft.
It can be accessed by setting: center_scale_delta=1

Using this mode all of the stats, with the exception of +hit rating and +expertise rating, will calculate 2 DPS points for every stat. One, with the stat increased by half of the corresponding value in the constructor (i.e. gear.crit_rating for crit rating) and the other decreased by the same amount.

So for example, using the default ranges, for the crit rating scale factor coefficients, it will run the simulator X times with the crit rating increased by 125 (250/2) and then X times again with it decreased by 125.
It then calculates the scale factor by taking the DPS of the first minue the DPS of the second and dividing by the full range.

The upshot of this method is that the scale factor given is more appropriate to that of the profile given, rather than at some point better than the profile given.

Both modes can have their uses.

*edit* Beaten.

Warlocomotif · 03/29/09, 11:59 PM

Just want to say that you forgot to include the 0/56/15 scale factors.

• Zakalwe · 03/30/09, 12:03 AM

That's on purpose - including demonic pact in the scaling factor runs artificially boosts the results for spell power and crit rating, because it increases the DP buff and the DP uptime. If you want scale factors for 0/56/15, just use the 03/52/16 ones, they shouldn't be substantially different.

Darkmantle · 03/30/09, 2:56 AM

Originally Posted by Althor

I recently added a new way to calculate scaling coefficients to SimCraft.
It can be accessed by setting: center_scale_delta=1

Using this mode all of the stats, with the exception of +hit rating and +expertise rating, will calculate 2 DPS points for every stat. One, with the stat increased by half of the corresponding value in the constructor (i.e. gear.crit_rating for crit rating) and the other decreased by the same amount.

So for example, using the default ranges, for the crit rating scale factor coefficients, it will run the simulator X times with the crit rating increased by 125 (250/2) and then X times again with it decreased by 125.
It then calculates the scale factor by taking the DPS of the first minue the DPS of the second and dividing by the full range.

The upshot of this method is that the scale factor given is more appropriate to that of the profile given, rather than at some point better than the profile given.

Both modes can have their uses.

*edit* Beaten.

Will that mean a doubling of the time to generate scale factors since you need to do 2 runs with high iterations?

• Zakalwe · 03/30/09, 3:25 AM

Originally Posted by Darkmantle

Will that mean a doubling of the time to generate scale factors since you need to do 2 runs with high iterations?

Yes. Though if you're short on time you'll always have the option of using the old method, of course.

Grevann · 03/30/09, 4:20 AM

Originally Posted by Zakalwe

That's on purpose - including demonic pact in the scaling factor runs artificially boosts the results for spell power and crit rating, because it increases the DP buff and the DP uptime. If you want scale factors for 0/56/15, just use the 03/52/16 ones, they shouldn't be substantially different.

I'm not sure I follow this logic, there are many talents that increase the value of spell power, or crit. How does Demonic Pact work differently than Shadow and Flame to increase the value of spell power?

And isn't increasing the importance of Spell Power and Crit kind of the point of Demonic Pact?

Perhaps you mean there's something in Simcraft itself that doesn't work quite right, if so, please explain.