Maurice2u

Worthy of note. Some of our number crunching via theorycraft has been in favor of this Hemo-Combat build, and some math has been quite against it's use. Yet, oddly enough, if we compile all of the posted actual test results, we are seeing a large amount of topping the meters and/or exceeding DPS of the standard Combat build.

IE: Are we missing something in the math? The math against this Hemo build essentially says that Combat should "blow the Hemo away" in DPS, and the math in support of the Hemo build simply says "it can be on par with Combat, situationally better, often worse without extreme gear/buffs", This would suggest the average of our in practice results should be quite a bit more skewed down (against Hemo performance).

Yet, as we go over about 10 of the WWS logs given by our friends here, we just aren't seeing that from in-game results. Even when the player is posting results vs themselves with each build on sequential raids. I feel like either we are missing something in our Energy Cost/Cycles, or the total kickback of Serrated Blades.

From all the math I've seen here, the Hemo build should be "at best" just on par with Combat or a bit below, but having a kickback to the raid in Hemo charges. Particularly in the 10mans, and with some AP being well below 3500. Is there something else I'm not seeing that is keeping our math from matching up with our logged results?

Oscarvil's post and logs above, along with previously posted logs, really highlight this disparity between what our math is saying, and how it's really performing.

 

Maurice2u

Glad hammers would likely be very poor. The specializations really help even out the DPS here and are hard to sacrifice if trying to compare it to standard Combat builds (that will almost always have the specialization). If leaving those out, it really comes directly down to the Haste proc of the Dragon"X" hammers vs the DPS of a specialization. I'm pretty sure we've math'd out the Haste proc loses, but someone else will have to confirm or deny for me please.

 

Oscarvil

Thanks for the reply Maurice. I'd just like to say my AP is usually around 2000-2500 raidbuffed depending on procs, which is about half the amount required to make SS and Hemo equivalent. Obviously they are not equivalent going by the logs but somehow serrated blades makes up for the loss in damage that CP gives.

Perhaps it's because I am at a relatively low gear level and that the difference between the 2 specs is minimal so i'm just seeing the effects of the RNG in the differing damage levels. I have taken enough stats to know that my damage on a fight like Gruul can fluctuate by 50-100dps for no good reason apart from randomness of the encounter. I'd say that Magtheridon shows more consistent dps week to week but my comparison above wasn't an accurate sustained dps number due to having to cut every 3rd or 4th cycle off in order to click a cube.

I was going to make a post about making some actual estimations of hemo vs CP theoretical dps as opposed to endless anecdotal evidence and in the meantime people did the calculations. I concur with your statement that something funny is going on if what the calculations show is true (which they appear to be).

 

Maurice2u

Here's one vantage that may close up the gap between the math and the viewed results:

Boss fights are almost never Tank&Spank.

Now before we all say "duh", let's look a little closer. Math damage has a built in presumption that we really have not accounted for. It presumes continuous use of it's damage cycles through multiple finishers to the Nth degree, so much so that CP energy returns are assumed to have a set amount of DPS given back over X-time.

Reality is, a boss fight often has:
A)Knockbacks
B)Adds that must be targetted on timer
C)Time stops
D)Mass AoE's requiring DPS pauses for avoidance
E)Teleport/Relocations
F)Other

All of these things bring melee DPS into small "extended burst" pockets, that have the added penalty of removed openers. At any point where energy is regenerated to 100% prior to regaining the main target, or combo points are lost due to targetting of something else requiring fast DPS (adds), then the CP energy return over time conversion to DPS is broken. However, the % added DPS from Serrated Blades within X-time interval is always going to favor the Hemo build. Over a finite period of time, the white damage delivered will simply be more for one w/ Serrated vs without. (for same weapons poison damage and windfury procs can be assumed to be equal between the two builds and not a factor). Haste effects also seem to scale better towards benefitting Serrated Blades vs. Combat Potency.

Now. Understanding that our DPS comes in these interval pockets, the remaining damage is a function of our yellow damage from specials, added to our finisher damage excluding SnD delivered within that finite pocket of time. With respect to SS vs Hemo. Our math would suggest that SS is almost always going to win. The amount of AP required to turn this over, combined w/ the lack of slow weapons that fall under our current best damage specialization (swords) simply puts this out of reach for most players. When everyone has a Glaive of Azeroth, maybe we'll feel different. So here, SS wins the DPS boost.

So now we come to the last of the three categories of damage: finisher input. Here that "limited time to execute" can play a large factor. What we may need now is math that illustrates the amount of time required to reach a 2 and 3 full cycles for each build. IE: how long does it take to get SnD going, execute a 5pt finisher, and then another 5pt finisher for each build using it's ideal cycle combination? My gut tells me Hemo is faster in the short term, but I'm deferring to the real mathcrafters here.

Further compounding this "limited pockets of execution time" phenom is the points wherein a full 5pt cannot be reached. Here a rupture boosted build has the benefit of dropping a 3pt or 4pt finisher that will do more damage than it's competitor anyway (30% boost) without negative effects. The effects of CP effectively wasted as both regain full Energy prior to regaining the target, and also being able to regain 5CPs faster than normal Combat in short periods of time. (as previously noted, the regain of CPs by Combat Potency depends on long-term engagement where no energy regen is lost due to being capped @ 100 or 110 energy without it being utilized).

So to sum it all up, the nature of the constant movement, engage/disengage fighting for melee DPS may have some significant effects on the standard math we sometimes use because it assumes "over the long term" without necessarily accounting for losses due to wasted energy regen. Once again, this does not prove one build "better" than the other. What it does address is potential holes in our math process that might explain the gap we are seeing in theory vs. application. It is simply one avenue of possibility as to why standard Combat builds have the highest "possible" DPS, but based on the encounters we are using them in, don't necessarily show a large dominance in actual in-game testing.

 

Faytte

Unless something has fundamentally changed in TBC the majority of boss fights have the rogue in an 'extended' situation, and even where there might be some movement, its not like the original shaz fights from MC, or running out on baron geddon (although granted now adays if someone did that they might just heal the melee through the pulses). So what are we talking about...trash/adds? I'll plead alot of ignorance to the new fights here but I cannot suspect that things are holistically that different.

Farewell, remorse: all good to me is lost; Evil, be thou my good.
~Paradise Lost (bk. IX, l. 171)

 

Melnor

I would also attribute some of PvE Hemo's succes due to human nature. If you are trying to prove a build is worth of specc'ing for dps purposes, you will unconciously/knowingly pay more attention and try to be perfect in cycle usage.

In other words since you respec'd you are most likely trying harder to make sure you get hemo to look good.

I'm not agreeing/arguing that Hemo is better than combat because I haven't spec'd it personally, I'm just trying to account for the "human factor" of the build.

 

Persona

Here are some WWS numbers from last night for a Kael/Rage/ and our first naj'etus kill using the 11/21/29 spec. Seems to be holding up pretty well.

Wow Web Stats

I know I can do more DPS with combat swords but with the added DPS from the debuff it would be really close.

 

Kezz

Or you could try to use the correct hemo build.

 

Persona

I have tried the 11/28/22 spec and the 11/21/29 spec does more DPS for me.

 

Kezz

Yeah, I see you claiming that, but nothing supports it whatsoever.

 

Persona

I honestly don't care if other use or try it. I'm just putting the information out there that I have found. It works better for me so I'll keep using it. I still plan to spec back to full combat swords though until I can stack some haste gear.

 

Kezz

I just want some information to back up your statements, maybe I would switch to your build. How is 115 AP better than Sword Specialization and WE?

 

• Shifft

You'll gain a TON more than 115 AP from 4/5 deadliness..200 at a bare minimum (2500 AP raid buffed).

 

Persona

Sword spec is good but is getting yet another small nerf when the patch comes out. WE is also a nice talent for bosses but, I happen to be a human rogue so I get +5 to weapon skill already. My hit rating is high enough that losing WE doesn't affect my damage much at all. As for the 8% AP, it also buffs your raid buffs so you get alot more than just 124 AP (144 for me). Think if it as getting almost another unleashed rage buff. AP really boosts white damage and thats what these specs are based around.

 

• Vulajin

In all the fights I've done in BC thus far, I can't think of a single one where a 17-20 second cycle can't be repeated at least once during one of those pockets of action identified by Maurice2u. On Gruul I get plenty of time between Shatters and, if I want, I can just Cloak and run back up and hit him during the Shatter anyway. On Magtheridon, I get 100% DPS time. Same on Hydross (other than during transitions). Same on Void Reaver. On Lurker it's at least, what, 90 seconds between submerges? On Karathress, you get about two minutes per target (assuming you take exactly the enrage timer to win). On Morogrim, I don't switch off.

These aren't the old days of running a 47-second 3/5/5 cycle. Whether you're 11/28/22 or 19/42/0, you're going to be running some variant of Xs/5r (see my post on the previous page about this). These cycles take around 20 seconds to run. In fact, by virtue of Combat Potency, the combat build will be running an even faster cycle than the Hemo build. If any build is capable of sustaining a cycle in a "quick burst" situation, it's the combat build, not Hemo.

 

◊ songster

While we're talking about bursty fights - what should you do if your DP stack drops, either by unlucky chance or by forced time off target? Should you wait for it to build itself back up, or does it ever make sense to use Shiv to build it?

 

HarpoonSeal

The assertion that serrated blades scales better than combat potency got me wondering. So I made a few bounding lower assumptions and tried out the math.

1. MH and OH weapons have roughly equal dps and are 2.6 and 1.5 speeds respectively

2. A sinister strike will do at least as much average dmg as a white MH attack (which has been proven in one of the other threads since a white MH attack is hemo without lethality)

3. The extra energy is used solely for a sinister strike and not for making a more efficient cycle (for instance, using ss at 5 combo points because potency procced 3 times in a row between energy tics)

I hope we can agree that these are fairly conservative lower bounds and that any number derived from these will be the absolute minimum bound.

So, here's the math:

CP averaged out gives 40 energy every 13.33 offhand attacks

During this time, you will get 7.69 mainhand hits (13.33*1.5/2.6)

With two weapons of the same dps, then your offhand attacks can be converted into "mainhand equilant attacks" equal to 75% of your mainhand attacks (If this is confusing, think about a 1.4 speed and a 2.8 speed weapon set. Every 2.8 seconds your offhand will have done dmg equal to 75% of a mainhand swing)

Thus, you get a free sinister strike for every 13.46 mainland equivalent strikes (7.69 * 1.75)

This translates to a dps gain of 7.4% (if you assume a sinister strike does the same amount as a normal white hit. This is the absolute lowest bonus possible, as SS has been proven to do more damage at currently achievable ap levels.

Using actual numbers that I grabbed from vladia's anatheron attempts: a 2.7 speed MH gave 696 average dmg on an ss and 598 average on a hemo. Redoing math for 2.7 speed gives 12.96 mainhand equivalent attacks per free SS. Each free SS is worth 1.16 white attacks. Thus the gain is 8.95%.

My conclusion is that CP scales much better than serrated blades, especially since this analysis can be directly applied to haste scaling (every 13.33 extra attacks granted though haste would be a free SS).

 

McTurok

I would be interested to see how Syphon of the Nathrezim, 2.80 speed and Swiftsteel Bludgeon, 1.50 sped with Improves haste rating by 27 would do as hemo.

Sure I understand swords get the proc but let's leave that out for now.

Heck, how about that mace from kael'thas that gives 10 energy.

EDIT: to add, maybe these weapons could rival sword spec in end-game pve just because of their haste and procs?

 

Maurice2u

The one thing that stands out to me the most from your log is how you have over 100DPS more than 2 people who put out a higher total contribution. This is not very good data for the sheer fact that out of the 20 highest total damage contributers in your log, you have the 2nd lowest %DPS time. Meaning, you simply were not contributing to DPS for a VERY large portion of the time compared to some of the other DPS on this run.

Simply raising your activity time to 30% (looks like) would have easily set you as 2nd or even 1st on the totals. Were you being pushed out of continuous DPS a lot, took a afk break, had a disconnect, etc?

---------------------

On another note, I totally object to the idea that "people with X build are trying harder, so their numbers look better". The implication being that people without it are somehow just casually sucking their way through their run is a bit unrealistic, and potentially insulting.

--------------------

As far as the disjointed pockets of DPS discussion, it is simply a thing to look at as an input to the raid damage done over a full raid. I never concluded it to be the end-all-be-all, or even a reality. I said it is a possibility. What we are really looking for here is to find out why our "math" is giving one expectation, but our logged results are lining up particularly well with that expectation. And then to find how to close the gap between the two, and understand what's going on, not just go ..... "uh-uh, my math says this does more, so screw what data your runs are showing". That's not useful.


Harpoon, you seem to be a good theorycrafter, so what would "you" contribute to the disparity. If CP does scale better, and our AP amounts are not reaching the amount where Hemo individual strikes average higher damage than Sinister Strike, and the DPS contributed back to the raid is only about 83DPS, and a large amount of targets are bleed immune (negating Rupture benefit), what is making this build perform so well in the logs we are seeing?

All those things being true, this build "should" basically be sucking. Somewhere in the 6-10th spot in damage, even for a well geared and well played rogue because the build has many losses, and no advantages (with those math theories holding true). Well, since that's not holding up on testing ...... SOMETHING is being overlooked. What is it?

 

sp00n

To everybody considering choosing the 11/21/29 spec, be aware that according to the current theory, 5 weapon skill decreases your chance to miss against a boss level mob by 3%.

So unless you are a human, or have enough weapon skill rating, or have enough hit rating (363 to never miss without any skill), you should really consider taking at least 1 point in WEx.

 

Trazhenko

I don't think you can completely ignore the human element.

I don't know about trying harder with one build vs. another, but I know that sometimes I try harder when I'm feeling pressure for my #1 spot, but that's typically an issue that only arises on trash clears, where maybe I'm (lazily) not using all my cooldown abilities every time they are up.

On bosses I tend to go hard no matter what, but even there you run into some issues: "Do I use haste pots on (insert farm content boss name here) this time? Do I use Tea? Do I take a little extra damage here and eat my healthstone for more time on target? Do I bother to sprint from this add to that one?"

Ideally, you'd like to think that you play pretty much the same all the time, but that's simply not the case. I don't think that he suggested people were "sucking their way through their run", but there is a tangible difference between merely "playing well" and "doing everything you can". It's not that outrageous to think that someone excited about a new build would be closer to "doing everything they can" than they were last week running the same build they've had for 6 months, easily topping meters, and scoffing at the competition.

One week I went to SSC after doing some farming, and didn't realize that I still had SS (untalented, lol) on my backstab key. Sometimes when I feel lazy I farm that way. I proceeded to get #1 damage on the trash clear to hydross, and on the hydross kill, though by a very narrow margin, and I was trying like crazy because I couldn't figure out how people were keeping so close, especially since my CP generation was ON FIRE. I feel like an idiot playing that long without realizing how terrible it was that my "backstab" crits weren't even breaking 1K. I was definitely putting more effort into that trash clear and boss kill than I was the rest of the night after I fixed the button.

 

Maurice2u

Yeah, it's been noted that part of the benefit of 11/28/22 is that it has "so much" of the standard combat build's same talents/benefits. CP and SA are the tradeoffs.

Disregarding WE and/or a weapon specialization will significantly hamper a Hemo build. Kind of the whole point is of this discussion is to find the gear/buff level that minimizes (or removes) the gap between 20/41 and a Hemo variant with the exact same gear/buffs. The closer the core talents of the two are together, the better.

This is not only true for keeping the DPS high, but also reduces the number of variables in figuring out what are the key contributions and detractors. In the Combat tree we have a few generally accepted concepts. Keeping the variation down to basically (Serrated Blades + Hemo energy cost + Hemo charges DPS back to raid) vs. (Combat Potency + Aggression/Vitality + SA) will help us eventually narrow down exactly what's going on between the two.

 

Maurice2u

I understand your argument. I am taking for granted that people posting their results and discussing, especially those posting results using both builds, would not be popping tons of different consumables and using better weapons and having a totally different buff input to skew the data. Taking for granted people are actually trying to learn about how the game mechanics are working, and doing as best they can to make equal ground statistical data. If we are gonna just say "people are popping Relentless Assault Flasks, haste pots, and Thistle every chance they get for creating these logs and not using them during combat spec'd", then we might as well just close the thread and say we can't count on any data acquisition to do us any good here.

I'm going to fall on the side of belief guys here are really trying to figure something out, not just make x-build or y-build (or themselves) look better. I'll grant someone may be a touch more focused while doing runs for testing, but I cannot see that accounting for the disparity between our math, and their in-game results. It would be a "much" smaller contributer.

 

Oscarvil

I can only comment for myself but I know that I don't want to show that Hemo is better than Combat, I want to make a fair comparison between the two builds. I don't even compare my dps between fights whilst I'm playing, I just slurp it into the combat log and parse it then do some post-analysis. To me the difference in the two builds is minimal in game because for both of them I'm just concentrating on spamming my direct damage ability (ss/hemo) whenever I can and deciding on how I should manage combo points and cooldowns (vanish/bf/ar) in order to put out max dps while not pulling aggro.

Some of the factor of being human shows in the way that in my experience dps on the same fight with the same gear and spec varies by 50-100dps as I stated earlier.

Maurice I think you lose a lot of information from Persona's log when you just look at the overall dps/contribution. Clicking on both the Kael and the Naj'entus logs individually give a much better comparison of Persona to other players in his raid. Who cares about trash?

Also I don't agree with your thoughts that it's the fight style that seems to be giving hemo the edge because of the fact that yellow damage contribution is so low compared to SS. Even with 10 bonus hemos from having energy refill to full a bunch of times you're never gonna catch the damage SS is putting out. Also, as Vulajin pointed out the definition of a sustained dps fight seems to have changed dramatically since 1.0 where long cycles were commonplace, and sustained now means 35-40 seconds on target as opposed to 1.5-2mins.

In addition to these reasons, my VR log shows my hemo build beating my SS build where I never ran out once, even though I lose the bonus from imp Rupture and was wearing a resist trinket over my poison vial (which put out ~1% damage on the previous log). I would have expected from all the math my damage on this fight to definitely be lower. (Random dps fluctuations excepted).

What we're missing doesn't seem to be the bonus we get from refilling our energy bar.

 

Maurice2u

Agreed. That discussion was simply to say "we need to look for other angles". I even noted in the comment that it was not necessarily enough to make a difference, or even true, for that matter.

I'm looking back at Persona's log after this. Ultimately, we're saying the same thing though. What's posted doesn't provide an adequate sampling of information.

What I'm looking for, simply, is some math that actually is "in tune" with the in-game results. Particularly like your Void fight, the disparity just doesn't add up.