Super and thank you very much. I guess I am wondering whether clipping the fifth AM tick will be de rigeur in both your TC as well as recommended real-world practice...?

Definitely agree about miscalculation and inconsistent mechanics...even if you're fairly certain you know all of them it can be very difficult to account for them without doing some programming or heavy statistics.

Also, standard deviations with fire being a problem, you'd want to compare that to arcane as well ....if I had time and actually kept using matlab (...I switched careers), I'd just put in the mechanics and have the thing simulate a couple hundred fights and do some statistical work.

Thanks for your work on the thread (and the whole mage forum in general)!

Well my plan has always been to at least have the base calculations be crystal clear. What I mean by that is roughly what the ffb thread spreadsheet does. Its simplistic, but the calculus is accurate, and most importantly (to me), it is very easy to follow. The goal is somewhat to open up mage TC and allow mages to understand why a spell scales well or whatnot.

I intentionally do not model hot streak, or combustion, or any cooldown basically (lust, IV) mostly because I want the base calculus to be something immediate and simple to figure out. I mean, if I do a spreadsheet that supports hot streak, combustion and all cooldowns, at that point, why use said spreadsheet over rawr or any other tool ? I redid every calculation from scratch. I rechecked multiple times the calculations to be sure there were no mistakes, even though they are kept to a simple level. I am fairly confident in its exactitude. Moreover, its always nice to have some individual confirmation that other tools appropriately model the numbers -- it should somewhat match what I have.

Quite unfortunately, arcane is far more annoying to model using a simplistic spreadsheet like I did. I want the calculus to be something you can easily understand from looking at 2, 3 or 4 cells in excel and you have it all in front of you. Now the annoying part is that arcane needs to take DPM into consideration as part of its dps cycles. As a result, while it would be nice to know the DPS/DPM of the different spells, I know my spreadsheet won't go past that line and tell you exactly what you want to know -- which cycle is optimum for the given fight duration ? how long will my mana last ? how much mana do I consume during lust ? But at least, ignoring cooldowns (AP, lust, IV, POM) I think I can establish a baseline of data good enough to keep it somewhat simple.

Same problem here with those factors, I have my own sheet similar to yours for FFB and fireball, but modeling arcane is an entirely different subject.

I look forward to checking out your sheet and thanks again!

That's an interesting question. It comes down to what kinds of questions you want to answer. In my case I'm trying to predict whether a talent change or equipment change will help or hurt various real-life scenarios. This is not the same question Magegraf tries to answer except with very broad strokes, and it overlaps with what RAWR does, but the fit isn't perfect. I've got mana on the spreadsheet but it is mostly there as a benchmark. If one talent rotation costs 2x the mana of another, I know I'm going to burn out faster, but how much faster, I need quality time with a target dummy to find out for sure whether it's likely to be "long enough" or "not enough" and thus relevant.

Also the spreadsheet is portable in a way RAWR (which requires an install) is not, and Magegraf, while convenient, doesn't use the same rotations I do in actual play so it's good for general purpose "how much better is my equipment than last week" and a sanity check to my spreadsheet but not good enough for the finer points of measuring gear against other gear. They also don't let me model AOE, burst and similar real world considerations very well, where my spreadsheet lets me see the tradeoff when I, say, go to arcane/fire and see an increase in single target dps but see my aoe and burst aoe dps drop and my mana shoot up.

On the other hand, my spreadsheet only works for me, as it's partly based on empirical results from WWS on how I actually cast spells, as opposed to theoretical rotations. What Manly did is a lot more usable by others. Which is why his is worth posting and mine is not.

I modeled combustion as a higher crit rate (or more accurately, 2 extra crits within its cooldown, when compared to the expected number without it), I modeled icy veins as haste divided by uptime, similar to what I did with clicky-trinkets. Hot streak uses the crit rate to get a rate of hotstreaks that I use in my final rotation formula to know how many pyroblasts to include and to estimate pyroblast DOT uptime. Living bomb I assume it's ticking 12 seconds of every 15, which is slightly conservative compared to real life.

What I don't assume in my model is stacking cooldowns, rather I just assume they come up automatically in the rotation (decent because they're macroed, and don't have same cooldowns so they're rarely totally stacked after I fire off the first one) For me, I get more uptime if I do it that way than if I try to second guess the cooldowns and save some for when they might stack. Dunno if I'm costing myself significant DPS or gaining it, but my performance in the field comes close enough to what I calculate that it's good enough.

I'm not a statistician, and either way it's not relevant. For any given FB spam the percentage of damage that represents bonus damage from a Critical will be significantly less than for any similar FFB. It's not exactly rocket science to work out that one spell that crits for 254% of base is less penalized (and consequently less buffed) from non-stable RBG than one whose spell crits for 344% base.

Plus, as the absolute crit rate will be higher for a FB spec, relative fluctuation of RNG will affect it less. In layman's terms, when your crit rate is 45% adding or losing 10% is a smaller change than when it's 40%, add or lose 10%.

Someone correct me if I'm wrong here but RAWR doesn't require an install - at least I never have.
Just unzip the files somewhere (Anywhere. Mine are to a folder on my desktop) and run the rawr.exe executable.

Correct. No install is required, beyond .NET, which is already on any recently updated copy of XP, or any copy of Vista. You could run it directly from a flash drive if you wanted.

No, no, and no. Recently there have been a few points of discussion. There's two questions, essentially.

First: To ignore, or not, MB procs.

Second: How to correctly "arcane-shatter".

Ignoring MB procs basically reduces the game's rotation to AB*3/AM/ABarr where the ABarr is shot before the 5th AM Missile hits, making it benefit from the 3*AB buff.

This rotation ignores whether or not MB has procced, which it most likely will have, after 4 spells which could proc it. (chance MB will not proc after 4 spells: 0.8^4=41%), is high in mana cost and seems to offer the best DPS in the arcane repertoire.

Correctly Arcane-Shattering is not really much of an issue to discuss, as it is an issue to practice. The key is successfully getting the ABarr to Launch before the 5th AM pulse Hits. If you do not succeed in this, the ABarr will be an unbuffed cast, resulting in inferior DPS to a conventional cycle which does not shatter but does MBAM whenever it's up. The problem is, raid environment, ambient latency and graphical delay may mean clipping the 5th missile is more systematically reliable than releasing the ABarr button when the 5th pulse has left but not landed. This is even truer for short-range where travel-speed is limited.

To summarize what you asked:

There is no definite answer, if there was we'd have created the Mage: Arcane Thread already. It's not exactly clear which rotation is superior and why, and there is no set-in-stone "always do X before Y and N when M procs". It's not clear at all. Watch this space.

Hy!

Im back from hours of dummie testing (on PTR). First i made 15 Circles with 0/51/20. With glyph's, molten fury, arcane int, and all cooldowns (IV, Mirrorimage, Managem..) and both trinkets (spider + sundial) but without Evocation.

With FFB i got an average of 2700 DPS and 148 Sec. fight duration.

After that i changed to 51/3/17 (http://talent.mmo-champion.com/?mage...=031900000000#)

I tested with the same(!) glyphs because no glyphs available on the ptr . Also with Molten Fury, Int, Trinkets and all cooldowns (+ Arcane Power).

I reached an average of 3000 DPS, but was OOM after ~ 70 Sec.

Dont forget to calc additional focus magic, 5% more arcane blast dmg with glyph and glyph of arcane power!

So this specc should be tested on Patchwork on live Servers this week

(sry 4 spelling, fast written, no time at work :x)

This question also pertains to the initial CD usage as well, i.e. do you AB spam ignoring procs until you are at max_mana - mana_gem before popping CD's and gem and then starting the rotations, which was standard for the old 40/0/21 spec.

Good to know. I was and am understandably nervous about downloading anything to my work computer and installing it.

According to the latest Simcraft on Page 1, it does include Arcane Blast as a glyph and shows Arcane to be the top dps Mage spec. I cannot seem to find how long it takes the fight into account for, but assuming a normal 5-7 minute fight with Arcane doing half way decent on mana it should stay as the top DPS spec.

Like Arastonachis posted in his results, Arcane is clearly winning but mana regen is fairly taxing at the moment.

Top enough that your guild is willing to sacrifice 10% spell crit?

Another question: what's the status of gems/enchants? Are we gemming for some spirit and/or intellect like 40/0/21? (assuming you can't get on-demand innervates like Fireangel lol)

No, but not every mage has to go Arcane.

According to the Simcraft SP>Int>Haste>Crit for 57/3/11. So to answer your question, I would still gem for SP once hit capped and then probably look for Int/Haste next. Although with mana being as terrible as it is, it might just actually be worth it to gem for Int rather than the next best stat, haste. Luckily Arcane doesn't work like TBC where the debuff has to countdown to come off.

I think all of the discussions about an Arcane spec assume there is a mage in the raid who is putting up Imp. Scorch or Winter's Chill. Clearly 10% less crit for 5-8 people is more important than 200 more DPS for 1 person. Many 25 man raid I am in have more than 1 mage and I will gladly respec for the greater good if a situation comes up where we need the crit debuff.

You're right. Not every mage HAS to go Arcane. But maybe every mage WANTS to go Arcane.

That's the issue. In my guild there's only 2 mages. It's not unreasonable to think both want to go Arcane, is it?

So just go 51/20/0

I worry less about wants and more about needs. We run with 3 mages typically, I would tell at least 1 of them to go arcane and the other 2 as FFB (one to scorch if the other goes down). In your case, if neither of you rarely die I would just trade off FFB and Arcane between the two of you each week, otherwise you both might want to stay FFB as a 'just in case'.

EDIT: Or take Manly's advice

Thanks. However upon checking simcraft I noticed a few things:

It seems simcratf is using the glyph without it stacking which means that the arcane numbers are possibly higher than it is currently showing.

I also thought this patch was making 18/53/0 (FB Spec) as viable as 0/53/18 (FFB Spec) yet SimCraft is accounting only 20/51/0. Is there a reason for the 2 points in arcane instead of fire?

In regards to your second point/question, where would your 2 points go into 18/53/0 over 20/51/0? I could see a case for flame throwing but in reality the 2 points are much better spent on Totw/Regen.

Assuming Incineration is not included in the spec, Flame throwing is an argument (I personally like having extra range on my spells) but couldn't incineration also be one? I mean it is extra crit on scorch and since you'll be using it anyway might as well try and get more crits for more HS procs.

Maybe I am thinking all wrong here.