Sinister Primal Diamond Meta Bug.

Bug Report
A lot of Elemental shamans are reporting around a 27% uptime for the proc on the meta. According to the adjustments made for each spec/class elemental shamans should be seeing a uptime around 47%. I can link World of Logs reports later giving the uptime if that would be helpful.
Warlocks are seeing a 35% uptime with it. Elemental Shamans have the highest coefficient modifier of all classes which should be giving us a 47% uptime.

Sinister Primal Diamond – 1.18 base RealPPM on damage/absorb of harmful or periodic spell. No ICD.
- That base proc rate is multiplied by an additional coefficient:
0.761 Arcane
0.302 Fire Mage
1.387 Frost Mage
0.625 Affliction
0.598 Demonology
0.509 Destruction
1.891 Elemental
1.872 Moonkin
0.933 Shadow
obv I know the coefficents. But you fail to realize why the coefficients are there. classes deal damage different ways. and the coefficents are to equalize the dps output of the meta per spec while also based on how fast/slow you cast. a class like ele shaman that do not scale badly off haste would not expect a 45% uptime. The reason the coefficent is so high is you have 1 dot and LB is a longish cast compared to some classes with 2+dots and a main cast spell here is our logs with a warlock and a moonkin on heroic meagara which is a good patchwork style fight even for some classes should be higher uptime than usual due to multidotting. and remember Moonkin has nearly the same coefficent as ele shaman.


Except Blizzard has explicitly stated that real ppm is not based on the method you just described.

"--Regardless of how you’re attacking or healing, slow or fast, with DoTs or direct heals, whatever, you can expect to get the same proc frequency, on average."

Taken from:

http://blue.mmo-champion.com/topic/240336-windsong-and-elemental-force-information/

The original poster is correct, the legendary meta gems are not giving the expected uptimes based on the reported rppm values.
I'm seeing 18% uptime across the board. Doesn't matter single target or AoE.

http://www.worldoflogs.com/reports/rt-qo8dmv73nk8pz2yq/details/138/?s=6789&e=7390
Lock up time for meta on heroic Ji-kun Tempus Repit 15 259.4 51.7 %

My uptime for meta on heroic Ji-kun Tempus Repit 12 86.8 17.3 %

http://www.worldoflogs.com/reports/rt-4x47z1dvputpfhwv/details/4/?s=4824&e=5325
Heroic Horridon - http://worldoflogs.com/reports/jd614m3q3dmj23mq/dashboard/?s=604&e=1181

Elemental Shaman - 27.6%
Boomkin - 22.4%
Shadow Priest - 22.8%
Destruction Warlock - 22%

Heroic Megaera - http://worldoflogs.com/reports/rt-4wjgpy2qrosc628a/?s=4366&e=4821

Elemental Shaman - 23.6%
Boomkin - 25.9%
Shadow Priest - 38.5%
Destruction Warlock - 21.5%

Heroic Durumu - http://worldoflogs.com/reports/rt-4wjgpy2qrosc628a/dashboard/?s=9072&e=9522

Elemental Shaman - 34.7%
Boomkin - 18.2%
Destruction Warlock - 25.4%
Reports from another player

Tuesday 16.4.2013, Durumu and Iron Qon on HC, rest on normal.

Durumu: 24.8%
Primordius: 38.6%
Dark Animus: 10.9%
http://worldoflogs.com/reports/g1uk3xq7be11mz6a/details/1/?s=8181&e=8741

Iron Qon: 25.7%
Twin Consorts: 20.7%
Lei Shen: 24.5%

http://worldoflogs.com/reports/h1unuvxcg8cm9iub/details/4/?s=4494&e=5068

All 5 fights done on HC, wednesday 17.4.2013.

Jin'rokh: 17.7%
Tortos: 25.9%
Council: 27.8%
Megaera: 35.2%
Ji'kun: 14.4%

http://worldoflogs.com/reports/rt-setozdfm6xjbcj6j/details/7/?s=466&e=700
Just want to bump.

Our Moonkin, over the entire night, saw a Tempus Repit uptime of 27.4%, when their expected uptime is 48%.

http://worldoflogs.com/reports/rt-m0hcwlcora9eir6n/details/2/?enc=wipes&boss=69078

While I don't particularly understand the inner workings of RPPM programming, the mass of numbers in the 25-32% range for every spec may suggest that the coefficients per spec are not being properly applied, and only haste scaling is working properly.

obv I know the coefficents. But you fail to realize why the coefficients are there. classes deal damage different ways. and the coefficents are to equalize the dps output of the meta per spec while also based on how fast/slow you cast. a class like ele shaman that do not scale badly off haste would not expect a 45% uptime. The reason the coefficent is so high is you have 1 dot and LB is a longish cast compared to some classes with 2+dots and a main cast spell here is our logs with a warlock and a moonkin on heroic meagara which is a good patchwork style fight even for some classes should be higher uptime than usual due to multidotting. and remember Moonkin has nearly the same coefficent as ele shaman.
http://worldoflogs.com/reports/rt-u05uaqc500cm8w8q/details/4/?s=12189&e=12667

Heroic meagara
our moonkin 24.6% uptime
our destro warlock 28.5% uptime


Considering the moonkin has 3x the coefficient of a destro warlock, this result is not expected.
I just want to reference another thread on MMO-Champ from the Mage forums.

http://www.mmo-champion.com/threads/1283425-Sinister-Primal-Diamond-for-FIRE/page2

Fire mages are reporting ~25%-35% uptime, which once again is expected of a 1.000 coefficient on the metagem, and not expected of a 0.302 coefficient as was reported.

This bug is killing Moonkin, Elemental and Frost Mage (the specs with >1 coefficient), and massively buffing specs with <1 coefficient.
Is it even worth using the legendary gem until this is fixed?
Catparty it should still be an upgrade over Burning, but a pretty minimal one for a class like Elemental that gets GCD capped easy. Depends on your ilvl.

Can we please just get confirmation that either
a) uptime is supposed to be roughly the same for all classes, in which case everything is fine
or
b) current uptime is not intended and will be fixed
Can we please just get confirmation that either
a) uptime is supposed to be roughly the same for all classes, in which case everything is fine
or
b) current uptime is not intended and will be fixed


This would be nice
My boomkin is also experiencing the same issue.

Throughout all of the LFRs I've logged this week (two full clears) and our main heroic raid, I've seen uptimes generally in the area of 20-25%, when it should be 40-44% as purported by WrathCalcs. I'm sure SimCraft has a slightly different number, but the point still stands: it's way too low compared to the simulations.

If I remove the 1.872 boomkin modifier from WrathCalcs, expected uptime drops to 23%, which is exactly what I'm seeing.

I know this is a small sample size, but it does agree with what everyone else is saying here.

Links to the logs I made public:

http://worldoflogs.com/reports/y9vo5ss4gei9l40h/details/22/
http://worldoflogs.com/reports/4iu68eigtsfi1an8/details/7/
http://worldoflogs.com/reports/rt-skz1kqf8ivaac12s/details/3/
http://worldoflogs.com/reports/a5ulvh02ywuyhkw4/details/2/
http://worldoflogs.com/reports/1slvwc0v56a9fvo6/details/0/ (dummy test)
http://worldoflogs.com/reports/rt-vjo83qo3gagrelvq/details/4/
If I remove the 1.872 boomkin modifier from WrathCalcs, expected uptime drops to 23%, which is exactly what I'm seeing.


Sounds like someone forgot to enable the coefficients for certain specs on the server side, if this is what you're seeing. Not too surprised given the trouble they've had with the metas since they became available.
So, who wants to talk about the random number generator in this thread?
I would. Could you please explain to me how much RNG is needed in order for hundreds of elemental parses to be around 25-30% when the expected rate is 48%?
More logs for the fire:

http://www.worldoflogs.com/reports/bhlyw34liz6wm7o5/details/0/
http://www.worldoflogs.com/reports/bco35lyj2zeeyeuj/details/0/
http://www.worldoflogs.com/reports/vp7ak6bi8rfio8sz/details/0/

These are 20-minute long controlled dummy tests, where haste is static and only one Wrath is cast per 10 seconds. This is for ease of replication and simplicity. PPM is then calculated by PPM=(observed proc chance)*6/(1+haste%). For the three tests, PPM was 1.13, 1.23 and 1.14, respectively. For boomkins, without the modifier PPM should be 1.18, and with the modifier it should be 1.18*1.872=2.21.

You can see the three tests are hovering around the 1.18 mark instead of 2.21, which implies the boomkin modifier is not in effect. Obviously this will require a larger sample size, but for a controlled test this is a significant result.

Will update when I do more repetitions, trying to get other boomkins to do the same. I suppose Ele shamans should be able to do the same with a Lightning Bolt every 10 seconds.
Just wanted to add i too have been averaging ~25% uptime on all fights. Ele is suppose to have the ~47% uptime because this massive amount of haste is pretty lackluster to our spec compared to other specs. Seeing my lava burst having a cast speed of .5 seconds is just silly.

Watching the meta proc at the pull of every boss, with lust, and literally netting 0 dps gain from the meta proc during ascendance, again, just silly. No other caster class deals with with being GCD capped like an ele. With our current place on non-cleave fights, this meta actually puts us further and further behind the other specs already doing really well.

Please fix the coefficients.

I actually thought my low uptime was more so because there is no ICD on the meta. I regularly get back to back procs. 3 seconds into the proc, booom, another proc. This can definitely lead to lower uptimes. Classes with lower proc rates wouldnt experience the over lapping proc rates nearly as often, because of the lesser proc rate chance.

So the modifiers may be correct, but without an ICD on the meta proc, its hurting ele / boomkin.

Join the Conversation

Return to Forum