Anaxie's Totally Revolutionary SoO Retribution T16 BiS & Theorycraft

[Deleted]

Not quite satisfied yet. Anaxie page 1 saying we should gem exp / mastery after reaching ~40% haste unbuffed, not str / mastery. That means before reaching ~40% we'd have to gem exp / haste instead of str / haste. Logic, right? Esp. if all of the 3 secondary stats weight higher than str. gemwise (80 str * 5,56 = 444,8 / 160 haste * 3,39 = 542,4 / 160 crit * 3,38 = 540.8 / 160 mastery * 3,17 = 507,2). gemming exp instead of str frees up secondary stats to reforge into, which all weight higher than str. no matter which one it is. So my question still remains: why do all gem str. / haste before reaching the 40%, if it looks obviously wrong?
Do I miss an important point? Please give advice.

Who gems Str/Haste? I asumed it was basic knowledge that it is wrong? You always want to fill up sockets with secondary stats, always.

[Deleted]

Who gems Str/Haste? I asumed it was basic knowledge that it is wrong? You always want to fill up sockets with secondary stats, always.

You're right, but keep in mind that you much likely have a lot of exp on your gear + many red sockets. I guess, you will often end up with half exp/haste and half str/haste gems (at least for my gear that is the case).

40%-haste-"cap":
I'm at 41% haste (561 ilvl with heroic Sha Weapon, t15 4p and Thok N WF trinket) and crit is at 0.71, haste is at 0.75, mastery at 0.76 - so it seems to be the soft cap you all talking about. But how does it look like with t16 4p? For now, if I chance a haste gem to a mastery gem, haste pulls ahead. How does that look like in better gear? I guess that we will end up not using many mastery gems...

[Deleted]

Wouldn't you not just reforge the excess expertise away into mastery or crit? Then gem expertise/x ? or just ignore socket bonus and gem straight haste, mastery or crit

[Deleted]

Yes, that obvious and I did that, but there is still not that much room for exp/haste gems. If I get the Immerseus bracers instead of the Juggernaut bracers it could be better, but you don't always have a choice.
And ignoring socket bonus is not the way to go: 2x 80str/160haste + 120str bonus (520) > 2x 320haste (480) (for my gear)

[Huntingbear_grimbatol]

You're right, but keep in mind that you much likely have a lot of exp on your gear + many red sockets. I guess, you will often end up with half exp/haste and half str/haste gems (at least for my gear that is the case).

40%-haste-"cap":
I'm at 41% haste (561 ilvl with heroic Sha Weapon, t15 4p and Thok N WF trinket) and crit is at 0.71, haste is at 0.75, mastery at 0.76 - so it seems to be the soft cap you all talking about. But how does it look like with t16 4p? For now, if I chance a haste gem to a mastery gem, haste pulls ahead. How does that look like in better gear? I guess that we will end up not using many mastery gems...

in BiS gear it's estimated that you'll use ~10 exp + haste gems, if you avoid the mace from Nazgrim and you're not human or dwarf for exp racials you can prob get more than 10 exp gems. All this means you'll open up a lot more room for mastery on your gear making thok trinket even more powerful as the trinket does nothing to strength, the "math" about this is that 160exp = 160mastery (or haste) and thok's trinket increases that 160 stat by 7-8%. So in reality the choice you have is 80str or 173 haste/mastery, not a hard choice. This is also why I'm not ever gonna want Nazgrim mace unless I'm the last one that doesn't have a heroic wep for mainspec, it's not bad by any means... you just upgrade it for any of the other 2h str weps we have this tier. If you're human or dwarf the mace is stronger because of the exp gain, but then you'll get less secondary stats which is stronger. We're only talking about ~4-500 rating difference (combined mastery/haste) here so it's hardly gamechanging.

[Deleted]

The scale factors I posted before are with Thok N WF trinket, but I gemmed exp before I got this. It's a nobrainer. And yes, the Nazgrim mace is shit.^^ We have to avoid exp where we can to make use of the ~10 exp gems.

[Deleted]

The scale factors I posted before are with Thok N WF trinket, but I gemmed exp before I got this. It's a nobrainer. And yes, the Nazgrim mace is shit.^^ We have to avoid exp where we can to make use of the ~10 exp gems.

if we must avoid the expertise of Nazgrim's 2H mace, then human's BiS weapon is Sha 2H sword, right?

[Aiya]

40%-haste-"cap":
I'm at 41% haste (561 ilvl with heroic Sha Weapon, t15 4p and Thok N WF trinket) and crit is at 0.71, haste is at 0.75, mastery at 0.76 - so it seems to be the soft cap you all talking about.

Which tool is giving you these results? It seems more like a local haste plateau.

I have quite a similiar gear setup (http://eu.battle.net/wow/en/characte...3%BDa/advanced) (heroic Sha sword, t15 4p and Thok normal wf trinket) and simcraft 540-3 with 50k iterations shows that haste is still leading - altough I'm at 42,4% haste (18,000 haste rating).

These are my normalized ratings:
Str: 1.00 AP: 0.45 Exp: 0.71 Crit: 0.73 Haste: 0.82 Mastery: 0.77 Wdps: 1.95

Maybe Galakras' trinket and t16 4pc will change this, but at this time and with this gear setup there is no valid evidence for a 40 percent breakpoint.

[Deleted]

Which tool is giving you these results?

simcraft 540-3
(I still can't post links, so search for Rúcina <Abstinentia> - Blackrock-EU to see my gear)

I guess the "breakpoint" everyone is talking about is only the point where mastery pulls ahead of haste. There is no reason why haste could be shit if you are over 40%, but you likely want to gem mastery until the scale factors are equal again. Your mastery scale factor is higher because you have more haste than me.

[Fredzilla]

Well the galakras trinket would theoretically increase the value of mastery, since mastery is the stronger stat during CDs (compared to haste). And outright saying to avoid the nazgrim mace just because it has expertise is a terrible thing to say without any clarification. No matter what other stat weights you get, weapon dps/damage is always the most important stat for us, so if its WF to your normal, or heroic to your normal WF, or w/e the case may be, if its higher ilvl, you want it. Beyond the fact that it has a huge amount of mastery anyway. You're not going to lose tons of dps just because you have to throw in some str/haste-or-mastery gems instead of exp.

[Deleted]

simcraft 540-3
(I still can't post links, so search for Rúcina <Abstinentia> - Blackrock-EU to see my gear)

I guess the "breakpoint" everyone is talking about is only the point where mastery pulls ahead of haste. There is no reason why haste could be shit if you are over 40%, but you likely want to gem mastery until the scale factors are equal again. Your mastery scale factor is higher because you have more haste than me.

You're reading it wrong. His haste is showing up as 0.82 compared to mastery at 0.77, and he questions why this is, since the general assumption is that we stop at 40% haste (and he has 42%).

[Deleted]

Nobody said that you should use a lower ilvl weapon to avoid the Nazgrim mace. For weapons is ilvl > all. You're right, I should have said this. Sorry for that.

Edit: And reading is hard. The "breakpoint" (that isn't a breakpoint) seem to be different for everyone because of racials/gear. Everyone has to simlulate his or her own gear to check if mastery is ahead and (if yes) then switch a gem. (If you do it that way, you will much likely end up with more than 40% haste)

[soamazzing]

You all are putting far to much stock into simcraft. Simcraft is a great tool but it can not be the only one used to optimize gear. The head of the ret portion of simcraft has admitted on several occasions that his main spec is prot and that he does not spend enough time on the ret module to be totally accurate. Doing you own testing with your own gear is really the only indisputable way to measure if something is better than another thing. 5 minute dps tests several times with the same buffs, reforge, armor ect (do like 10 of the 5 minute tests to get a delta) and then change things around and do it again. It's not the easy way to do it but if you want to squeeze the most out of your gear choices and gemming/reforge choices, testing on your own is key.

[Deleted]

DPS tests are not accurate, unless you can do like 1000 of them. You can test a set bonus like this, but not the scale factors of haste and mastery. But what should we do? All we have is our experience and simulation tools. If the actual dps not matches the simulation at all, you can test something else. In the end the only thing that matters is how YOU do the most dps.

[Deleted]

Who gems Str/Haste? I asumed it was basic knowledge that it is wrong? You always want to fill up sockets with secondary stats, always.

Where did all the crazies come from all of a sudden? So You think I should be gemming expertise over strength when i'm already over-capped on expertise?

[Fredzilla]

DPS tests are not accurate, unless you can do like 1000 of them. You can test a set bonus like this, but not the scale factors of haste and mastery. But what should we do? All we have is our experience and simulation tools. If the actual dps not matches the simulation at all, you can test something else. In the end the only thing that matters is how YOU do the most dps.

This, pretty much. Data you gather yourself from such tests don't matter in the slightest unless you get thousands of them, otherwise RNG will play far too big a factor for any results you get to be remotely accurate, as far as any assumptions you make off it are concerned. Small sample sizes (anything less than 4 digits, really, in this setting) don't give you the whole picture and never will.

And yeah, the only time you start gemming exp over strength is when you can then reforge to more of a better stat as opposed to reforging for exp. This depends greatly on your own stat weights and how much exp you are at, and what you'd be at by changing out gems. You wouldn't use an exp/haste gem over str/haste if like half the exp is going to be wasted.

[soamazzing]

DPS tests are not accurate, unless you can do like 1000 of them. You can test a set bonus like this, but not the scale factors of haste and mastery. But what should we do? All we have is our experience and simulation tools. If the actual dps not matches the simulation at all, you can test something else. In the end the only thing that matters is how YOU do the most dps.

Simcraft is guaranteed to be wrong regardless of the iterations if the values it uses are not totally optimized and accurate. Per Theck, he doesnt concern himself to much with the details (trinket uptimes ect) on the ret profile due to lack of time. He is open to others stepping in and helping perfect the ret module but atm it is a rough guideline at best.

[Fredzilla]

Simcraft is guaranteed to be wrong regardless of the iterations if the values it uses are not totally optimized and accurate. Per Theck, he doesnt concern himself to much with the details (trinket uptimes ect) on the ret profile due to lack of time. He is open to others stepping in and helping perfect the ret module but atm it is a rough guideline at best.

We don't really have any other viable options for stat weights though. And, even if the values within simcraft aren't right, it's still going to be less wrong on average at higher iterations than lower ones (hooray normal distributions form larger sample sizes). You could theoretically get individual tests closer to what might be right (assuming the values it's using are still wrong, which we don't necessarily know they are), but there's no way you would know that yourself. It's the best tool we have for this still just because it's capable of doing so much math with so little effort.

[soamazzing]

DPS tests are not accurate, unless you can do like 1000 of them. You can test a set bonus like this, but not the scale factors of haste and mastery. But what should we do? All we have is our experience and simulation tools. If the actual dps not matches the simulation at all, you can test something else. In the end the only thing that matters is how YOU do the most dps.

Simcraft is guaranteed to be wrong regardless of the iterations if the values it uses are not totally optimized and accurate. Per Theck, he doesnt concern himself to much with the details (trinket uptimes ect) on the ret profile due to lack of time. He is open to others stepping in and helping perfect the ret module but atm it is a rough guideline at best.

[Deleted]

Where did all the crazies come from all of a sudden? So You think I should be gemming expertise over strength when i'm already over-capped on expertise?

Equip Spark of Zand or the Timeless trinket instead of Feather, replace your gems with haste/hit? Also you have Nazgrim mace, obviously you're above cap. As soon as you drop it, you'll be fine.