I'll follow the pattern of using spreadsheets to try to model DPS classes. This is continuation of what I started here:
http://forums.elitistjerks.com/viewtopic.php?id=6819
I put it on its own thread, since discussion here will probably focus mainly on the spreadsheet itself, and there's no reason for the OP to be clogged up with line-by-line calculations. I didn't want to edit it out however, since it might help explain what I'm doing.

Here's what I have so far:
http://www.sigilguild.net/hamlet/Mage.xls
It's not all that versatile yet--the stats are only for Human, and I haven't looked at Fire spells at all. But I wanted to get some input early on.

It's not all the user-friendly yet either. Basically, you can enter a gear/talent/buff setup in the top left box. The top middle box computes a bunch of relevant quantities about your damage/mana use with certain spells (only Frostbolt and Arcane Missiles for now).

The bottom middle box computes your average DPS and mana usage with a few different spellcasting patterns, in the manner described in the thread linked above. The bottom right box checks what your best option is for DPS at zero net mana usage.

Finally, the top right box shows your theoretical damage output in various length fights (you can customize the fight lengths, and how many Evocations and Mana Gems you want to allow yourself). In each case, it checks two options:
1) Chaincast FB11 until OoM, then use your best constant-mana option for the rest of the fight,
2) Same, but with Arcane Missiles worked in on all the clearcasts,
and reports the better total.

At the moment, you can enter two parallel setups; it runs the first on page 1 and the second on page 2, and compares the two in the bolded box in page 1.

-------

There are some things I want some thoughts from other Mages about.
It's harder to accurately predict our DPS in practice than it is for, say, a Rogue, due to the mechanics of casting. For them, white damage works automatically, and exact timing of the specials is not critical because the rate at which they can use them is limited by energy regen anyway, so theoretical DPS should be very close to actual DPS.

Right now, I try to account for a few specific things:
The unavoidable delay in changing spells when you Clearcast. Right now, you can enter how much time you think you lose doing that.
Full-regen ticks that sneak in after Clearcasts. Again, right now, you can enter how many you think you get.
Overall casting lag. For the moment, this is simply modeled as a uniform increase to cast times.

Any other ways to try to improve a model of our actual DPS?

Finally, things I haven't worked in yet:
NW Focus. I don't have this yet; any thoughts on the particular timing intricacies would be appreciated.
Timers (AP, PoM, Trinkets)
Fire spells in general. I've never played Fire in raids, can anyone give a basic description of how you tend to use Fireball/Scorch in practice?

-------

A few preliminary results:
Low-rank Frostbolt combined with Clearcasted Arcane Missiles seems to be the best way to regen back up while maintaining overall DPS. In other situations, though, the value of working in Arcane Missile on Clearcasts is little to nothing. It seems the overall best plan is to chaincast Frostbolt R11 until dry, and then use the cycle mentioned above.

In general, Arcane Missiles is highly marginalized. Since such a powerful regen is available by using low-rank Frostbolt, the differential benefits of Int, Spi, and +dmg(non-Frost) are small. +dmg(Frost), +hit, and +crit are strongly dominant, even in long fights.

well, my mage is only level 40, but i'm already glad i specced arcane upto improved counterspell. 25 mana for a rank 4 arcane missile to fire off has saved my ass too many times to count as well as allowed for reasonable mana regen.

my problem is indicisiveness for 1.11. I love a lot fo the arcane talents but i'm trying to play without PoM so i get used to having to escape situations whilst casting. I know most mages at the moment swear by it but i feel what with 1.11 offering several substantial raid bonuses for a fire/frost spec i'm probably doing myself a favour.

i'll admit spreadsheets make me go cross eyeed (i'm a literature man ot one of science), but its damn nice Hamlet (what numbers i've typed into the talent setups :))

Just to bump something - when I parse myself on level 63 MC mobs, my frostbolt only has roughly a 5% resist rate, on all of them. All have Curse of Elements on them. I'm only wearing +1% to hit at the very most on these fights.

Does Curse of Elements have an effect on Frostbolts innate resist rate? I know it sounds strange, but I just can't reconcile a 17% resist rate with what I see in MC, especially when I last parsed myself on it. (Admittedly a while back).

This is interesting. A player by the name of Xyruul from Warsong (US) did some tests the other day by casting 1000 rank 1 frostbolts against a target. You can see the thread (might not last too long) here.

Essentially he found that, at least in the case of frostbolt and against other players, the cap from +hit gear (minimum 1% 'miss' rate) was calculated after resistances were taken into account. Essentially for frostbolt, +hit gear will act like spell penetration gear, at least against players.

That started me thinking that it was possible the reverse would work as well. Namely, would spell penetration and curses work to lower the frequency of 'level based' misses, at least in the case of a binary spell like frostbolt?

Your post seems to suggest that.

It's still an open question as to what the case is with fireball and the other 'normal' damage spells. I'll probably start investigating this myself and post my own findings.

If thats the case, then + to hit gear for frost mages is utterly overrated, as the cap will be at around 3 or 4% tops before you get the 99% hit rate, if CoE does work that way.

Bah, am running BWL tonight, not the best environment to parse as CoE isn't on the mobs a lot of the times, and other stuff is always going on, but I'll try to see what I can find measure.

If you are getting a roughly 4-5% resist rate with CoE up, and -resist works in the way you're suspicious of, and the bosses are in fact level 63 with a natural miss rate of 17%...if your miss rate is between 4% and 5%, this would tend to suggest that the MC bosses have roughly 98-102 (lets call it 100) frost resistance, assuming the:

.15(resist score)/(caster level) = mitigation %

is true.

If your wore roughly 30 points of spell penetration, it would be interesting to see what resist rates you experienced. (I'd do it, but the warlocks in my guild seem to have difficulties finding their CoE buttons. Curse of agony has a double meaning for me.) Of course it's entirely possible idiosyncracies of Curse of Elements, and not the general idea of -resist/spell penetration, is what is causing these results.

Hah. Now that I think about it, that's the one I could test for ;)

Just as a clarification: You want us to enter the bonuses from Gear in the top left box, correct? Not total stats.

Dont you have a mage Wodin? :P

This will be interesting to see the results of tho.

Not all mobs in MC are level 63, though. Also, a bunch of those mobs have negative Frost resist, like Firelords, Fire Spawns and Ragnaros.

It used to, but after 1.9, most frost mages reported seeing their resist rates go up in MC (particularly vs. Molten Destroyers).

Many, many moons ago, Eyonix made this post. It suggests that the formula for resisting a binary spell (outside of Arcane Focus or Suppression) is as follows:

chance to hit = (x + to hit)*[1-(target resist)/(5*caster level)*0.75]

where x=0.96 for targets at or below your level, 0.95 for targets 1 level higher, 0.94 for targets 2 levels higher, etc. This check is a single check, not a double check (i.e., the system doesn't roll to hit and then to resist for binary spells). From everyone's experience with old CoE and firelords/Ragnaros, it's commonly accepted that the value of (target resist) can be negative and it does indeed decrease the resist rate of binary spells versus the level cap.

All tests suggest that outside of Suppression, Arcane Focus, etc., chance to hit is capped at 99%. Inside of Suppression, Shadow Focus, Arcane Focus, etc., it's well known that no resist cap applies and it's possible to get a 100% hit rate on spells.

Yes, but he's been very explicitly not a raid character for a reason. :) I know the basic rules and the interesting pvp variations on how spells work, but PvE I admittedly only know the basic damage mechanics that have been laid out(ie how hit/crit/resist/+dmg work). The exotic things like the ordering between level-based resists and stat-based resists I'm still unsure of.

Neat, we need to start a WoW spreadsheet web ring or something!

Just another question: Why the two spirit boxes? Perhaps I don't quite understand the mechanics of Evocation, but if the original spirit box is +Stats anyway, where's the need for both boxes?

My Spirit is 50 higher than usual during Evocation, because I swap in Dancing Sliver (w/ Mighty Spirit) and Dragon Finger of Spirit. This is easy to do and makes a large difference to Evocation (marginally, 1 Spi adds 18.4 or 19.6 to Evocation, depending on how well "primed" it is with the 5SR).

Speaking of weapon-swapping, I'm pretty sure that people like me who use AV offhands can switch them around without losing any time, by "hiding" the global cooldown inside spell cast times. As far as I can tell, though, this isn't really significant enough to be worth doing.

EDIT: D'oh, I just realized my profile isn't in my sig. I'll put it there now, for reference. The 2 setups in the chart already are my current and "wishlist" profiles (note that the huge increase in DPS is mostly due to the new talents, and less to gear upgrades).

I used to be able to do this, but I can't seem to do this now. I can swap weapons while channeling Evocation, but when I'm, say, casting Frostbolt, I don't seem to be able to change weapons. And believe me, I've tried... it'd great to swap to the Sliver+Dragon Finger of Spirit as you cast one last Frostbolt, so you can get 2-3 Evocation ticks in FSR.

Hmm, to be honest, I've never actually tested the weapon-swapping and the 5SR in conjunction, only separately, so maybe you can't get quite as much as I'd thought.

I use ItemRack to handle the weapon swaps. I'll watch it carefully next time I chain a non-instant cast spell directly into Evocation, but you're probably right; you miss the first tick.

NW Focus procs as soon as the spell casts (not when it lands, like clearcast does). As far as timing goes, everyone who gets 8/8 NW for the first time has to go thru an adjustment phase. Every mage attempts to time their next spell the moment their current spell finishes casting, but this is the exact same time NW proc goes off, and many mages find they have fired 2 frostbolts off simultaneously, but hardly realize it (takes some getting used to). In pure theorycraft, this shaves 1 second off the 2.5 second talented frostbolt cast time, but does nothing else (ie, the global cooldown becomes the cast time of a frostbolt when NW focus occurs). When chain casting, one just has to accept that the regular "hit button every 2.5/3" seconds will change once in a while based on when NW Focus procs.

Ok. So what do I need to know about the average behavior of someone who's casting away, all while dealing with the simultaneous chances of NW Focus and CC?

0.1*0.1 = 0.01, so 1% chance they both activate at the same time. But a 20% chance on every spell that it is either free or instant cast (presuming you're using Frostbolt, Fireball, or Arcane Missiles).

The new tier 1 frost talent is going to appear in nearly every mage build though, an innate 6% chance to hit with frost and fire spells.

Ah crap, I just realized that Clearcasting can only proc when you hit. This throws a wrench in the whole works.

EDIT: I don't even want to think about partially-resistable Fire spells yet.

Ugh, and Ignite (and Combustion + Ignite) are going to be messy. Will probably have to take the simulational approach. As soon as I've been away for long enough to stand the sight of MATLAB again, I'll give it a go.

Ok, got my 8/8 today. Will continue to work on this as soon as I have time--it now has more interest for me, as I'd to find out what it predicts as to how the set bonus will compare to +127 more spell damage from non-set gear. And, as always, whether to go Frost or Fire after the patch.

For reference, any intuition from current Mages on wich would push more DPS: the gear in my profile, or the same suit with full NW swapped into the appropriate slots?

None of the mages in our raid is particularly enamored of the bonus - mostly because of things like hit and crit that they lose out on.

The rough theorycraft I came up with (spamming all r11 frostbolts) came out to about a 4% dps increase. If you swap in fireballs on procs instead that should go up a bit more. Being generous to the set maybee a 6% increase. Remembering rough equivalency from your spreadsheet 3 +dmg ended up as about 1DPS, soooo you're looking at roughly 400 DPS*.06 vs 127*.33 which works out too - 24 DPS vs 41 DPS.

Even if the clearcast procs turned into a raw 10% DPS increase, you're pretty much at the level of +dmg (assuming you have the 648 that comes base in your spreadsheet) that makes netherwind less valuable than straight up non-set gear. (especially if you factor in lost hit and crit).


Hopefully that's useful as a baseline for you to work from.