Simcraft Stat Weighting Progression by iLvl

[Zrog]

So I've been running my own Simcrafts since about iLvl 640, and I've noticed that the weighting changes sharply in favour of Mastery > Multi after about 650-ish iLvl.

... this is from memory, so bear with me until I can get home and look up my saved Sims.

(Normalized)
Note: Int and SP are always 1.0 and 0.9
640 iLvl = Crit 0.6; Multi: 0.41; Mastery: 0.41; Haste: 0.40; Vers: 0.38

645 iLvl = Crit 0.62; Mastery: 0.42; Multi 0.40; Vers: 0.39; Haste: 0.38

655 iLvl = Crit 0.65; Mastery: 0.61; Vers 0.41; Multi: 0.41; Haste: 0.38

I was happy with a Frost/Fire dual spec because I could get away with Crit/Multi gear and still do decently, but lately Mastery is severely lacking when I go Fire, as evidenced by my low Ignites (from other thread).

I also noticed that at very low gear levels (615 - 630), I was very Mastery-heavy; it just seemed to be the gear that was available at the time (Heroic dungeon rare gear), with Multi coming more often on low-epic gear.

Thus, it seems like the "natural progression" of spec, per Blizzard's itemization and spec-to-item interaction is Arcane --> Frost --> Fire/Arcane

Thoughts?

Zrog

EDIT: Actually, my mastery might be pretty good, at 21% unbuffed. Not sure what the issue is...

[Wilderness]

I don't think its the best idea to draw conclusions from specific sims about different ilvls of gear - stat weights are definitely influenced by what gear you have available, and thus what stats you have a lot or a little of. So if you have heavy MS gear with little Mastery, then Mastery is going to come up in relative value when you run those sims.

So I don't agree with your conclusion that the "natural progression" for mages is to be Arcane at low levels, Frost at medium ilvls, and Fire beyond that.

[Zrog]

You didn't experience a high amount of Mastery gear at low levels?

Edit: I meant to say that at higher iLvls, depending on specialization, Fire OR Arcane would probably be superior to Frost, due to scaling (Arcane for single target, Fire for multi-target).

PS - I agree that more sims would be required, and I was hoping someone else had been running some over time, to make a comparison with mine and my experience.

[Wilderness]

You didn't experience a high amount of Mastery gear at low levels?

Not really - its been a while since I last looked through all the normal and heroic dungeon gear, but from what I remember there was a metric ton of different drops with different stats on them. I know all the 615/630 mission gear that you can get can be very random as well. You personally may have ended up getting a bunch of pieces with predominately Mastery on them, but I don't think that's representative of the gear as a whole. Nor would that alone be a reason to play Arcane over other specs - due to the balancing between specs and how Arcane/Fire have generally scaled better than Frost, at low gear levels Frost has almost always been the best option.

PS - I agree that more sims would be required, and I was hoping someone else had been running some over time, to make a comparison with mine and my experience.

What sims would you like to see? There are sims on the main page of Simulationcraft that represent the different specs at different gear levels - what else are you looking for?

[Zrog]

I guess I found the Simcraft page to jump too far between each iLvl bracket, for starters. I'm also the type of player who believes that one spec is never optimal for all situations, and it will often pay to collect / save gear for more than one spec. The question is, what stats do I want, and what (off)spec should I be aiming for?

While I agree that there's a value in seeing the difference between Heroic gear vs. Full Highmaul vs. Full Mythic, it doesn't really give an idea of what it's like when you're halfway between and wondering where one's optimal spec is, and what kind of DPS one should be able to pull at that gear level and spec compared to other mages.

Simcraft ranking shows an "optimal" spec for each gear level for single-target fights, but there should be an approximate point where one spec exceeds another, based on not only the spec but also the boss fight type.

I note that the specs always rank Arcane - Frost - Fire, regardless of gear level. I think that's relatively new - they used to be in different orders based on gear level. I find this disappointing - are we all destined to play Arcane to be "optimal"? There's quite a large DPS gap between each spec, in each sim...

EDIT: I notice that they've disabled the Tier-piece bonuses in the recent Sims, meaning that these current posted Sims are only indicative of the "best spec" if your gear level is currently the same as what was simulated, and it doesn't predict what the spec rankings will look like once set pieces enter the picture.

EDIT: Am I off base here? Is there no way that Simcraft can be used to predict where one should head in terms of gear and spec? For example, at what point should I start to transition toward Fire gear, when I'm currently a Frost mage?

[Frosted]

There is more to choosing a spec than just secondary stat distribution.

Example: Highmaul gear is extremly heavy in MS, yet arcane out performs it on single target fights. Thinking that because you have high mastery (or crit. or MS) means that you should be playing that spec doesn't make sense.

How are you scaling the item level? If you're changing gear (and thus changing the secondaries you have), then I do not believe you conclusions are valid, since altering secondary stat use will change their values, and will act as a confounding variable here.

Itemization may effect spec parity to an extent, but there are a lot more factors that go into it than just secondaries. Also not sure why you would ever choose a spec based on stat weights. The value of interest is DPS, not how a spec is scaling with stats at some particular ilvl.

I really don't know what you want, actually. Are you expecting that between heroic and mythic ilvl some crazy jumps in DPS are going to happen, such that like at 675ilvl it's frost >fire>arcane, then you see what the sims on the front page of simulationcraft.org show as you go to 685? You're talking like you expect there is be large shifts in between the ilvls that are typically run. I'm not sure this is the case.

If you want to see like, the exact ilvl where fire overtakes frost, or something like that, then you can just do something like using the default normal mode profiles, and scale_to_itemlevel argument to scale in 1 ilvl incriments, plot the DPS of all three specs as a function of ilvl, and see when they switch around in position.

You could then repeat this with tier bonuses turn on, and show how the tier bonuses effect things.

Stat weights would then be used as some indicator of why these changes are happening.

[Zrog]

That's a good suggestion, Frosted, with the incremental iLvl increase. I just don't know if I have that much time to run that many Sims...

However, I guess that's the only way to really know, otherwise it's just conjecture.

[ZerotuL]

That's a good suggestion, Frosted, with the incremental iLvl increase. I just don't know if I have that much time to run that many Sims...

However, I guess that's the only way to really know, otherwise it's just conjecture.

I wouldn't put alot of trust in simcragt. It calculates DPS under certain conditions and circumstances. You can perform much better then simcart shows.
That's true- a lot of MS gear falling here and there, but i consider it overrated due to it's "chance" nature. If compare it to haste or mastery it doesn't work all the time- it has chance to be involved into DPS rotation. Haste and mastery do their job 100% of the time you dps the boss.
I think MS will lose it's current priority with top or almost top current gear and we will return to classics of fire mage craft-crit/haste<->mastery
Right now MS is a good stat to stack but later on you will have a lot of headache changing MS gear to something better.

All above is purely my own consideration of the current state and future of fire mages. Feel free to criticize and provide suggestions.

[Frosted]

Why would you not put trust in simulationcraft?

You're going to need to provide something a bit more compelling than because "it calculates DPS under certain conditions and circumstances." Simulationcraft doesn't "calculate" much.

Do you also consider crit overrated due to it's "chance" nature?

[Komma]

I wouldn't put alot of trust in simcragt. It calculates DPS under certain conditions and circumstances. You can perform much better then simcart shows.

This is as meaningful as saying "I wouldn't trust numbers because reasons".

As with all "simulationcraft is not good" complaints, you're invited to present your reasons as to why they're bad.

That's true- a lot of MS gear falling here and there, but i consider it overrated due to it's "chance" nature. If compare it to haste or mastery it doesn't work all the time- it has chance to be involved into DPS rotation. Haste and mastery do their job 100% of the time you dps the boss.

If for some reason you think that "chance" reduces the value of an object, I recommend that you stay away from all proc trinkets, from all specs, and in fact, from this game in general. Most spells have damage ranges, so no, haste and mastery don't "do their job 100% of the time" either.

Claiming that a stat is "overrated" because it is a "chance" is a ridiculous argument, because over any reasonable duration, the "chance" evens out. You cast more than 500 fireballs in a raid night. Not getting a multistrike is less likely than randomly typing and recreating the entire works of Shakespear.

With your last few posts, you've shown that you don't really understand how current Mage theorycraft is done. I recommend reading more on how things work before offering opinions based on your own feelings.

[Herpie]

Pretty much what Komma said. While i agree that its sometimes good to ignore sim/theory crafting in favour of common sense to choose another talent or glyph to compensate for certain bossmechanics that prevent you from performing optimally, things like statwheight are too complex (for me at least) to make an accurate estimation from just using common sense... I need a guideline. Haste and mastery for example do not have 100% "uptime" as suggested. Fire Mastery will only have value after you crit, and haste also increases in value with high crit. Crit and multistrike will both increase the value of ms and haste. And than there is the flatout dmg increase from versatility. I wouldnt know where to start determining which stat would be best to take because of these interactions if it wasnt for the peope running simulations.

[Xenost]

Is there a breakpoint in crit where the last change (nerf of pyroblast, increased value of crit) becomes a dps gain ?