[Mage] TC after 2.3 - Page 70

Finkum · 12/06/07, 12:19 AM

To back up this argument with some (very simple) maths:- using the following formula

FBlast coeff x +dmg + Fblast base = (Fball coeff x +dmg + Fball base) / 2
0.43x + 725 = (1.15x + 719) / 2

and solving for x (your spell damage), it turns out you need 2520 spell damage for pure fireball spam's dps to equal the dps given by including fire blast in your rotation (assumes zero haste, does not account for incinerate).

With 4t6, this becomes

0.43x + 725 = 1.05 * (1.15x + 719) / 2

and solving for x gives 2006, still well above typical spell damage levels.

Augmenting the formula to allow for spell haste (no 4t6, y denotes spell haste rating):

0.43x + 725 = (1.15x + 719) / ((3 / (1 + (y / 1570))) / 1.5)

Substituting in my own raid-buffed +dmg, x = 1100, it turns out that I would need ~325 spell haste rating (~21% haste) before pure fireball spam wins out.

Scruffy1207 · 12/06/07, 12:59 AM

I need a little help with ultimate DPS rotation for a mage. I am a 41/20 arc fire spec. 1026 Fire DMG unbuffed, 28.21% chance to crit. any Ideas?
if this armory link works there ya go. it was broken when i copied the link. http://armory.worldofwarcraft.com/ch...maul&n=Scruffy

Tempestra · 12/06/07, 2:36 AM

Originally Posted by Scruffy1207

I need a little help with ultimate DPS rotation for a mage. I am a 41/20 arc fire spec. 1026 Fire DMG unbuffed, 28.21% chance to crit. any Ideas?
if this armory link works there ya go. it was broken when i copied the link. The World of Warcraft Armory

http://elitistjerks.com/f31/t18441-m...tional_thread/

There's a sweet informational thread for a reason. Sorry - I really try to avoid saying "this is a stupid question" but... this thread exists for this specific purpose.

Rouncer · 12/06/07, 9:41 AM

Originally Posted by Finkum

To back up this argument with some (very simple) maths:- using the following formula

FBlast coeff x +dmg + Fblast base = (Fball coeff x +dmg + Fball base) / 2
0.43x + 725 = (1.15x + 719) / 2

and solving for x (your spell damage), it turns out you need 2520 spell damage for pure fireball spam's dps to equal the dps given by including fire blast in your rotation (assumes zero haste, does not account for incinerate).

With 4t6, this becomes

0.43x + 725 = 1.05 * (1.15x + 719) / 2

and solving for x gives 2006, still well above typical spell damage levels.

Augmenting the formula to allow for spell haste (no 4t6, y denotes spell haste rating):

0.43x + 725 = (1.15x + 719) / ((3 / (1 + (y / 1570))) / 1.5)

Substituting in my own raid-buffed +dmg, x = 1100, it turns out that I would need ~325 spell haste rating (~21% haste) before pure fireball spam wins out.

That is really interesting, I really didn't think that including Fireblast in the rotation, especially with Empowered Fireball, would scale nearly that well.

So would it be worth pulling a point from somewhere and taking the one point in Imp Fireblast or should I just plan on using Fireblast in the rotation whenever it is off cooldown? Also revises my current thoughts on speccing points into Incinerate.

Something like this perhaps,

WorldofWarcraft.com -> Info -> Classes -> Mage -> Talent Calculator

Edited to add:

At 1500 damage you would need 190 haste rating before pure Fireball spam beats adding in Fireblast. Which seems to imply that it really is worth placing that point in Improved Fireblast since having static 190 haste rating and being hit capped with 1500 spell damage isn't feasible with the currently available gear choices. Even with the Skull you will just use pure spam while the proc is active and then add Fireblast back into the rotation when the buff fades.

Dustwhisper · 12/06/07, 9:44 AM

In a deep firespec which is 10/44/3 you have 4 points to waste. People usually put the first three in pyroblast, blastwave, dragonsbreath and coincidentialy the last point fits perfect inn Imp fireblast to make fireblast be ready for a 2fireball+fireblast rotation

Pintofbrew · 12/06/07, 10:30 AM

Originally Posted by Dustwhisper

In a deep firespec which is 10/44/3 you have 4 points to waste. People usually put the first three in pyroblast, blastwave, dragonsbreath and coincidentialy the last point fits perfect inn Imp fireblast to make fireblast be ready for a 2fireball+fireblast rotation

That will be small consolance given the incoming need for Icy Veins. Funny how the amount of haste needed to make Fball>Fblast is exactly the amount given by IV.

I find it ironic that we re-discover Fblast an improvement over Fball when a few months back data was posted advocating against it in the old thread.

Rouncer · 12/06/07, 10:35 AM

Originally Posted by Dustwhisper

In a deep firespec which is 10/44/3 you have 4 points to waste. People usually put the first three in pyroblast, blastwave, dragonsbreath and coincidentialy the last point fits perfect inn Imp fireblast to make fireblast be ready for a 2fireball+fireblast rotation

I was talking about for 2.3.2 and I don't think many people will be keeping the 10 points in Arcane once Icy Veins becomes available.

Dustwhisper · 12/06/07, 11:18 AM

Originally Posted by Pintofbrew

That will be small consolance given the incoming need for Icy Veins. Funny how the amount of haste needed to make Fball>Fblast is exactly the amount given by IV.

I find it ironic that we re-discover Fblast an improvement over Fball when a few months back data was posted advocating against it in the old thread.

From what I remember that old data advocated to not weave it inn at something like 1500-1600 spelldmg. The only reason I don't always weave it inn is mana and ping since it seems to slow down the sequence due to new castingsystem.

◊ manly · 12/06/07, 11:28 AM

All of this assumes 0 lag and a working server side spell queue. You guys need to run the proper tests on the new cast system to make sure it will output 0 latency on instant casts, since that would invalidate your conclusions.

Cardynal · 12/06/07, 11:54 AM

From what I've noticed, 1 point isn't sufficient to do a 2 fireball/ 1 fireblast rotation anymore with the new casting mechanic. When I've tested it out, the cool down isn't up when you're able to start casting the fireblast...which is at about 1/5 of the fireball cast bar left. With 1 point, you have to wait till the fireball is nearly complete before you can cast fireblast. With 2 points, the cool down comes back up just in time to cast it when my fireball is 2/3 of the way complete.

What this seems to do is start your global cool down before you even finish your fireball cast or at least as soon as you finish the cast, which means you can start your next fireball sooner.

Also, has anyone else noticed that Scorch is currently giving 2 stacks per hit on the Crusade Card? This may be old news but I just noticed it this week since I hardly pay attention to that stacking anymore.

Kilon · 12/06/07, 11:54 AM

I were wondering if the Fireball-Fireball-Fireblast rotations are still beneficial to your dps if you have enough haste to cast the 2 Fireballs before the cooldown ends (assuming you'll always lose at least a very little cast time to latency). Would you attain higher damage by just waiting out, i.e. casting 3 Fireballs prior to the Fireblast, wait until the CD finishes (I'm talking about the occasional 0.05-0.08 second, not for every Fireblast as I still encounter the issue of latency), or drop some other talent point to reduce the Cooldown on Fireblast by another 0.5 second?

Cardynal · 12/06/07, 12:13 PM

As posted above...there are really only 44 points in fire worth getting...blastwave and dragons breath aren't worth all that much as from my experience...they get me killed while AEing quite often from their larger amount of threat. Depending on your haste, you could drop all 3 points in imp fireblast.

Duodecimal · 12/06/07, 12:51 PM

If replacing clearcasting with the new mana gems is a wash, how does the extra 30 mana per second or so cost of weaving Fireblast factor into sustainability?

In my spreadsheet it nearly doubles mana consumption. Is that worth a ~3% dps gain?

Kasi · 12/06/07, 1:05 PM

It really depends if you have a shadow priest. It's probably much the same with me doing a 3 LB, 1 CL rotation. If you have a shadow priest, you probably won't be able to run oom as a fire mage no matter what you do (well other than spamming flamestrike). If you don't then yeah you probably won't be able to fireblast at all, and might have issues just maintaining a full fireball spam.

Pintofbrew · 12/06/07, 1:21 PM

As for latency/new casting mechanics affecting when you can start Fblast, who cares?

If Fblast is indeed more than 0.5 of a fireball, I'm not bothered if I can fit one for a 2*Fball,1*Fblast rot. Why not go 3*Fball,1*Fblast rotation instead? No excess points to be dumped in Imp Fblast and no question of whether there's some funny check serverside, blah blah. Though I do concede that manly has a point on whether the server handles the start of the next Fireball correctly in terms of timing.

Granted, 2Fball,1Fblast will be better than 3Fball,1Fblast on theortetical terms, but I wouldn't risk having server side inject added down-time because it won't handle cooldowns correctly.

Cardynal · 12/06/07, 2:08 PM

Originally Posted by Pintofbrew

As for latency/new casting mechanics affecting when you can start Fblast, who cares?

If Fblast is indeed more than 0.5 of a fireball, I'm not bothered if I can fit one for a 2*Fball,1*Fblast rot. Why not go 3*Fball,1*Fblast rotation instead? No excess points to be dumped in Imp Fblast and no question of whether there's some funny check serverside, blah blah. Though I do concede that manly has a point on whether the server handles the start of the next Fireball correctly in terms of timing.

Granted, 2Fball,1Fblast will be better than 3Fball,1Fblast on theortetical terms, but I wouldn't risk having server side inject added down-time because it won't handle cooldowns correctly.

You are being rather dismissive for a theory crafting thread. Have you actually tried it yourself to see what i'm talking about with the GCD starting earlier when using fireblast with 2 points in it? Even it if starts .1 seconds earlier...that's 4 seconds of dps time gained over a 5 minute fight....which is another fireball. This is theory crafting after all...not oh it's good enough.

I'll test this more this weekend to see if it does infact start the GCD quicker.

Vannik · 12/06/07, 2:16 PM

Avg fireball = ((1-crit%)*((904+713)/2+ dmg*1.15)*1.13 + cit%*((904+713)/2+ +dmg*1.15)*1.13*2.1)
Avg fireblast = ((1-crit%-0.04)*((730+865)/2+ +dmg*0.43)*1.13 + 2.1*(crit%+0.04)*((730+865)/2+ +dmg*0.43)*1.13)

This is the average value including ignites for critting that your fireball and fireblast will hit for. (A crit is 150% of spell damage + 40% of the total crit. Yes, I know you can't count crit damage as instant damage, but these crit %'s are only true as you look towards infinity, so if you do your calculations based on your exact crit% this fact will be negligible.

With these numbers you need 3200+dmg 30% crit to make fireball spam = fireblast rotation. I agree with the fact that due to spell queing and latency you'd need 2 points in fireblast so you can que it like you can fireball.

ebbv · 12/06/07, 2:48 PM

DISCLAIMER!
So this isn't really that meaningful, I was just curious what the numbers would be. Obviously more thorough testing needs to be done. I just wanted to get the ball rolling.

I'm wearing some haste gear so my Fireball Cast time should be 2.87sec. I realize I could have simplified this by taking off the haste gear but I am curious about how the haste gear impacts it too, so here we are.

35.890 Your Fireball crits Dr. Boom for 3417 Fire damage.
38.796 Your Fireball hits Dr. Boom for 2176 Fire damage. 2.906 (+36ms)
41.828 Your Fireball crits Dr. Boom for 3536 Fire damage. 3.032 (+162ms)
44.859 Your Fireball hits Dr. Boom for 2150 Fire damage. 3.031 (+161ms)
47.671 Your Fireball hits Dr. Boom for 2203 Fire damage. 2.812 (-60ms?)
50.687 Your Fireball crits Dr. Boom for 3242 Fire damage. 3.016 (+146ms)
53.796 Your Fireball hits Dr. Boom for 2191 Fire damage. 3.111 (+241ms)
56.609 Your Fireball hits Dr. Boom for 2302 Fire damage. 2.813 (-60ms?)
00.359 Your Fireball hits Dr. Boom for 2113 Fire damage. 3.750 (I flubbed this one)
03.343 Your Fireball hits Dr. Boom for 2210 Fire damage. 2.984 (+114ms)
07.171 Your Fireball crits Dr. Boom for 3355 Fire damage. 3.828 (I flubbed this one too)

Average Latency 92ms

The first thing that jumps out besides the 2 I messed up (which we'll just throw out), is the 2 which have shorter than expected cast times. I believe this is a feature of the new cast system and with enough practice and perhaps some adjustment to Quartz once we understand it better, may become repeatable.

Anyway the above Fireball spam is easy to understand and obvious.

15.625 Your Fire Blast crits Dr. Boom for 2040 Fire damage.
16.437 Your Fireball hits Dr. Boom for 2166 Fire damage. 0.812
21.156 Your Fireball hits Dr. Boom for 2258 Fire damage. 4.719
23.968 Your Fireball crits Dr. Boom for 3428 Fire damage. 2.812
26.156 Your Fire Blast crits Dr. Boom for 2099 Fire damage. 2.188
27.203 Your Fireball hits Dr. Boom for 2185 Fire damage. 1.047
31.734 Your Fireball hits Dr. Boom for 2225 Fire damage. 4.531
34.125 Your Fire Blast hits Dr. Boom for 1329 Fire damage. 2.391
34.546 Your Fireball hits Dr. Boom for 2049 Fire damage. 0.421
39.640 Your Fireball crits Dr. Boom for 3217 Fire damage. 5.094
42.500 Your Fireball crits Dr. Boom for 3349 Fire damage. 2.86
42.687 Your Fire Blast hits Dr. Boom for 1339 Fire damage. 0.187

Here is one example of Fireball, Fireball, Fireblast combat logs using the same gear with 2.87 cast. Note that it is not as intuitive to read due to travel time mixing up the time of damage dealt vs. cast time. Let's break it down.

A - Fireball
B - Fireball
C - Fireblast

Because of Travel Time C will hit before B, and this is what obfuscates the combat log. We expect a single rotation with my haste gear to take:

2.87*2 + 1.5 = 7.24sec

Let's look at what the rotations actually took from another Ball, Ball, Blast log. This time with rotations labeled:

17.578 Your Fire Blast crits Dr. Boom for 2123 Fire damage. C0
23.406 Your Fireball hits Dr. Boom for 2138 Fire damage. A1 5.828
25.906 Your Fire Blast hits Dr. Boom for 1351 Fire damage. C1 2.5
26.328 Your Fireball crits Dr. Boom for 3288 Fire damage. B1 0.422
31.046 Your Fireball hits Dr. Boom for 2199 Fire damage. A2 4.718
33.484 Your Fire Blast crits Dr. Boom for 2029 Fire damage.C2 2.438
34.062 Your Fireball hits Dr. Boom for 2210 Fire damage. B2 0.578
39.218 Your Fireball hits Dr. Boom for 2314 Fire damage. A3 5.156
41.515 Your Fire Blast hits Dr. Boom for 1374 Fire damage. C3 2.297
42.031 Your Fireball hits Dr. Boom for 2055 Fire damage. B3 0.517
46.953 Your Fireball crits Dr. Boom for 3262 Fire damage. A4 4.922
49.218 Your Fire Blast hits Dr. Boom for 1402 Fire damage. C4 2.265
50.046 Your Fireball crits Dr. Boom for 3467 Fire damage. B4 0.828
54.890 Your Fireball crits Dr. Boom for 3470 Fire damage. A5 4.844
57.203 Your Fire Blast hits Dr. Boom for 1310 Fire damage. C5 2.313
57.812 Your Fireball hits Dr. Boom for 2208 Fire damage. B5 0.609
03.390 Your Fireball crits Dr. Boom for 3303 Fire damage. A6 5.578

So if we take a given set we should have the time of that rotation. Note that any given A will contain the GCD from the C of the set before it. Let's see what these 5 rotations actually took.

A1 5.828
B1 0.422
C1 2.500
T1 8.75 (+1.51*)

A2 4.718
B2 0.578
C2 2.438
T2 7.734 (+0.510)

A3 5.156
B3 0.517
C3 2.297
T3 7.97 (+0.730)

A4 4.922
B4 0.828
C4 2.265
T4 8.015 (+0.775)

A5 4.844
B5 0.609
C5 2.313
T5 = 7.766 (+0.526)

* - Slow first fireball, so let's throw this out.

Ignoring the first set, the last 4 still each took an average of 635ms longer than we'd expect to complete the rotation. That is to say that our 7.24sec rotation becomes 7.875sec rotation in reality, and that means that 8% of your cast time is being eaten up by latency.

By comparison with my Fireball spam example above each Fireball averages 92ms latency, that means each Fireball goes from 2.87sec to 2.962sec on average. That's just 3% of cast time eaten by latency, or almost 3 times more efficient than Fireblast rotation.

This is why Fireball Spam is better DPS than Fireblast rotation even though on paper it seems like Fireblast should be better.

I encourage someone else to do some more thorough testing, I just wanted to get this out there.

Vannik · 12/06/07, 3:04 PM

Originally Posted by ebbv

DISCLAIMER!
So this isn't really that meaningful, I was just curious what the numbers would be. Obviously more thorough testing needs to be done. I just wanted to get the ball rolling.

I'm wearing some haste gear so my Fireball Cast time should be 2.87sec. I realize I could have simplified this by taking off the haste gear but I am curious about how the haste gear impacts it too, so here we are.

Average Latency 92ms

The first thing that jumps out besides the 2 I messed up (which we'll just throw out), is the 2 which have shorter than expected cast times. I believe this is a feature of the new cast system and with enough practice and perhaps some adjustment to Quartz once we understand it better, may become repeatable.

Anyway the above Fireball spam is easy to understand and obvious.

Here is one example of Fireball, Fireball, Fireblast combat logs using the same gear with 2.87 cast. Note that it is not as intuitive to read due to travel time mixing up the time of damage dealt vs. cast time. Let's break it down.

A - Fireball
B - Fireball
C - Fireblast

Because of Travel Time C will hit before B, and this is what obfuscates the combat log. We expect a single rotation with my haste gear to take:

2.87*2 + 1.5 = 7.24sec

Let's look at what the rotations actually took from another Ball, Ball, Blast log. This time with rotations labeled:

So if we take a given set we should have the time of that rotation. Note that any given A will contain the GCD from the C of the set before it. Let's see what these 5 rotations actually took.

A1 5.828
B1 0.422
C1 2.500
T1 8.75 (+1.51*)

A2 4.718
B2 0.578
C2 2.438
T2 7.734 (+0.510)

A3 5.156
B3 0.517
C3 2.297
T3 7.97 (+0.730)

A4 4.922
B4 0.828
C4 2.265
T4 8.015 (+0.775)

A5 4.844
B5 0.609
C5 2.313
T5 = 7.766 (+0.526)

* - Slow first fireball, so let's throw this out.

Ignoring the first set, the last 4 still each took an average of 635ms longer than we'd expect to complete the rotation. That is to say that our 7.24sec rotation becomes 7.875sec rotation in reality, and that means that 8% of your cast time is being eaten up by latency.

By comparison with my Fireball spam example above each Fireball averages 92ms latency, that means each Fireball goes from 2.87sec to 2.962sec on average. That's just 3% of cast time eaten by latency, or almost 3 times more efficient than Fireblast rotation.

This is why Fireball Spam is better DPS than Fireblast rotation even though on paper it seems like Fireblast should be better.

I encourage someone else to do some more thorough testing, I just wanted to get this out there.

Wow. . . with 1250 dmg and 70 haste yes, even mathmatically it will go to fireball's favor. Do you have one point in fireblast? Probably. The most effective way to analyze this is just to figure the time between fireblasts. But you should have 2 points in imp fireblast if you want to better take into effect of the spell queing and that will also account for some haste to a point.

beta4Life · 12/06/07, 3:04 PM

Originally Posted by Rounced

At 1500 damage you would need 190 haste rating before pure Fireball spam beats adding in Fireblast. Which seems to imply that it really is worth placing that point in Improved Fireblast since having static 190 haste rating and being hit capped with 1500 spell damage isn't feasible with the currently available gear choices. Even with the Skull you will just use pure spam while the proc is active and then add Fireblast back into the rotation when the buff fades.

1500 spell dmg is definitely feasible raid buffed. I believe that I sit right around 1540 fully buffed (imp DS, flask, food, oil, and shammy totem), with 145 passive haste +skull of guldan (which is equivalent to about 29 passive haste used every cooldown i believe). I find that squeezing fireblasts into my rotation doesn't boost my dps enough to even notice it (due to the unpredictable nature of crits), the mana inefficiency isn't really a problem with evocate, but if you think about it, your fireblast would have to increase your dps by more than 60ish, in order to make up for the 10 seconds you have to spend evocating every fight because you throw the fireblasts in.

I am not saying that fireblast isn't worth doing for everyone, but once you reach the end of the game, and have 1250+ spell dmg unbuffed, with a decent amount of haste, it is unlikely that fireblast will do anything for you.

Just to preempt any questions about my low hit, i always have a draenei elemental shaman in my group, making the hit cap for me only 9% (4% from shaman, 3% from EP).

ebbv · 12/06/07, 3:24 PM

Originally Posted by Vannik

Wow. . . with 1250 dmg and 70 haste yes, even mathmatically it will go to fireball's favor. Do you have one point in fireblast? Probably. The most effective way to analyze this is just to figure the time between fireblasts. But you should have 2 points in imp fireblast if you want to better take into effect of the spell queing and that will also account for some haste to a point.

I obviously have 1 point in Fireblast, yes. Even with 2 points, the way that latency interacts with instant GCD spells like Fireblast is what kills it. Also don't quote long posts if you're just going to write a 3 line response, please.

◊ manly · 12/06/07, 3:31 PM

Well, your methodology is somewhat flawed. The combatlogs show only local timestamps. This means the order in which you see things listed have no correlation with the order it ran on the server. The only way to know if the 'server side queue' does allow for 0 latency is to cast as many fireballs as possible, and compare the total time taken vs predicted time.

The best way to test this out would be no haste gear, downrank to the first rank of fireball with 3s cast time, and spam that.
Then do the same with ice lance spam.
Then do the same with downranked fireball/icelance.

In any case, it isn't productive to point out that 1500 damage is possible in the top end gear. If you like doing that, you can have fun posting in the wow forums. The issue that will have by far the most impact of the theory is whether or not zero latency is possible. Going through very rough numbers, if you gain, say, 3% more dps with fireblast every 3 fireballs (ignoring gear and whatnot), but at the same time it costs you an added 100ms lag every 1.5seconds, then it is readily apparent that fireblast is not the way to go.

ebbv · 12/06/07, 4:07 PM

Right, that's what I was doing a rough test for. To see how the fact that Fireblast doesn't interact well with the new cast system impacted latency in the rotation. And I think even with the tiny little example I used, it's readily apparent that Fireblast doesn't play well.

Tomorrow is a non-raid night so if nobody else has taken up the task I will do some more thorough testing then.

Rouncer · 12/06/07, 4:59 PM

Originally Posted by ebbv

DISCLAIMER!
So this isn't really that meaningful, I was just curious what the numbers would be. Obviously more thorough testing needs to be done. I just wanted to get the ball rolling.

Lots more stuff :-)

Wondering if the issue has to do with the possibility that they added a server-side "queue" to the casting system. Well not so much a queue as a fudge factor. There was talk on the UI forums by Blue that they were thinking of allowing a bit of leeway into the system as to when the server would receive the request.

So the server would be eligible to receive a new request say up to 250ms before it was actually free to cast a new spell and then would cast that spell exactly when freed.

Not sure if it ever was incorporated into the system or not but since you say you had one point in Improved Fireblast that would mean that the server would only allow the next cast every 7.5 seconds but since you have some haste involved, 2.87 fireballs, you would be adding 0.26 seconds into the equation since that would be necessary to complete the cooldown with only 1 point in Improved Fireblast.

So your average rotation ended up being 7.875 and the cooldown for Fireblast with 1 point is 7.5 so for the 3 spells you actually only added 0.375ms or 0.125ms per spell which doesn't look too much different then the 0.092ms you had with Fireball spam and could be accounted for by the Human factor especially in such a small sampling.

What I'm suggesting is that the Fireblast was eligible to be cast even with only one point in Improved Fireblast due to the "fudge" factor in the casting system allowing the cast request to hit the server and be responded to even though you actually shouldn't have been able to cast due to Fireblast still being on cooldown due to the haste.

Someone should hop (I'm at work atm) on the test server and either repeat the test with no haste gear in place or take 2 points in Improved Fireblast and see if the same latency issue is still there.

clairecakes · 12/06/07, 5:06 PM

So the guild is making serious attempts on Al'ar and then we'll be moving back to SSC and working on Hydross. Now the idea of deep fire or deep frost spec doesn't really appeal to me because that means respecs between boss attempts. I've been arcane/frost and done decent damage (and jaw dropping amazing when a deep frost mage puts up Winter's Chill) but I was thinking of maybe trying a 40/5/16 spec. I can use the fireballs as filler on hydross and go back to frostbolts for al'ar.

Would that work? Or am I over thinking/ignoring things?

My big goal is to get my other piece of T5 spec arc/frost and woo another mage to spec deep frost.