I ran some uptime tests this morning on berserker. A single 15 minute test ended at 44.71% uptime. I need to run a much longer test to get better results...but it is looking like a little better uptime than I had anticipated.

Death's Bite is the superior OH. Armageddon isn't BAD necessarily, the difference in crit rating on the two is pretty significant, but depending on your hit rating haste will be more of a dps increase.

But I'd say just to be safe, enchant the Death's Bite.


Was this with just berserking on your OH? If so, I'd be interested in seeing some WWS or Recount or something of the tests.

This was berserking on my main hand, I was running a simple addon that keeps track of uptime for a specific buff (The buff name is actually Berserk).

I am definitely going to wws my next test. This was just quick and dirty.

I know a lot of the mechanics of this enchant have already been disucssed, but I want to see for myself and I thought it would be helpful to post the results. I plan on running several tests:

naked with mh berserking (no oh)...30 mins. stratight dps...no abilities used
naked with oh berserking (no mh)...30 mins...straight dps...no abilities used

hit soft capped and expertise soft capped mh berserking (oh massacre)....30 mins....straight dps...no abilities used
hit soft capped and expertise soft capped oh berserking (mh massacre)...30 mins...straight dps...no abilities used

hit soft capped and expertise soft capped mh berserking (oh massacre)...1 hour....(using BT and Slam when instant)

I have been trying to decide on trinket options. My current choices to decide between are [Meteorite Whetstone],[Bandit's Insignia], & [Mirror of Truth]. I recently acquired the Bandit's Insignia and switched the whetstone out to use it. My DPS dropped however amd I am not sure if it was due to the trinket choice or other gear changes.

I would go with the Mirror because of the crit and better on hit proc. I would recommend using Landsoul's Spreadsheet available here. You can enter all your gear and then switch out just the trinkets and see which raises your dps the most. This is probably the best tool out there for quick, accurate answers and is extremely easy to use.

If you're really talking about ideals I would think the best situation is just 2xBoH, as it isn't unique. But assuming you're trying not to be too greedy, I think it mainly depends on your hit rating. A lot of DB's item points are going into hit rating, and if you are much over cap the stat isn't necessarily as valuable as what armageddon offers.

Thanks for the input. I think I will hold off another few days in hopes Blizzard might release some more info on the upcoming haste changes, and after that I will just enchant the DB. I can always attempt to afford another Massacre enchant if I need to switch.

I am a little hesitant about not using a DB because of what another poster mentioned with the HR loss. At the moment I just fell below 361 (currently 339 hit rating) by using t7 heroes legplates. I have almost nothing socketed/enchanted for hit at the moment (except my boots but I dont really consider 1 gem to be gemming, when it is there to fill my meta req). I know that 361 isnt truly the cap anymore, and it is slightly lower, and will be even a little lower with the patch. I am really trying to aim for getting t7.5 Helm/Shoulders soon, and If I do that I am out 94 Hit rating. To recover from this I was thinking about getting the [Bladed Steelboots], and maybe using the [Sphere of Red Dragon's Blood]instead of the [Incisor Fragment] (or preferably [Grim Toll])

My armory should be fairly accurate, what are some thoughts about the gear swapping I mentioned, and if it brings me down too low in hit (mid-high 200's) would that just be hurting my dps, even though AP, Crit, Armor Pen would all be going up?

Fuzzy or napkin math need not apply here to making bloodsurge calculations. There are way too many different possibilities per player and too many assumptions people make. When you make assumptions, you overlook everything else. Please stop making them.

The big thing: Cooldown overlaps. We talk here about damage loss. There is no fight in WoW that lasts an infinite amount of time. If fights were infinite, then your DPS could be calculated by using a periodic rotation, and your gems and gear could be optimized based on that calculation. Fights in WoW are not infinite howerver, and for better results, you cannot base those calculations on that assumption. To do it differently, we must make calculations for rotational DPS based on moments. Theoretically, if the fight was finite in length, a different ability priority and decision making process could be formulated that could possibly be better. The only way to be absolutely certain is to find a means to calculate both ways and then compare them. I believe I did a pretty good job at formulating an infinite fight length with a best periodic rotation, prioritizing whirlwind and only losing one second on a bloodthirst every 16 seconds while weaving in your slam procs. However now it is time to figure out the best decisional DPS based on a finite fight length and see if it could possibly be better. Maybe hitting slam over whirlwind or bloodthirst or in any different feasable combination could be better. Maybe someone somewhere can prove this and then people could practice it instead of randomly hitting abilities based on the new Slam! proc.

The problem is, that for me to prove that I can do this, the fight must be finite and there must be a moment where you are losing some rotational DPS by prioritizing one over the other. Based on the decisions that you did make, there will always be a damage loss that occurs after the end of the fight. These are the abilities that you could have used but couldn't because you prioritized one over the other, thusly preventing damage that you could have done at the end of the fight. This damage was pushed back so far that the end of the fight happened (we can assume that the end of the fight happens at 20% because your rotation changes there) before you could use the abilites.

The theory is, Whirlwind and Bloodthirst have cooldowns. Take for example, whirlwind. The next whirlwinds can only be hit in periods after when you decide to hit the whirlwind before. Baseline, whirlwind does damage at its total strength every 8 seconds. If you delay that by hitting one ability over another based on a decision, over the spam of the entire fight, there are going to be a certain amount of whirlwinds that will get pushed off the fight length that you could not make. We call this damage loss. Take for example, a 160 second fight. Normally you could hit 20 whirlwinds during this fight. However, if you prioritize using your Slam! before your whirlwind, in fear that the next whirlwind will overwrite it, 20 times and delay each whirlwind by 1.5 seconds, then you will lose a specific amount of whirlwinds that you could have done if you didn't prioritize Slam! this way. Instead of one whirlwind every 8 seconds, you have now all 20 whirlwinds happenning every 9.5 seconds. This means you can only fit 160 / 9.5 = 16.842105 whirlwinds. Since you can't have part of a whirlwind, and we're not assuming infinite averages at this time, you have to drop off the decimal. Therefore, you are only using 16 whirlwinds instead of 20, for a 4 whirlwind loss. This is the cost, or damage loss for hitting Slam! at the same time your whirlwind cooldown comes up. This idea is not so bad. You are trading the power of 4 whirlwinds with the power of 20 Slams. I for one would take the 20 slams, but the 4 whirlwinds are completely unrecoverable. This example is highly improbable because the Slam! proc has a chance to occur, and also it can only be once for 5 seconds after you get the proc.

The main problem is that we are trying to figure out if letting a Slam! proc expire or become overlapped is less penalizing than pushing off whirlwinds and bloodthirsts. The problem becomes more complicated the shorter the fight is because the results become unreliable and too different between fight sizes. However, we need a long enough fight length to be able to make these decisions on a reliable basis. This then creates a complete paradox. We get more reliable results the longer the fight length, which suggests infinity, but we can only do the calculations on the assumptions the fight will end, so we can subtract the abilities that are pushed off. This is still completely possible with the power of abstract thinking. Okay, imagine a fight that was infinitely long, but had an end. Imagine cross sections of happennings called moments. Some moments can be different than other moments in that within each moment a certain decision is made. There can only be a finite number of similar moments. There are only 3 abilities that we are dealing with here and only so many ways to arrange them. We have to take each moment and analyze the damage losses of hitting one ability over another by pushing the corresponding delayed abilities off the end of the fight. Based off of those analyses, only some are obviously relevant to the problem at hand. This is the kind of thinking you have to do to make any remotely accurate claim regarding bloodsurge now.

A problem I ask of you all or whoever wants to run some possibilities and present their analysis on any relevant rotational moment:

Use:
Whirlwind @ 8000
Bloodthirst @ 6300
Slam @ 5400

**numbers are calculated based on best in slot gear and best possible buffs, using crits and deep wounds damage, rounded up to the nearest 100

Im using 4/5 UW at the mo, its pretty good rage over time (Kel'thuzad : Glutfive); yeah imp zerk rage too!

Does anyone happen to have any numerical simulation of the rage/second provided by UW assuming a normal ww/bt/slam/hs rotation with 3.6 speed weapons?

Being a dwarf using a mainhand mace, I am finding I trivially get over the expertise cap and have already been able to drop 1 point in weapon mastery, with the 2nd point dropping in the future with the Sarth+2 hat. I am however at a loss of where to put the spare points.

I tried out 1 point in Enrage last Naxx and found it quite lackluster. Ignoring trash, Enrage was recorded only 13 times over 2930 seconds of boss combat in WWS (I realize that this does not count Enrage refreshing itself) - giving only a 5% uptime of a 2% damage increase per talent point. This leaves either UW or perhaps imp demo shout as the options for the spare points.

Thoughts?

landsoul:

In my experience on live, there are very few moments when you have to make a decision between delaying whirlwind or losing the slam proc.

The majority of the time you can choose to not delay the whirlwind and still slam afterwards.

Only if whirlwind is lighting up within the last 1.5 seconds of the slam proc do you have to make a choice and usually you would have already used the slam.

In other words, unless you're global cooldown locked by for an external reason you never have to make this choice.

5 pts in UW grants 15 PPM on white attacks and heroics only for each hand.

This averages to 30 rage per minute modified by your differring miss chances, and your use of heroic strike, and is increased by haste effects proportionally.

I want to start off by linking the spreadsheet I use for my math, it is a beta one but fairly accurate and gives me the information I need. I made it myself, so would enjoy any feedback. It is not done and I actually have done the formulas for the DPS gained for each talent point just have not implemented it in this version yet.

MEGAUPLOAD - The leading online storage and file delivery service

Now I have some questions that I can't seem to support other than what I see in game,

I feel that before I can focus on any other stats I require
-15% hit (12% hit + 3% with precision)
-35% crit

This does not seem to give me a huge DPS increase however it gives me much more consistency. I know our current issue is largely not enough rage like we had at the end of BC, and these numbers seem to allow for the most heroic strikes.

I have a little over 3,800ap and 34% crit and 14.5% hit (with precision).
I range from 3,500dps (10 mans) to 6-7,000dps (25 man loatheb)

With these numbers you can then stack STR/AP until over 7,000 without crit or any other stat out weighing strength. (This is assuming expertise capped as well)

Now I know the lower the hit % the better, because once we have enough rage to heroic strike, hit only affects offhand swings and the points used on hit % can then be used on STR etc. I just don't feel my offhand does enough until I reach these base numbers.


So my real questions,
-15% hit to much? If so what about rage and heroic strikes, are you getting enough rage often enough without the hit?
-35% crit to much? If so, how is your flurry uptime and does lower crit effect your deep wounds?

Also one last point,
Attack Power - How much is to little to be stacking other stats? Is stacking 35% crit with only 3,800AP detrimental?
Or is 3,800ap a decent amount for now.

I have a whole bunch of math that fights for my arguments, but I just am not so confident that I should be ignore haste, armor pen etc as much as I am.

Kamacho, the issue with Slam procs is the fact that in the next patch, if you do not use a Slam proc before using BT/WW, you risk overwriting your current proc with a new proc, and what they want to calculate is whether the DPS lost from overwriting Slam procs outweighs the DPS loss from delaying the BT/WW cycle.

I know the hit cap for 2 handed weapons buti wonder does that change if your dual wielding them. Or is it capped at the same lvl as if your wielding 1 2 handed weapon

I'm still not sitting at an "infinite rage" setting where I can always Heroic 1-2x in between my current instants counting random Bloodsurge procs. Until that point happens, after the change I'll simply follow this priority;

WW
Slam, else BT
Slam, else Heroic if above ~60 rage
Slam, else Heroic if above ~60 rage (depending on if Heroism/Speed potions are active)
Slam

I don't look at it as how many Bloodthirsts or Whirlwinds will you lose total if you keep pushing them back. The fights are too short for that right now and I look at the actual damage per second of these abilities. If I'm losing 1.5 seconds on a Bloodthirst that would hit for 3000, that's a loss of 140 DPS but if I'm landing a Slam that hit for 2300 I've still gained 1390 DPS within that 1.5 second loss (Yes, I know that's an oversimplification). When they implement longer fights, or more fights that require movment then maybe this wouldn't be the best course of action, but until they do I still think the fights are too short to have to worry about whether or not you should hold off on Slamming simply because it's shaving into the GCD of your other instants. The only thing I would question right now is on multi mob fights and whether or not a Slam will eat into your WW. In that case then I would prioritize WW over the Slam for sure.

Note: 3 meetings interrupted the writing of this post so it may possibly be complete and utter trash.

Look at the problem from a different direction. Instead of looking at the DPS lost look at the maximum possible DPS gained if we are perfect and see if it's worth trying to attain it.


One part of this equation is pretty much static - the cost of delaying WW or BT by one GCD.

At 6300 dmg BT does 1,575 damage per 1.5s. At 8000 WW does 1,333 damage every 1.5s.* That's the cost of delaying them. *WW is 1,333 not 1,500 becuase it already can't be used optimally - it can only be used once every 9 seconds even though it's cooldown is 8 seconds. So clearly delaying BT is much more costly than delaying WW.


Best case scenario (aka rage unlimited) over the course of a 10 minute (600 sec) fight and assuming a single target you are going to have

99 BTs
148 WW *hits* (74 total WWs)
218 HS's (using a swing speed of 2.75)

So in an ideal 10 minute fight you have a maximum of 465 opportunities to proc Bloodsurge. If we assume an average crit rate of 50% (this may or may not be a representative number to use) then any given BT, WW or HS has a 10% chance to proc Bloodsurge. The chance of getting back to back 10% procs are 1/100.

That means that in a 10 minute fight you are, on average, only going to have 4.65 actual back to back Bloodsurge procs.

If we assume that we anticipate them perfectly and only delay the cycle for just those 4.65 procs then we (using WW since it's the least costly to delay) at best lose 1,333*4.65 or ~6,200 total damage and gain 4.65 slams (per Landsouls numbers) = 4.23*5400 = 25,110 slam damage for a net gain of 18,910 dmg or 31.52 DPS in a 10 minute fight

At 5,000 DPS on average, 31.52 DPS = 0.63% DPS increase.

So we have the potential to gain .63% DPS if we perfectly utilize back to back bloodsurge procs and if the stars align perfectly and it's always WW we delay. Seems pretty much a dead end to me given the chances of being perfect here are probably something less than 0.0000001%. Continuing on anyway...

Now if we look at it this way:

The potential damage gained/cost of delaying WW = 18,910/1.333 = 14.33.

That means if you are wrong more than 14.33 time over the course of 10 minutes you've lost DPS.

The next question is how many times per minutes would we actually have to make a decision

Per minute on average we are going to get in 21.8 HS, 10 BT and 6.67 WW or roughly 38.5 attacks per minute that can proc BS.

Now looking at a little over a 1 minute span of the cycle you can see that:

BT is going to follow HS or WW by 1 GCD roughly 7 times
WW is going to follow HS or BT by 1 GCD roughly 6 times

Edit: Skip the next 3 paragraphs and go to the next edit mark

So roughly 13 times per minute (~130 times per 10 minutes) you would be faced with the decision to delay or not. Given that the chance for a back to back proc is only 1% (given a 50% average crit rate on melee attacks) that means, on average, only 1.3 of these 130 decisions per 10 minutes will actually be a back to back proc.

You basically have to guess with a 99% accuracy to get the maximum possible gain.
If you delay every time you lose as much as 7*1,575 Bloodthirst Damage and 6*1,333 Whirlwind damage or a total of 19,023 damage per minute. Less the amount you gain from the extra slam you get in on dbl procs.

I just don't see how the back to back proc scenario is remotely worth worrying about.

EDIT: See the meeting note at the top

Ok so 13 decisions per minute with a 20% chance to proc BS. That means we have 2.6 decision per minute or 26 decision per 10 mintues - on average. Given the 1% chance at a back to back proc and that we know that 14 wrong guesses is the breaking even point it still doesn't look like this is remotely worth worrying about.

Oh! Thank you for clearing that up, that makes sense. Ouch.


Edit: Meant to edit the lower statements into this one, was grabbing a quote and spaced out.

Ok, so you can lose a technical 317 dps (based on your 19,023) assuming optimally bad conditions. But as you stated there will be around 1% chance for this to happen, on 26 chances per minute.

So our only fear is the offchance, which it will happen and probably fairly often, that we get that 1% double or triple proc (using both WW and BT before Slam!)

So I agree, we will face double and triple procs, however I feel based on your numbers we will have so little control over it that we just cant focus on and to prioritize it last for the most part.

Is that more or less correct Dark?

Basically yes.

It really boils down to just what your chance to proc Bloodsurge is. It's based on your melee crit percentage. I used 50% which may or may not be the best number to use but it's probably close enough and it has the advantage of being nice and easy to work with.

50% of your Heroic Strikes, Bloodthirst and Whirlwinds will crit. 20% of those crits will proc Bloodsurge. That means any given HS, BT or WW has a 10% chance to proc Bloodsurge.

The probability of back to back occurrence of something that has a 10% chance to happen is 1%. The chance of back to back to back is .1%.

You can't know you are going to get back to back procs. You can wait to see if you do but the chance you are actually going to see one is only like 1-2%. And each time you hold BT or WW it's costing you damage and that damage loss adds up quickly.

Now it's certainly worth making a special case for Recklessness. Reck->BT->Slam->WW->Slam is probably the way to go. I put WW last because I suspect it eats 2 reckless charges and not one when dual wielding.

Is this accurate? From reading the notes I took it as being a 20% chance on all *hits*, normalizing it for the most part and making it more useful at lower gear levels.

I'm also thinking this may be erroneous; I'm pretty sure I can reck->ww->bt->slam on live, each attack burning 1 charge. I'll test for certain when I get home tonight.

No disrespect intended of course, but I want to make sure I'm not misguided either in my interpretation of PTR notes and talent trees.

I completely disagree. Execute will be the king still. I definitely see vast improvement in execute phase. I do ~9k dps on patchwerk during execute phase which is what gets me close/equal to 6k at end (early dps is as low as 4.5k). Just look at dmg meter where warriors really arent too high then skyrocket past everyone in execute phase.

Key to execute effectiveness is simple - its much better then heroic strike. Thats where the rotation breaks. Instead of doing a heroic, its much better to do an execute, and then you are left without rage at all. It feels most of the time unless your MH crit, when your gcd after execute lights up, you do have enough rage for another execute, but not enough for a BT/WW. Thats my case anyway. Therefore i execute again, and so on so on. Now what i of course do, is use WW/BT IF at any point im at high rage (basically even if im at 25 ill WW). But saying heroic strikes are amazing is pushing it too far.

Besides 10k BT with DW, requires some good procs (greatness/MoT) to happen. Executes at full rage are at 13800 for me, or a bit above 17k with hysteria (which actually is a bit strange, since hysteria is supposedly adding 20% - those are the actual numbers though, i guess ill have to look at it closer).

Heroism is almost always up for execute phase (there is very little reason to not have it up there), so its kinda pointless to discuss what happens without it.