Lets all get in the way back machine and go back to a basic physical science class we probably took in highschool our freshman year. In that class you probably took mass readings on cubes of metal, compared between each other, did some simple standard deviation math, took a nap, made it through the semester.

A concept that hopefully was at least mentioned in that class was the idea of what I call the "Little Green Men" theory, which explains how you can prove or disprove things in a scientific manner.

You cant prove something doesn't exist by testing, you can only prove that you haven't found it yet. You can't definitively say that there are no little green men out there in space with ray guns coming to suck our brains out, but you can say that we haven't found them yet and thus it is not likely that they exist.

However, if there is definitive evidence that little green men do exist and they do in fact have rayguns and are in fact coming to suck our brains out, video footage interviews with the aliens and people laying on the side of the road with their brains sucked out, then you can prove that they do exist, regardless of how many people look into a telescope and see no flying saucers.


What this means, in less amusing terms, is that any assumption we make about the way game mechanics work, needs to be proven in some way. Once you have proven the mechanic, you need to document the proof, without documentation the proof is absolutely useless. To quote a great nerd Clifford Stoll (author of the Cuckoos Egg), "If you didn't write it down, it didn't happen".

What this *DOESN'T* mean, is that people need to prove you wrong when you have an assumption. You need to first prove yourself to be correct, before they can disprove your conjecture. Until you have proven yourself to be correct, the little green men may exist, but you have nothing to indicate they do other than "Hey, I said they did, isn't that good enough?", which isn't proof, since I can with exactly the same amount of credibility (if not more) say "Hey, I said they didn't! isn't *that* good enough?".


What I have set out to do is prove the Tooltip-TheoryCrafted mechanics we are familiar with in the World of Warcraft Mage, and document them in a way that can be referenced when they are called into question. Each experiment is done under controlled conditions that are recorded with the trial itself. This thread is basically intended to be my signature so other people can reference what I have done if they like, or I can just click through my own sig and see it all together in one place.



Interesting Links:

TheoryCrafting Tools
TheoryCraft-o-Matic: http://zaldinar.bounceme.net/tcom/
Rawr.Mage: http://elitistjerks.com/f75/t37825-rawr_mage/
MageGraf: http://www.magegraf.com

Important WoW Forum Threads
Mage FAQ: http://forums.worldofwarcraft.com/thread.html?topicId=75311165
FrostFire Bolt: http://forums.worldofwarcraft.com/thread.html?topicId=12661303549
Spell Hit: http://forums.worldofwarcraft.com/thread.html?topicId=7475660772
TCoM Thread: http://forums.worldofwarcraft.com/thread.html?topicId=11693100528
Treatsie on Scaling: http://forums.worldofwarcraft.com/thread.html?topicId=2038122089
How Stats Work: http://forums.worldofwarcraft.com/thread.html?topicId=7716836049
World of Mechanicscraft: http://forums.worldofwarcraft.com/thread.html?topicId=108235377
When does frost out DPS fire: http://forums.worldofwarcraft.com/thread.html?topicId=3547591903
Is the fire only elitism over: http://forums.worldofwarcraft.com/thread.html?topicId=4184984246
Triple Point Graph: http://forums.worldofwarcraft.com/thread.html?topicId=93003149

Important EJ Threads
FrostFire Bolt: http://elitistjerks.com/f75/t37364-frostfire_bolt_thread/
Frost Raiding: http://elitistjerks.com/f75/t37133-raiding_frost/
Pre-Raid Gear: http://elitistjerks.com/f75/t37146-best_non_raid_mage_gear/
Raid Gear: http://elitistjerks.com/f75/t42476-rawr_optimal_mage_sets/

-{ Arcane Tree }-

Arcane Barrage
( http://forums.worldofwarcraft.com/thread.html?topicId=10043085137&pageNo=3#56 )
- Coefficient: (2.5 / 3.5) = 0.7142


Missile Barrage
( http://forums.worldofwarcraft.com/thread.html?topicId=10043085137&pageNo=2#22 )
- Reduces base cast time to 2.5 seconds and period to 0.5 seconds, normal haste rules apply after wards

Netherwind Presence
( http://forums.worldofwarcraft.com/thread.html?topicId=10043085137&pageNo=2#22 )
- Stacks multiplicatively with Haste as expected


-{ Fire Tree }-

Ignite
( http://forums.worldofwarcraft.com/thread.html?topicId=106669217 )
- Provides 40% extra damage in a 'bank' system that preserves previous crits
- Two bugs exist, one that increases damage, one that reduces damage

Frostfirebolt
( http://forums.worldofwarcraft.com/thread.html?topicId=10043085137#5 )
- Coefficient: (3.0 / 3.5) = .8571428
( http://forums.worldofwarcraft.com/thread.html?topicId=10043085137&pageNo=2#36 )
- Coefficient with Empowered Fire: (3.0 / 3.5) + .15 = 1.0071428

Pyroblast
( http://forums.worldofwarcraft.com/thread.html?topicId=10043085137#5 )
- Coefficient: 1.15 Frontend

Burnout
( http://forums.worldofwarcraft.com/thread.html?topicId=10043085137#6 )
- Crit Gain:
5/5 Burnout: 1.75 = 175% crits
5/5 Burnout 5/5 Ice Shards: 2.25 = 225% crits
5/5 Burnout 5/5 Ice Shards 5/5 Ignite: 3.15 = 315% crits
( http://forums.worldofwarcraft.com/thread.html?topicId=10043085137&pageNo=3#54 )
- Increases crit damage of all spells except Blizzard

Blastwave
( http://forums.worldofwarcraft.com/thread.html?topicId=10043085137#20 )
- Kickback range: About 10 yards, *APPEARS TO* (as in, not proven, but indicated) increase as you move farther away from the caster.

Living Bomb
( http://forums.worldofwarcraft.com/thread.html?topicId=10043085137&pageNo=2#37 )
- DoT Coefficient: 0.2 per tick
( http://forums.worldofwarcraft.com/thread.html?topicId=11693100528&pageNo=7#134 )
- Kaboom Coefficient: 0.4
( http://forums.worldofwarcraft.com/thread.html?topicId=10043085137&pageNo=3#57 )
- Glyphed crits do not benefit from burnout, interact with combustion, or generate ignites. Do gain from CSD

Assorted DoTs
( http://forums.worldofwarcraft.com/thread.html?topicId=10043085137&sid=1&pageNo=3#43 )
( http://forums.worldofwarcraft.com/thread.html?topicId=10043085137&sid=1&pageNo=3#58 )
- Fireball, No Spellpower, Yes Playing with Fire / Firepower / Arcane Instability
- Pyroblast, Yes Spellpower (0.05 per tick), Yes Playing with Fire / Firepower / Arcane Instability
- Flamestrike, Yes Spellpower (0.122 per tick), Yes Playing with Fire / Firepower / Arcane Instability
- Living Bomb, Yes Spellpower (0.2 per tick), Yes Playing with Fire, Fire Power
- Frostfire Bolt, No Spellpower, Yes Playing with Fire / Firepower / Arcane Instability / Piercing Ice


-{ Frost Tree }-

Elemental Precision
( http://forums.worldofwarcraft.com/thread.html?topicId=10043085137&pageNo=2#35 )
- Gives 3% hit bonus to frostbolt.

Ice Lance
( http://forums.worldofwarcraft.com/thread.html?topicId=10043085137&pageNo=3#57 )
- Glyphed crits for 4X damage on level 83 boss targets, 3X on same or lower level targets


-{ Misc }-

AoE Spell Caps
( http://forums.worldofwarcraft.com/thread.html?topicId=10043085137#18 )
- Frostnova, Blastwave, Dragon's Breath, Cone of Cold, Arcane Explosion, Living Bomb = 37,500
- Flamestrike = 75,000
- Blizzard = 25,000 per wave
( http://forums.worldofwarcraft.com/thread.html?topicId=10043085137&pageNo=2#22 )
- Hellfire = 25,000 per tick
- Rain of Fire = 50,000 per tick
- Seed of Corruption = 50,000 per kaboom
- Shadowflame = 6500 per cast (NOTE, this appears to likely be an OVERSIGHT)

- Summary: General rule appears to be 25000 DPS limit by cast time or damage frequency.

AoE Spell Coefficients
( http://forums.worldofwarcraft.com/thread.html?topicId=10043085137#19 )
- Arcane Explosion, Cone of Cold = (1.5 / 3.5) / 2 = 0.2143
- Blast Wave + Dragon's Breath + Frost Nova = ((1.5 / 3.5) * 0.9) / 2 = 0.1929
- Flamestrike DD = 0.2427
- Flamestrike DoT (per tick) = 0.1220

Pre Casting
( http://forums.worldofwarcraft.com/thread.html?topicId=2968113556 )
- '/stopcasting' macros are not required post 2.3 for optimal DPS
- Quartz style addons still provide the best possible DPS time

One Roll Two Roll
( http://forums.worldofwarcraft.com/thread.html?topicId=443186981 )
- Spells use a two roll system, first checking if they hit, then if they crit

Multiplicative Bonuses
( http://forums.worldofwarcraft.com/thread.html?topicId=108301152 )
- Typically are all multiplied together to form a composite multiplicative
( http://forums.worldofwarcraft.com/thread.html?topicId=10043085137&pageNo=3#40 )
- Except for Spell Impact, which is additive after Firepower
( CITATION NEEDED )
- And the FFB Glyph, which is additive after Firepower as well

-{ Known Bugs }-

Ignite

( http://forums.worldofwarcraft.com/thread.html?topicId=106669217 )
- If two fire spells crit at effectively the same time sometimes only one generates an ignite debuff, this results in a net loss of DPS.
- If a fire spell crits at effectively the same time as an ignite tick happens from a previous ignite, the tick is not counted as used and is recycled into the ignite bank, this results in a net gain of DPS.
- In typical conditions you will encounter more gained bugs than munches.

Spell Impact/FFB Glyph
( http://forums.worldofwarcraft.com/thread.html?topicId=10043085137&pageNo=3#40 )
- Are applied additively instead of multiplicatively ( 1.1 * 1.06 = 1.166 vs 1.1 + 1.06 = 1.16 )

Firepower
( http://forums.worldofwarcraft.com/thread.html?topicId=10043085137&sid=1&pageNo=3#43 )
- Does not increase the damage of the Living Bomb DoT ***FIXED in 3.2ish***

Flamestrike
( http://www.wowhead.com/?search=flamestrike )
- Ranks 1-6 have a 2 second cast time, ranks 7-9 have a 3 second cast time

Shatter Combo
( Reported for due diligance, blizzard has stated that they are 'aware' of the combos and keeping track of them )
- It is possible to launch a Frostbolt->Icelance combo on a target and have both gain the frozen benefit even if the Frostbolt will break the Nova effect

Combustion
( Also reported for due diligance, no active blizzard acknowledgment has ocurred )
- it is possible to launch a Cast Time spell (Fireball/Scorch/Flamestrike) and follow up with an instacast (Fireblast, Blastwave, Dragon's Breath) and get the current crit benefit of Combustion for both spells even if one should consume all the charges of the combustion

Placed a sticky request. Get to work people.

And yet theorycraft is all but useless because you can't calculate the human factor. Hence why I tend to break theorcrafted DPS, and my gear isn't even that good.

Not to doubt or hate, but how does this "prove" anything really? You could take a sample of 10,000 frostbolts and do the math to get a number, but that is not to say that this number would not change the outcome given 1,000,000 casts.

To relate to your statement, you have not proven that there are little green men, simply given evidence that they exist (like finding a laser gun with a green fingerprint on it!!!).



Regardless of that, I guess I would be interested in knowing if you DISPROVE anything that has been a "known-assumption" used in calculations. If everyone assumes something is true, you are simply adding a sample to further that knowledge. Now if you find something that ISN"T true that everyone has assumed is, that could be very nice.

Aside from that, I have taken everything you showed to be common knowledge....assumptions that I have taken as truth (and will continue to until they are proven otherwise).


Just curious, what was your sample size on these numbers? and did any of your conclusions not agree with any of the previously accepted assumptions?

---

Q u o t e:
Tagged cause people will post in here about serious stuff and I won't.

---

You're stoopid.

---

Q u o t e:
And yet theorycraft is all but useless because you can't calculate the human factor. Hence why I tend to break theorcrafted DPS, and my gear isn't even that good.

---

I am guessing that theorycraft could pretty easily factor in your maximum possible DPS on any particular fight to the point that you would be unable to beat that number.

Keep in mind that numbers used on much of the theorycraft are below raiding mage level stats. things like 800 +dmg...what real raiding mage has less than 800 dmg? But that keep it so that others can relate to it even if they are not PvE mages and such.


Another thing to think about is the ratio. If you and another mage are equally geared and skilled...the ratio of how much damage one may do compared to the other should stay the same as the theorycrafting numbers.

Example:
theorycrafted
Person 1 (deep frost) - 1000 dps
Person 2 (deep fire) - 1050 dps

actual
Person 1 - 2000 dps
Person 2 - 2100 dps.

the ratio stays the same, just the numbers might differ.

---

Q u o t e:
Not to doubt or hate, but how does this "prove" anything really? You could take a sample of 10,000 frostbolts and do the math to get a number, but that is not to say that this number would not change the outcome given 1,000,000 casts.

---

You're absolutely right that the average solved values will differ over those long periods, but what is useful is using the minimum and maximum recorded values in a given sampling. When you reach enough datapoints (for example, 1845 on the ice lance coefficient test) you can safely say you have seen the entire range of the spells output, then you solve for the min and max, and if the numbers come out even, you know you have an established good range.

---

Q u o t e:
Regardless of that, I guess I would be interested in knowing if you DISPROVE anything that has been a "known-assumption" used in calculations. If everyone assumes something is true, you are simply adding a sample to further that knowledge. Now if you find something that ISN"T true that everyone has assumed is, that could be very nice.

Aside from that, I have taken everything you showed to be common knowledge....assumptions that I have taken as truth (and will continue to until they are proven otherwise).


Just curious, what was your sample size on these numbers? and did any of your conclusions not agree with any of the previously accepted assumptions?

---

Check the threads for sample sizes, they contain all the pertinent info for the results posted in this thread.

I disproved the established Pyroblast coefficient theory, I disproved back in 2.0.6 the way that we thought the Improved Coefficient nerf worked.

---

Q u o t e:
And yet theorycraft is all but useless because you can't calculate the human factor. Hence why I tend to break theorcrafted DPS, and my gear isn't even that good.

---

While true that you cant calculate the human factor, if you gave me accurate information about your spell rotation and gear / spec / buffs / other party members adding debuffs to targets, I could generate a maximum output line that you couldn't break without breaking the conditions of the model.

The "breaking of the theorycrafted DPS" you are observing is because most theorycrafted models rely on the average damage of a given spell, where over any given sampling you may exceed that, although as your sampling grows the tendency towards being on average grows aswell.

This is serious business.

---

Q u o t e:
This is serious business.

---

lol

Since you seem so interested in theory and falsifiability, never forget what Einstein said: "No amount of experimentation can ever prove me right; a single experiment can prove me wrong."

---

Q u o t e:
Since you seem so interested in theory and falsifiability, never forget what Einstein said: "No amount of experimentation can ever prove me right; a single experiment can prove me wrong."

---

And you're more than welcome to do so, I love being proven wrong or having someone find a kink in my models. What I have done with these bits of research is established my theories with data supporting them, if you can generate data that goes against what I have found, then we can get some peer reviewing going on and do a true scientific process to decide what the story is.

This isn't an end all "This is absolutely the way it is", this is "This is how it appears to be, and here is data to back it up". Mechanics change and aren't listed in patch notes, Mechanics have hidden interactions that I may not have encountered in my testing (IE Wrans Master of Timing ignite bug that I hadn't observed on my own until he pointed it out to me).

---

Q u o t e:


And you're more than welcome to do so, I love being proven wrong or having someone find a kink in my models. What I have done with these bits of research is established my theories with data supporting them, if you can generate data that goes against what I have found, then we can get some peer reviewing going on and do a true scientific process to decide what the story is.

This isn't an end all "This is absolutely the way it is", this is "This is how it appears to be, and here is data to back it up". Mechanics change and aren't listed in patch notes, Mechanics have hidden interactions that I may not have encountered in my testing (IE Wrans Master of Timing ignite bug that I hadn't observed on my own until he pointed it out to me).

---

Okay, but then you have to stop claiming that you've proved anything. ;)