Grai

*deleted* Post in the wrong thread.

 

Wizeowel

It's not just that it's an early post in the thread, it's that the thread is in fact a consolidation of knowledge from the older theorycrafting threads which had grown too large for easy digestion. Vontre kindly made a post summarising the conclusions on various matters so that people didn't have to ask in the theorycrafting and repeat the discussions. The thing is that the "AE or FS in MH" question was done to death, and while a consensus has been reached, it seems that you have both come too late to the forum to realise this.

This was the question. Not "about AOE theorycrafting in general", yes specifically "what do I do on Hyjal trash". Explicitly stating he wasn't threat capped. I marked it for moderation since it was in fact answered in the first post.

To summarise what we've covered so far.
* AE is better than FS on Hyjal trash.
* Leading with FS doesn't give more damage. As I explained previously, if you know the right moment to start casting FS, then you also know the right moment to cast AE. In fact you are more likely to know the right moment for AE since it's not a prediction about the situation 3 seconds into the future.
* Yes flamestrike suffers from click lag. No vague "it's worked pretty well for me" please -- post a WWS proving it doesn't.
* Yes a lower tps spell with the same damage as a higher tps spell in a threat capped situation will result in more overall dps. [(tank total threat) / (threat per cast)] x (damage per cast).

So please both of you: look at the cold hard facts. Stop bending the scope of the discussion to suit your needs.

 

solbergb

I believe I was responsive to his question in a way that the first page was not. It didn't cover # of mobs, for example which would help a newbie understand why it's better in a raid situation than when you try it out solo vs 2-3 mobs and it does less damage than your other options.

Nor did I post any misinformation although I did not structure it in the most comprehensible way. Lets cover some other points in summary, as I can't object to any that you posted.

1. Flame strike is lower TPS than Arcane Explosion, if you are not within 10' of the mob and have the Burning Soul. Click lag is irrelevant, if .5 sec is too fast for you, assuming 1s or more just lowers threat further. Blizzard is also lower TPS, but it starts sooner, which can backfire. Flamestrike->Blizzard is a low threat/low DPS rotation if you keep your distance. Opening with a flame strike "for threat reasons" is focusing on "do I get eaten" not "do I do the most DPS".

2. Flame strike is less DPS than arcane explosion spam unless the number of mobs is VERY high, and only if the DOT tics the whole time. AE caps at ~16 mobs my gear level..however flamestrike (assuming 3 second cast, 1 second click time and DOT runs to completion) instead of doing 2.67 arcane explosions requires 33 mobs to break even - not very likely.

3. Arcane explosion spam is less DPS than mixing in blast wave and dragon breath unless those spells are AOE capped. The break even point for blast wave/DB is around 10 mobs and peak performance is about 7 mobs, at least at my gear level. The instant fire AOE spells are just a threat spike at 10+ mobs without doing more damage, even considering higher crit % for fire and ignite damage.

4. Master of elements usually means using the instant fire spells is similar DPM to Arcane Explosion.

Now people don't have to read the last two pages of discussion and can focus just on these points in this post and the last one. If you still think these points are in error, I'm interested to know why.

 

Wizeowel

That's fair enough and, in his response, he seemed to take only 1 part of your information away with him, namely: the dot part of FS keeps ticking and isn't affected by the AoE-cap. This is perfectly valid, but if you look back at my initial response you'll see I was objecting to the idea that leading with flamestrike is somehow safer and/or more dps. Furthermore the whole issue of click-lag was brought up by Kentosani himself when he mentioned needed 1 second to aim his FS. Then also your big calculation assuming only 7 mobs in Hyjal trash waves, well, we can't really prove anything if we keep moving the goalposts.

1. Let's look the numbers.

Arcane explosion for Kentosani.
722-808 per cast per mob (690-722 hits and 1035-1083 crits, 23.69% crit rate, 1.5 multiplier)
516-577 per second per mob (his cast time for AE is 1.4 seconds)
5157-5771 dps on 10 mobs
269-301 tps per mob (assuming salvation and arcane subtlety)
Sustained tank tps needed for arcane explosion: 273 (assuming 110%)

Flamestrike initial strike for Kentosani on 10 mobs
1067 per cast per mob (783 hits and 1644 crits, 32.69% crit rate, 2.1 multiplier)
410 per second per mob (cast time for FS is 2.8 seconds)
4104 dps on 10 mobs
320 tps (assuming salvation and burning soul)
Sustained tank tps needed for the direct damage part of flamestrike: 246 (assuming at range 130%)

Now the the dot part is 692 over 8 seconds. This is 173 damage per mob per 2 seconds. That's 67 tps (with salvation and burning soul) and would need the tank to do an extra 51 tps on each mob even if Kentosani is at range, at which point he'll lose time running in for the following DB/BW/AEs. As soon as he is in melee range, those numbers become: 290 for the initial strike and 61 tps for the dot.

So addressing the question of leading with FS and whether it's safer. If Kentosani crits even once on the initial flamestrike, which is likely, then he has an initial threat on one mob of at least 1287 with 102 threat incoming every 2 seconds after that. The tank would have to have already done 990 threat on that mob, and moreover he'd have to stay at range or see that number rise by 18%. However, if Kentosani would lead with AE then any crit would be 565 threat. The tank would have had to do 514 threat. As you can see, AE can safely start a lot earlier. My point about this was that as the consecrate goes down, you can see each tick coming on omen. As soon as you see over 500 threat on a mob which the pally isn't also meleeing, then you can press the AE button with confidence. In comparison, can you really seriously predict the best moment for starting to cast FS?

Conclusion: lead with AE.

2. I'm not sure how you did your calculations to get 33 mobs. Flamestrike is AoE-capped at 7830. For Kentosani, his normal hit from FS is 948 - 1070 on each mob. So he is AoE-capped on FS already at 7 or 8 mobs. His AE hits normally for 690-722 so would be capped at around 14 mobs. This is what I objected to in your calculations. You took the assumption of 7 mobs, which of course makes FS more dps (but even higher tps). Unless your guild does very specific single targetting this will only be the case for the last few seconds of each wave.

Conclusion: don't spam FS.

3. There's no question that DB and BW are higher dps than both AE and FS. So if you are not threat capped then you can mix those into any rotation (AE or FS or combination) providing you are not at range. The issue I had with this, I'll explain again. Your dps/threat/mana calculations were showing AE rotation without BW and the FS rotation with BW. Since you can mix these spells into both rotations I don't understand why you excluded them from your AE-spam calculation. If anything AE is more likely to include other point-blank range instants than FS since as both you and Ilyawen were arguing about the reduced threat from FS at range.

Conclusion: you have to be in melee range if you want max dps.

4. We're not talking about DPM. The waves are 2 minutes maximum, you can drink in between. Even if your raid dps is so poor that you go oom from Hyjal trash waves there are several options: mana gem, mana potion, mage armor. Even so, if you want to calculate in MoE returns, then I think you'll find that overall AE is still less than 10 mana per second more expensive (at least on Kentosani's gear it is 9 mps more according to Rawr).


So overall I would say that if Kentosani wants to maximise his dps safely on Hyjal trash, then he'd want to do the following: wait for the tank to have 500 threat on all mobs (I think this is 2 consecrate pulses); spam AE until his tank has 1600 threat higher than him, then hit blastwave; then continue with AE until tank has around 1200 threat higher, then go with flamestrike-4xAE rotation. Of course he'd have to aim his flamestrike a lot quicker than 1 second

 

solbergb

Nothing to object to in the above. I don't disagree with the conclusions for Hyjal trash.

Responding to this one:



This is a good point, although I think you have to consider the idea that you started earlier with AE to give the full picture. Lets assume that you'll have at least one crit on one mob on first AE, but you're not that likely to crit on same mob the second or third time. So one crit, and the rest normal the threat curve looks like this:

1. AE spam right when he begins casting flamestrike: AE spam = 1318 threat (565 immediately)
2. AE spem 1.5 sec after he begins casting flamestrike: AE spam = 941 threat (565 after 1.5s)

So it depends on how the tank is getting his threat, how steep the curve is. If tank is starting at zero and generates 1300 threat in 3 seconds then the flamestrike is ok, and AE spam #2 is ok but AE spam #1 is not ok.

If the tank has a flatter curve (say, 650 threat at time of flamestrike cast) then AE spam #1 is safe.

If the curve isn't a line, but is instead something like _/ (he's got some kind of burst effect or needs time to swipe all possible targets with multi-attack stuff as per bear) you might not be able to start any earlier, but when you can start at all, he's got more than 1287 threat.

Pally tanks have a fairly predictable threat curve...consecrate+passive damage as stuff beats on them. So probably it is mostly judging how steep the threat curve is, whether you can really AE earlier or not. If you don't know the steepness of the curve or the shape of the curve, backloading your own damage might be safer even if it costs you some potential DPS.

The idea with flamestrike as a threat tool is you can choose to stay at range if the threat situation still seems iffy, or you can move in slightly and burn your instant AOE's if threat seems under control (or if somebody who AOE's hotter than you has already opened up). You also have more time to react if you did pull aggro and might be able to get back behind threat curve with ice block, or dump threat with kiting+invis.



>2. I'm not sure how you did your calculations to get 33 mobs.


The DOT portion of the AOE isn't capped, and I said the numbers were with my gear. The initial damage of both flamestrike and arcane explosion are capped at that number of mobs, so what you're looking for is where the DOT exceeds the significant advantage AE has from both higher cap and faster cast time.

normalized for damage/second, we're solving for "number of mobs" in this equation (NM)
AE = arcane explosion damage cap, FS = flamestrike damage cap, FD = flamestrike DOT at my gear level
Done right haste should be worked in somehow on the damage capped figures.

Assuming you can get 3 arcane explosions off in an 8 second flamestrike rotation. (which is probable given some haste but slower than .5 second target time) you get

AE/(1.5*NM) = FD/(8*NM)+AE*3/(8*NM)+DOT/8
or

NM = (2.33AE - FD)/DOT


the "1.5" in the AE on the left should really be reduced for haste, but the AE on the right should not, as it is the DOT that limits the entire rotation and extra time saved by faster GCD's will likely be consumed in click lag.

The above also ignores crits, which is probably dumb because crits aren't capped and are gear sensitive. Including crits should favor fire, as crit rate is higher both in % crit chance and in ratio to base damage (with ignite). You'd add an additional parameter on left and right side of equation that would not be divided by NM, would end up in divisor and would both favor the FS side of the equation and get larger proportionally by gear via spell damage coefficient. I can't speculate as to how significant this is compared to the DOT without attempting the rest of the math.

With my gear I get 27 mobs from this formula assuming 3 arcane spam and 38 mobs assuming I only mix two arcane explosions in between flamestrikes. The 33 figure was splitting the difference, assuming sometimes I would get fast enough performance in click lag, sometimes not.

For him -

NM = (2.33*10100 - 7830)/692 = 22 (assuming 3 AE between each flamestrike, more plausible if his haste is significantly better than my limited amount and/or he's used to the rotation and has his keystrokes down pat)

NM = very high in Burning Crusade at pretty much any gear level. I have encountered very few instances where there were more than 16 mobs in a AOE group, but my experience is mostly 5-10 man stuff.

 

◊ Vontre

I suppose I should update this for 3.0 and WLK now huh? Oh wait nothing changed hahahahahaha!

Edit: I just updated the front post with stuffs. Horrah.

www.magegraf.com

Raiding is full of challenge. Sometimes there is fire. You have to not be in the fire.

"We agree with Communism." - Greg Street 2009

 

solbergb

You should probably modify this: "Beyond that, gear can be used to improve your hit chance to a maximum of 99%. Once you reach the 99% hit mark, further hit rating is completely useless and you considered "capped". For arcane mages, the cap is 76 rating. For fire and frost mages, that number is 164 rating."

you can go to 100%, arcane only gets 3% and arcane/fire mages won't have elemental precision.

The AOE section isn't precise enough with respect to blizzard. It's only highly effective if talented, especially deep frost talented. Frostfire builds will have some talent points that support it but not as much as deep frost and the other builds won't find blizzard a lot more effective than their other alternatives.

Does the frostfire DOT munching apply to the frostfire build too? Or do the GCD's from hot streak, living bomb etc get you off the cycle where it refreshes but never tics? Or am I just misunderstanding that section.