Deep Subtlety: Thread and Spec Necromancy - Page 51

Monstrosity · 03/19/09, 3:51 PM

Originally Posted by Armanewb

Going back a few pages, you can see the discussion between using and not using Rupture in HAT rotations. Otherwise, its just a matter of maintaining Slice and Dice, maintaining Rupture (if you opt to include it) and hitting Eviscerate. Using a combo point builder only when you are close to capping energy and CP <= 2.

Well that is what I try to do. I use SND glyph and Evi glyph on raids, spam evi when CP's cap with occassional Rupture thrown in but I cant seem to approach the kind of numbers many say they are doing. I am not sure what I am doing wrong? My gear probably shows my pvp gear in armory but I have decent Naxx gear from 25's have SR in MH and WD in OH, usually using DP/WP. Guess I need to be quicker on the trigger dunno.

Should I be using Tricks of the Trade?

Cooljo · 03/19/09, 5:02 PM

Originally Posted by Mavanas

I need some confirmation regarding what can proc HaT and what cannot, so if you have done some testing and know the answer, please let me know please:

Mutilate - procs 2 points if both hands crit
Stormstrike - same as mutilate
Searing totem - does not proc Hat
Windfury procs - proc HaT (someone mentioned earlier they do, but if someone could indepently confirm, it would be nice)
Rune weapons - proc HaT
Mirror images - do not proc HaT

Thanks for your help.

Just confirming:

Windfury definetly proccs HaT
Searing doesn't

streaka · 03/19/09, 5:09 PM

Originally Posted by Monstrosity

Should I be using Tricks of the Trade?

If you value your raids dps > yours, then yes. Also, remember that if your target is high on the threat list, he/she probably isn't the best candidate for your TotT. 15% damage increase for 6 seconds is a nice buff.

Mavanas · 03/20/09, 11:12 AM

Originally Posted by Cooljo

Just confirming:

Windfury definetly proccs HaT
Searing doesn't

Does that mean that flametongue also procs HaT? That's kinda overpowered then. It's like poisons proccing HaT.

OrdOrh · 03/21/09, 3:59 PM

Having been a lurker here for some time, I first and foremost want to thank all the people who put in so much effort into helping the rogue community. Now for my question =p

After reading through every page of this thread (yes all 31), I've come to the conclusion that there are three ways to spend your energy before it gets capped, assuming you get an erratic CP supply:

- Hemo
- Shiv
- Backstab

My question is (and I aplogise if I missed something out in the previous pages, 31 is a lot to digest =O), what are the best situations to use each of these combo point builders, and is it absolutely vital to use them in order to prevent energy from capping out (with < 2 CP)? Also, is it ok to pop Eviscerate with 4 CP, or is it necessary to always Eviscerate with 5 CP's?

Also, I've noticed the use of the terms sHaT, nHaT and vHaT - what's all that about? I plan to spec HaT for an upcoming Naxx run - I have 3x geared Survival Hunters and 1x Enh Shammy (also geared). In this case, what would be the best HaT build to use? Thanks to all for the help.

Mavanas · 03/21/09, 4:48 PM

I have done additional analyses of HaT ppm of various classes. My previous analysis was done based on counting the number of critical attacks of abilities that are known to proc HaT. However since then I have analyzed WWS report of a survival hunter and found as much as 20% of criticals not resulting in HaT procs due to internal cooldown of the talent. HaT proc loss can result when certain abilities that proc HaT can occur in less than a global cooldown. Here are some examples:
- explosive shot ticks
- heroic strike
- arcane missiles during arcane barrage
- starfire

In this analysis, I have examined 6 Patchwerk fights of 4 guilds, each lasting 2 minutes and 7 seconds (consecutive entries in the WWS record list). The total number of players I analyzed was 142. Because combat logs do not have an indicator for HaT proc, I had to use two analytical approaches to determine which criticals resulted in a HaT proc. In designing the two methods, I took into account latency issues.

HaT count A = Criticals - Criticals from different abilities within less than 1 second. The idea here is that even if due to latency the time between two consecutive attacks of same type was less than 1 second, they will still be counted.

HaT count B = Criticals - Criticals within HaT cooldown + Criticals from simultaneous abilities. HaT cooldown period is one second in theory, but due to latency different cut off points can be used. Lowering the cut off point will increase the number of abilities off global cooldown that are wrongfully assumed to proc HaT. Increasing the cut off point will increase the number of legitimate HaT procs that are not counted due to latency. For some classes, which do not have abilties that land off global cooldown, it is safe to lower the cut off point significantly to negate the effect of latency on the combat log. For arcane mages, the cut off was lowered to capture half of arcane missiles (arcane missiles land every 0.5 seconds during arcane barrage).

In most cases, both approaches gave very close results. To calculate the loss of procs to HaT cooldown, I used HaT count B as a more reliable measure of HaT in my opinion.

Here is the format:
Class (Observations) Criticals/second - HaT Count B - HaT CD Loss
1. Survival hunter with pet (16) 0.70 - 0.54 - 23%
2. Holy paladin (6) 0.34 - 0.33 - 3%
3. Elemental shaman (6) 0.42 - 0.29 - 31%
4. Enhancement shaman (8) 0.32 - 0.32 - 0%
5. Moonkin druid (6) 0.34 - 0.28 - 19%
6. Restoration shaman (5) 0.29 - 0.29 - 0%
7. Arcane mage (3) 0.41 - 0.28 - 32%
8. Blood death knight (3) 0.27 - 0.25 - 6%
9. Fury warrior (10) 0.35 - 0.25 - 29%
10. Unholy death knight (8) 0.25 - 0.24 - 2%
11. Mutilate rogue (5) 0.26 - 0.26 - 0%
12. Combat rogue (3) 0.23 - 0.23 - 0%
13. Fire mage (20) 0.29 - 0.21 - 26%
14. HaT rogue (4) 0.22 - 0.22 - 0%
15. Affliction warlock with imp (11) 0.22 - 0.20 - 7%
16. Bear druid (6) 0.31 - 0.19 - 39%
17. Retribution paladin (7) 0.17 - 0.17 - 0%
18. Protection warrior (3) 0.28 - 0.14 - 52%
19. Holy priest (6) 0.14 - 0.14 - 3%
20. Shadow priest (6) 0.08 - 0.07 - 11%

I parsed WWS reports in Excel, so if you are interested to see the raw data, send your requests in a PM.

sp00n · 03/21/09, 10:58 PM

Originally Posted by OrdOrh

Having been a lurker here for some time, I first and foremost want to thank all the people who put in so much effort into helping the rogue community. Now for my question =p

After reading through every page of this thread (yes all 31), I've come to the conclusion that there are three ways to spend your energy before it gets capped, assuming you get an erratic CP supply:

- Hemo
- Shiv
- Backstab

My question is (and I aplogise if I missed something out in the previous pages, 31 is a lot to digest =O), what are the best situations to use each of these combo point builders, and is it absolutely vital to use them in order to prevent energy from capping out (with < 2 CP)? Also, is it ok to pop Eviscerate with 4 CP, or is it necessary to always Eviscerate with 5 CP's?

Also, I've noticed the use of the terms sHaT, nHaT and vHaT - what's all that about? I plan to spec HaT for an upcoming Naxx run - I have 3x geared Survival Hunters and 1x Enh Shammy (also geared). In this case, what would be the best HaT build to use? Thanks to all for the help.

Well, you could have read a bit more thoroughly then.

General census is to use a combo point builder if you are below 3 CP. Which one is mainly up to you and the weapons you have, but you can check the spreadsheet if you're not too sure about it. For me, equipping 2x WD and Shiv was about equal (a tad worse) than CG+WD and Hemo.

As for the abbreviations, in the order you mentioned it is solo HaT, HaT with Envenom as finisher, HaT with Eviscerate as finisher. nHaT was mostly a theoretical build, which is likely not to happen since the removal of the Envenom glyph. Alternating between Envenom and Eviscerate might howerver prove beneficial once 3.1 hits.

I'd really suggest you check out the spreadsheet Mavanas posted in this thread, you can play around with the builds and there are also the two 'main' builds included. Your group seems pretty optimal for a HaT spec, though I don't know how Mavanas new findings change the numbers.

Mavanas · 03/22/09, 12:16 AM

I am pretty satisified with my modelling of HaT procs now. I used the WWS parses to build empirical distributions of times between attacks for each class. While crit/sec rates will change as gear improves, the distributions of time between attacks should last a while. This new approach also takes into account HaT proc loss to cooldown, so dps in survival hunter groups went down a little.

The latest version can be found at Rogue_Simulation_3.1c.xls Download File on FileFront. As always any feedback is welcome.

In the gear sheet, I've loaded approximate EP values for middle HaT (2 hunters, a HaT rogue and an Unholy DK), and highlighted best items in terms of EP values.

Cooljo · 03/22/09, 6:45 AM

Originally Posted by Mavanas

HaT proc loss can result when certain abilities that proc HaT can occur in less than a global cooldown. Here are some examples:
- explosive shot ticks
- heroic strike
- arcane missiles during arcane barrage
- starfire

That's strange... I made some tests recently with a sv hunter and an ench shammy. A triple crit explosive shot yielded 3 CPs for me and a dual WF crit (same speed weapons) also gave me 2 CPs. (Though I didn't take into account the "perhaps" buggy implementation of Flurry...)
Nevertheless my conclusion was, that abilities generating 2 yellow numbers for a group member of mine at the exact same time would ignore the internal 1sec CD of HaT. That would make sense if Blizzard implemented their event system in a bucket style fashion... but thats another topic.

But I'm 100% positive, that the 3 explosive shot ticks gave me 3CPs. Ping was between 100-200ms. I'm gonna repeat these testings... All that I wanted to say, that this seems a bit fishy and although your effort is highly appreciated, it could be in vain because of Blizzard's implementation of HaT...

Herb · 03/22/09, 7:58 AM

Originally Posted by spookz

PTR:
a) BasePPM * (1+0.5 (Imp.P.)) * (1+0.75 (Env)) = BasePPM * 2.625 -> 162.5% increase

b) BasePPM * 1 + BasePPM *0.5 + BasePPM *0.75 = BasePPM * 2.25 -> 125% increase

Option b) is exactly the same as it is on Live but if they go with option a) it encourages the use of (glyphed) Envenom even more but I don't think of it as overpowered.

Works according to option b) on the PTR ("exactly the same as it is on Live") - I've written a small, specialized log parser to verify, and tested with 2x [Dirk] (IP/DP), using only Mutilate and Envenom (as 51/13/7):

totalSwings = 1024 (MH/OH = 516/508)*

totalTime = 578991, totalTime /w Envenom buff = 379863

procs/triggers = 63/181 (0.34806629834254144)
procs/triggers /w Envenom buff = 264/512 (0.515625)

"procs" are actual IP procs, "triggers" anything that could trigger one (MH swing, Mutilate, Envenom) - for a 1.6 speed weapon, 34.3% proc rate is expected outside, and either 60.0% (a) or 51.4% (b) proc rate during the Envenom buff. If sample size is too small, I can easily extend testing.

*) the OH Dirk broke

Apps · 03/22/09, 11:41 AM

The easiest simultaneous ability to test with may be fan of knives. Everything I've seen shows it gives a combo point per crit - which is less useful than it sounds, since combo points on aoe packs aren't that useful.

Mavanas · 03/22/09, 1:41 PM

Originally Posted by Cooljo

That's strange... I made some tests recently with a sv hunter and an ench shammy. A triple crit explosive shot yielded 3 CPs for me and a dual WF crit (same speed weapons) also gave me 2 CPs. (Though I didn't take into account the "perhaps" buggy implementation of Flurry...)
Nevertheless my conclusion was, that abilities generating 2 yellow numbers for a group member of mine at the exact same time would ignore the internal 1sec CD of HaT. That would make sense if Blizzard implemented their event system in a bucket style fashion... but thats another topic.

But I'm 100% positive, that the 3 explosive shot ticks gave me 3CPs. Ping was between 100-200ms. I'm gonna repeat these testings... All that I wanted to say, that this seems a bit fishy and although your effort is highly appreciated, it could be in vain because of Blizzard's implementation of HaT...

I agree with you that each explosive shot tick can in theory proc HaT. I also think simultaneous abilities like stormstrike, whirlwind, and mutilate could create 2 HaT procs simultaneously. That is not how HaT loss can occur.

Consider this example from an actual WWS report:
20:02'10.536 Walo Explosive Shot hits Patchwerk for 4814 Fire. (Critical) (199 Resisted)
20:02'11.372 Walo Explosive Shot hits Patchwerk for 4814 Fire. (Critical) (199 Resisted)
20:02'12.427 Walo Steady Shot hits Patchwerk for 2813 Physical. (Critical)
20:02'12.528 Walo Explosive Shot hits Patchwerk for 5350 Fire. (Critical)
All of these are criticals and could potentially create HaT procs, but HaT 1 second cooldown will prevent the last explosive shot tick crit from proccing HaT. The times between attacks are 1.176, 0.836, 1.055, and 0.101. The first hit is an explosive shot that leaves a dot for the next two seconds. The reason the first two intervals are not exactly one second is latency. Then after first explosive shot tick, Walo (who is no longer restricted by a global cooldown) fires a steady shot, which lands 0.1 second before the last explosive shot crit. That's how the loss occurs.

It does not mean that those classes that have "off-cooldown" abilities like that will automatically have low HaT ppm rates. The fact that the hunter leaves a dot that can crit on its own and then goes on to do other abilities probably only increases his HaT proc rate, but if two abilities crit within less than 1 second of each other that benefit will be lost. When I talk about HaT proc loss I ackowledge that for these classes you cannot simply count criticals, you need to subtract some to account for the loss.

Mavanas · 03/22/09, 1:51 PM

Originally Posted by Herb

Works according to option b) on the PTR ("exactly the same as it is on Live") - I've written a small, specialized log parser to verify, and tested with 2x [Dirk] (IP/DP), using only Mutilate and Envenom (as 51/13/7):

totalSwings = 1024 (MH/OH = 516/508)*

totalTime = 578991, totalTime /w Envenom buff = 379863

procs/triggers = 63/181 (0.34806629834254144)
procs/triggers /w Envenom buff = 264/512 (0.515625)

"procs" are actual IP procs, "triggers" anything that could trigger one (MH swing, Mutilate, Envenom) - for a 1.6 speed weapon, 34.3% proc rate is expected outside, and either 60.0% (a) or 51.4% (b) proc rate during the Envenom buff. If sample size is too small, I can easily extend testing.

*) the OH Dirk broke

Your sample size should be sufficient. The confidence interval around 0.515625 rate is (.4714, .5597), so the 60% value is outside of the interval.

madsushi · 03/22/09, 2:25 PM

I asked this in a different thread, but has anyone tested the T8 4pc bonus (Rupture crits) or Primal Gore (the feral druid talented equivalent)? It may be the same as the Warlock/Priest versions but I cannot confirm.

If Rupture crits can generate CP (on a fairly regular basis, as well), this would make Rupture a more attractive finisher as a HAT Rogue.

spacewiz · 03/23/09, 12:27 AM

Just need a bit of clarification.
CG/WD is better than SR/WD? (since I don't have a second WD)

And since I'm new to HaT I was wondering what the best 4th group member would be.
I will be raiding with 2x sur hunters and 1 enh shaman. I have a pick of pretty much everything else but a feral druid and another rogue.