Talk:Spell damage comparison

Untitled
To do: Add blasting spells to AoE & Channelled spells. --Tbannister 11:38, 25 Jan 2006 (EST)

It looks like Hunter traps may be bugged. They should receive +spell damage modifiers but I've been told by a hunter player that testing showed they did not actually receive any bonus. Additional testing is needed to confirm result. --Tbannister 13:40, 16 Feb 2006 (EST)

Global cooldown is 1.5 seconds. Look at practically any spell on thottbot for confirmation: http://www.thottbot.com/?sp=25345 http://www.thottbot.com/?sp=10151 http://www.thottbot.com/?sp=18809 --Tbannister 13:34, 24 March 2006 (EST)

What formula is being used here to calculate DPS? I'm reviewing the warlock spell charts in particular.

The DoT formula seems to be damage / duration. This is fine as long as duration > cooldown, otherwise cooldown should be used.

What confuses me are the DD and Instant formulas.

DD & Instant should be: damage / (the greater of) cooldown, duration, or cast time. However this does not seem to be the case.

For instance: I don't see how Death Coil, which can be cast once every 120s for 470 damage could possibly have a DPS of 313.

My calculation for Death Coil would be damage / cooldown or 470 / 120 and so 4 DPS.

Could someone please explain?

--Tiris 14:13, 31 March 2006 (EST)


 * The reson cooldowns are not counted for DPS is that you can keep casting other spells while waiting on the cooldown. Death Coil alone my only add 4DPS over 2 minutes, but that is 4 DPS above what you will be doing normally. --Stfrn 17:45, 31 March 2006 (EST)


 * Fair enough - but that *is* the DPS for Death Coil. Casting Death Coil will do 4 DPS. Casting Death Coil in combination with other actions can give you any number of possible DPS readings... where does this 313 in the chart come from? If you want to know what sort of DPS those various combinations can give you - do you not first need to know that Death Coil's base DPS is 4, not 313? --Tiris 18:30, 31 March 2006 (EST)


 * Think of it this way - if you cast Shadow Bolt, it's got a (typically) 2.5s cast time. Those 2.5 seconds are being spent to gain the damage of the Shadow Bolt, because during the cast time, you can't do any other (direct) damage. Hence, the DPS for Shadow Bolt is damage / 2.5. Death Coil costs 1.5s, the global cooldown time, of no-other-direct-damage time. Rank 9 Shadow Bolt averages 510 damage, for 204 DPS over the 2.5 second cast time. Rank 3 Death Coil averages 470 damage, for 470 / 1.5 = 313 DPS over the 'cast time', which is just the global cooldown. The idea behind analyzing spells in this way is to answer the question "If I don't care about mana cost or tactical uses of spells, how can I pump out the most damage possible?" If you just care about damage, and you've got a choice between Shadow Bolt 10 and Death Coil 3, Death Coil has a higher DPS - 313 compared to 204. 1.5s later, it's time to ask the question again: "What can I cast to maximize my DPS?" Except now Death Coil isn't an option, because it's on cooldown. So you pick the spell with the highest DPS that's available.
 * Of course, spells with a large DoT component skew things a little bit, because people tend to talk about them differently. Corruption rank 7 only does 46 DPS as listed in the chart, but it's 46 DPS that runs in the background while you're doing other stuff. Doing the same direct-DPS maximization as above, we need to consider the cast time, not the running time, of Corruption. Assuming you don't have any talent points in it, it's a 2s cast time for 822 damage at Rank 7. That's 822/2 = 411 damage per casting seconds, showing that yes, it's worth your time to stick things with Corruption before you start bolting them, if you expect the fight to last long enough to get the full damage out of Corruption. --Esselte 03:04, 14 June 2006 (EDT)

Are channeled spells calculated as DoTs for +dmg bonuses?


 * My impression is that channeled spells count as normal casting time, but then the bonus is split over the whole effect even if the channel ends early. So most channels are longer then 3.5sec, but lose in other ways- AoE get only 33%, or the spell will end early, etc. --Stfrn 17:45, 31 March 2006 (EST)

Channeled spells do damage over the length of the channeling, they also receive lower bonuses from +damage. For spells that are AOE they start with the lower dmg bonus from being channeled, if they have any secondary effects like slowing from blizzard or the slower attack speed from hurricane, then they are hit with another slight reduction from +damage.

Error
Arcane Missiles rank 8 is unavaible from trainers

Verifying data, and a few questions
So I went through and ran the numbers on the +spell damage coefficients (referred to on this page as Scale Factor), based on the info over at Formulas:Plus_damage_and_Plus_healing, both because I wanted to see the how the spells that have both instant damage and a DoT portion deal with things, and to check some derived numbers (which I'll talk about in a sec). Most everything lines up, but there are a few exceptions, some of which I know the explanation for, and some of which I don't.


 * Moonfire: From the DPSSF values, this splits into 20%+60%, but I got 15%+65%, or 10%+5.43% in DPSSF terms.
 * Mind Flay: Listed here at 45%, but I got 57%. (60% from only being three ticks, x95% for having a second effect.)
 * Flame Shock: From the DPSSF values, this splits into 32.25%+50.4% for a total of 82.65%, but I got 21.3%+50.2% for 71.5%, matching the SF value listed here. DPSSF values would be 14.2%+4.18%.

Everything else matches up. I'd be interested to see the source data for these numbers, as it's quite possible that they're exceptions that need to be added to the other page.

Now, I got started on this because I wanted to add a column or two to the tables that would give a more conceptual version of the DPSSF values. The basic idea of +spell damage is that it's like attack power for spells, in that it adds linearly to your DPS, or it's supposed to. Looking at a the simple damage spells (being the ones that all tie for first place in DPSSF), it's clear that in the same way that 14 points of attack power adds 1 DPS to your white physical damage, 14 points of spell damage adds 4 DPS to your basic damage spells. I'm guessing the reason it's 4 instead of 1 is because melee classes get attack power from strength and agility, and so have a decent chunk of the stuff already without stacking up +attack power bonuses, caster classes have 0 +spell damage unless they specifically have equipment that adds to it, but to confirm that we'd have to ask the devs directly. Regardless of the reasons, I'd like to see a column with DPSSF * 14, which gives nice round numbers, or DPSSF * 14 / 4 (or DPSSF * 3.5), which shows the effectiveness of +damage on the DPS of the spell when compared to the basic 4 DPS per 14 damage that many spells get.

The other change I'd like to see is DPS split into DPS for the duration of casting and DPS from any DoT component, as well as a separate column for total damage / casting time. DPS is already shown as instant + DoT for hybrid spells, but it's impossible to distinguish spells that are purely DoT from those that are instant damage with some lasting secondary effect, such as Frostbolt.

I'm writing all this here rather than just going in and fumbling around on the page proper because there's obviously been a lot of work put into this, and I don't want to spend the hour or so on my end that it would take to make these changes if it's just going to result in "wtf, what're all these horrible changes? *rollback*" --Esselte 07:37, 14 June 2006 (EDT)

Were these questions answered?
I am seeing issues with the Scaling Factor of the druid spells also. Based on the information in the Formulas:Plus_damage_and_Plus_healing (which I just reworked) some of these numbers are innaccurate.

The spell hurricane for example has a SF of 31.7% because it gets a 5% penalty for having an attack speed reduction. The cast time is 10 seconds so it caps at 3.5 for purposes of the equation.

((3.5 / 3.5) - .05) / 3 = .317

Insect Swarm also has a 5% penalty for having a 2% decrease to hit ability, so its SF would be 95% not 100%. There are more than 5 ticks so it caps at 5 for the purpose of the equation.

(5 / 5) - .05 = .95

Moonfire is complicated. "189 to 221 Arcane damage and then an additional 384 Arcane damage over 12 sec." Usually people use an average of the low and high for the standard portion damage. Then you have to balance the +damage between the two halves of the spell.

( 189 + 221 ) / 2 = 205 standard damage portion average

384 total tick damage at 3 sec per tick for 4 ticks total

Portion to the Over Time part of the spell: 384 / (384 + 205) = .651

Portion to the Regular part of the spell: 1 - .651 = .348

Now apply the rules for Regular and Over Time spells. So we have to now consider the cast time of the spell for the Regular part and the number of ticks for the Over Time part.

( 1.5 / 3.5) * .348 = .149 Benefit Regular part of the spell

( 4 / 5) * .651 = .521 Benefit Over Time part of the spell

So rounding to whole numbers 15% for the initial part and 52% for the DoT part would be correct, not the other numbers listed.

--Mynia 17:30, 19 September 2006 (EDT)

Removed column DPSPM.
I removed the column DPSPM because it did not have any meaning. If you do not agree - please explain what use is there for such a strange value.

It's a comparison that utilizes DPS and DPM that "could" be useful for regular grinding or prolonged spellcasting, since while you would like a high DPS, you also have to take into account the DPM you're using up. Sure a nice 2-shot kill would be nice to do on every monster, but that probably means you're also spending up a high chunk of mana on every one of them. Pzychotix 01:17, 17 July 2006 (EDT)

Sorry for not signing my first message. Look. This values can not be usefull in any way. See example: Let us have two imaginary spells:

There is no difference, which of two spells I use because it's obvious that two casts of Spell#2 are equivalent to one cast of Spell#1. But, DPSPM values DO differ. So, what does DPSPM show? --Rotsor 01:35, 17 July 2006 (EDT)

Efficiency Column
I did some tests in excel for a new value which shows overall spell efficiency, the value is basically DPS*DPM, the higher the value the better the overall spell efficiency. I tried it on the druid spells Wrath, Starfire and Moonfire and it showed good results with Moonfire being the most efficient then Starfire then Wrath.

So maybe it will be a good idea if someone will edit the page and add this value to the tables. --Rani 15:06, 29 July 2006 (EDT)

Hello, Rani. I do not agree with you about the efficiency column. DPS*DPM is better than DPSPM, but not good enough also.

Let us have four spells now:

You can see that these spells have the same "DPS*DPM", but they have nothing similar in fact.

For example, you wil always prefer Spell#0 over Spell#1 because they both have a negligible mana cost, but their DPS differs alot.

Similarly, you probably will always prefer #2 over #3 because you cant afford 2000 mana spent every second. So, DPS*DPM does not show anything.

I propose another method for determining efficiency of a spell. This method is based on mana recovery rate. If you are casting a spell with a low DPS - then you are losing mana, which you could have recovered during the casting time.

Lets take our mana recovery rate as 100 mana/sec.

Then our table becomes as follows:

Here we can see that Spell#2 is the most efficient spell assuming you have a mana recovery rate of 100 mana/sec. And that is true.

For "mana recovery rate" i take the rate at which you recover you mana while not casting spells. So, if you always cast spells non-stop during combat and then recover mana out of combat drinking water - then your mana recovery rate will be high (regen + water). But, if you run out of mana during combat - then you should assume your mana recovery rate to be low (regen only).

Unfortunately we can't calculate efficiency for any given spell because different characters have different mana recovery rates. --Rotsor 23:46, 21 August 2006 (EDT)

In all this maths, I think a figure that would be useful and more understandable is a figure giving total damage possible from a set mana pool. For 6000 mana, how much damage could you throw on a boss, neglecting regen and mana potions and such like? Talk of spell efficiency is fine - but I fear in all the purity of this theorycraft that I for one got lost ;) --Clawsofdoom 13:42, 30 Aug 2006 (BST)

Clawsofdoom, the figure you are talking about is plain DPM. And it is currently present in the table. However, you can't neglect regen when comparing DPMs. In the example above compare spell#0 and spell#1 and you will see why.

To do 10000 dmg with spell #0 you will need 500 mana and 400 seconds of casting.

To do 10000 dmg with spell #1 you will need 1000 mana and only 200 seconds of casting. That means that you have 200 extra seconds to stand still and regenerate mana. And you can easily regenerate 500 mana in 200 seconds. :) --Rotsor 03:45, 5 September 2006 (EDT)

In my own numbers I have been using SQRT(DPS * DPM). The argument about this not disregarding insignificant changes is one of usage, not of theory. All of the spells in your example are equally powerful given different conditions. It is only true that you would prefer spells 0 and 2 under 'usual' circumstances, but given extreme circumstances spells 1 and 3 may be more useful. The 'efficiency' value given is perfectly fair in and of itself, and for most spells in the article it is a useful indicator.

Clarification
What is a 'sustained DPS spell'? I would presume it is used to mean spells with no significant cooldown, or reagent cost, or some other limitting factor to their use, but spells on the list don't agree with that theory. --Scudmarx 07:05, 6th November 2006

Incorrect Info...
I suspect the majority of this info is incorrect with the release of TBC. I know that many spell coefficients have been changed. Not to mention the addition of new spells. xD --Sky 01:05, 18 January 2007 (EST)

Priest Coefficients update: 3-20-2007
Here is a list of coefficients for priest spells: (found at http://forums.worldofwarcraft.com/thread.html?topicId=12881738&sid=1 )

42.9% Binding Heal (per target) 14.3% Circle of Healing (per target) 42.9% Flash Heal 85.7% Greater Heal 28.6% Prayer of Healing (per target) 42.9% Prayer of Mending (per charge) 20.0% PW:Shield 100% Renew

75.0% Holy Fire direct damage 25.0% Holy Fire dot 0.00% Mana Burn 42.9% Mind Blast 57.2% Mind Flay 110% Pain 65.0% Shadowfiend 42.9% Shadow Word: Death 71.4% Smite 100% Vampiric Touch

42.9% Desperate Prayer 100%Devouring Plague 0.00% Feedback 100% Shadowguard 171.4% Starshards 10.0% Touch of Weakness - Benedikta

Merge?
I disagree with the merge suggestion. This page is like Healing Comparison; it's not a mechanics page, it's a survey of spells. Of course, it really needs updated to be useful either way. Quamarett 03:35, 21 December 2007 (UTC)

Starting Update, Need HELP!
Base on Healing Comparison, I starting update the content. Ottokang 11:21, 30 December 2007 (UTC)

What's with completely taking down all the information that used to be there, rather than marking it as old and updating it piecemeal?

There's no place to get the full comparison for all classes now besides the history. The old information marked as out of date is much more useful than no information at all. Many of the changes are minor.

mes 20:11, 10 January 2008 (UTC)

Adding More information often
I am currently adding alot of information to this page so expect more info very soon. If anyone wants to double check my info which is the warlock table currently that would help.

Also I am creating the the SoC with three sections, if somone could give me feedback if there is a better way that would be great.

-Bagofdoom 21:49, 13 January 2008 (UTC) 10 Jan 2007

Scaling on spells
I am not really well informed about the current scaling percentages for spells, and since there are quite a lot of different numbers (for example the Pyroblast, some say it scales 100% others 115%) i dont feel the right person to add that here.

Could somebody else plz add that to the spells?

Doenerman 13:12, 12 January 2008 (UTC)Doenerman

Alternate Table Format
It seems like useful DPS numbers are: * DPSC - Damage per second of casting. * SDPS - Sustained DPS (Damage/Max(Cast time, Duration, Cooldown)) * DPS - Damage / Max(Cast time, Duration)

For each of those, as well as Damage and DPM, we could include scaling numbers (how much the value changes per point of +spell).

Example:

Druid
With this table I can do the math in my head to see, for instance, what the DPM for Wrath is at 800+spell

1.59 + .22% of 800 is about 3.35 DPM.

Erdluf 13:52, 14 February 2008 (UTC)