I took these stats from armory currently i am at:
Haste 6.43%
Hit +10.98% (+3% from precision)
Crit 17.27%
Expertise 26/26
My question is should i reforge my two trinkets who have hit on them into crit?
I really don't know how useful this post if going to be, as I am literally doing the math as I am typing this, so bear with me.
The thing about haste is that it also results in more rage which ideally is used for more HS's. I am going to keep this model incredibly simplistic because I'm lazy/tired.
(the numbers for the haste damage and what not were ripped from this log also note that the RB damage values were ripped from this log too.)
per 1 MH+OH hit combo, I gain 25 rage. this is 9.615 RPS
if I gained 1% haste (up from 0), my new attack speed would be ~2.57 (I'm using the non truncated number in my math)
this is 9.712 RPS
the gain is 0.09615.... RPS, which equates to 1 gained HS from that rage every 312 seconds (el oh el). and a brief glance at my logs suggests that my HS hits for ~18.3k, which is a gain of 56 DPS (it's really more than this due to HS crits giving deep wounds/flurry and what not)
Now, because of the previously mentioned change in attack time, I gain an extra MH/OH hit every 38.84 seconds. and for the sake giving mastery the best possible chance of being better than haste, I am only going to look at the scenario of getting two glancing blows from the gained attacks.
if I get a double glance from those two gained attacks I gained ~10k damage, which is a gain of 257.42 DPS
the net resulting DPS from that 1% haste is ~313 DPS
Now lets look at mastery.
(using enrage up times from this log note that I modified the uptimes to account for only the time that I spent DPSing, not total time in combat)
Enrage was up for 57.172% of the fight
Deathwish was up for 24.401% of the fight
this means that "on average" 1 mastery would result in
a ~0.6069% Damage increase from enrage
a ~0.2497% damage increase from deathwish
(note that is this is in terms of total damage, and it's not even going in the same country as the math involving cooldown stacking and what not)
added together this is a gain of ~0.8566% damage
now for RB gains 6.44% damage per point of mastery due to the +15% from war academy, and my RB's hit for ~20.5k, meaning that I would gain ~1318 damage per RB if I gained one more point of mastery. and on the nefarian log, during my DPS time I was averaging an RB every 10.7 seconds, which means the extra mastery would result in a gain of 123.178 DPS (before the above % increases mind you)
so, mastery gain is:
1.8566*(normal DPS + 123.178)
and haste gain is:
Normal DPS + 313
so let's find how much DPS you would have to to for them to be equal to eachother
1.8566*(normal DPS + 123.178) = Normal DPS + 313
1.8566*normal DPS+228.692 = normal DPS+313
bla bla bla
normal DPS = 98 DPS
so obviously it's possible for mastery to still be better than haste, but that was with giving mastery the best possible chance
looking at the double crit scenerio for haste, the DPS gain is 699.666 DPS, making the average DPS needed for them to be equal is 549.81
basically this a long way of saying that either my model really really sucks, or mastery is still definitely better than haste.
EDIT:
to recap a few things for the sake of clarity:
Things not accounted for that do alter the values of haste:
gained HS's critting, causing flurry uptimes and deep wounds procs
gained swings critting, causing flurry uptimes deep wounds procs
interplay with cooldown usage, IE if those extra attacks land during enrage/deathwish/CS/etc.
actually accounting for all the possible outcomes for the gained attacks instead of looking at best/worst case
Things not accounted for mastery:
cooldown stacking during deathwish (this is pretty damn important)
variances in RB usage (dependent on slam usage)
solid modeling of enrage/deathwish uptimes
Last edited by Thegreatme; 05-12-2011 at 02:23 PM.
Wouldn't your "haste gain" be Normal DPS + 313? not 257.42 since that was the gain with glancing blows? Obviously I'm not calculated any of this; I'm only going off of what my eyes tell me, but everything else seems correct.
Again, I'm at work so unfortunately I cannot access the tooltips of mastery and talents, but I still believe that mastery also affects the amount of damage done with white hits since the white hits are what cause you to become enraged. That should add a bit more emphasis on mastery and make the tired math above become even more complicated.
Unless I am missing something, I only see you adding the mastery benefits to RB and Deathwish (I'm not considering cooldown stacking as you already said that it's not included but it's also very important). What if you were to take your 6.44% increase that you added to RB and also add that to the white hits used in that log during that certain time? Would that more accurately portray the benefit of mastery without the things you listed on the bottom of the thread as not being accounted for?
Since you did most of the math already, that saves me from looking too much into it this weekend so Thank you TGM. Something I am curious about now, though, is how much rating does it take to get 1% haste vs. the rating it takes to get 1 mastery. That would also go into account at which is easier to stack for the greatest benefit. If it takes less mastery rating to get 1 mastery than it does haste rating to get 1% haste, that adds more weight to the mastery section. However, if it takes significantly less haste to get to 1% haste than the 1 mastery, then the tables might lean towards the haste a bit. Then, we can play the balancing game between the two, using the math above and the ease/accessibility of the ratings. (Meaning would we be able to hit that magic balance rating number by reforging to haste or mastery if we needed to)
I will be checking to see how much rating it takes for my character to get those numbers probably tomorrow.
Last edited by Crittable; 05-12-2011 at 12:33 PM.
White hits aren't abilities, they aren't affected by Unshackled Fury.
|TGM UI |Fury Warrior Guide | How To Use Landsoul's Spreadsheet| The Numbers Game |
"I am an elitist; I will never accept mediocrity and I openly show no respect for stupidity"
I just ran some numbers based on my own stats ( http://us.battle.net/wow/en/characte...rious/advanced )
-to the effects of reforging hit > haste. With my current gear it isnt possible to drop my hit closer to the cap than 8.5% without swapping out gear or sacrificing expertise needed to make the 26 cap, or CRIT which is still a better stat than haste or hit.
I traded 447 hit for haste. Haste contributes less to each stat % than hit per-point, so i lost 3.7% hit and gained 3.56% haste, and my weapon speed went from 3.36 to 3.22. I would be expecting to gain about 3 extra white hits per 10 mins (about 16 extra attacks but 13 additional misses).
What I found is that trading hit for haste (obviously still aiming for the 8% hit cap so as not to affect special attacks) results in a small DPS gain from white damage alone, at my current unbuffed CRIT rating (around 20% - at under 19% CRIT, there would be an expected DPS loss). There is an expected rage gain, since there is a net gain in the number of hits being landed.
The higher my CRIT, the more haste will benefit me over hit since this is where I have the most potential to make up any DPS lost through white misses. Since my CRIT gets buffed by around, what, 5% in raids?, there is little question that haste > hit for me. The faster attack speed has the added advantage of smoother rage generation.
At the moment this makes my stat priorities : hit (to 8%) > expertise (to 26) > STR > CRIT > haste (if crit 20+) > hit > mastery.
Welcome back to dump-stat status, mastery; you've been missed.
Note that for modeling these numbers, I used a single-roll hit-table with assumed values of 24% chance for glancing blows doing 65% damage of normal hits.
Okay, disregard my earlier conclusion about haste > hit... I did some testing on dummies combined with looking through my recent raid logs and realised something was askew with one of my assumptions.
I noticed that my auto-attacks were consistently critting about 5-6% less than my crit rating indicated I should be, and I realised this is because crits are calculated from hits only, not attacks (particularly attacks that miss or are glancing blows). Basically this means than crits from white attacks scale inversely to the amount of hit you have, and since crit was the only redeeming factor that made haste-scaling seem good in my previous model this conclusion is now debunked.
My model DID accurately show a relationship between hit and haste however, unfortunately this requires you to crit about 19.5% of the time on white attacks; to get this you would need to be hit capped @ 27% and have a crit of 26%+ since you will always do ~24% glancing blows to boss-level mobs meaning you only apply 76% of your crit.
I am eagerly awaiting the outcomes of TGM's haste vs mastery modelling, since it will determine at which point haste or mastery takes priority relative to hit before or after the hard hit-cap is reached.
Haste and hit - this doesn't factor in Flurry uptime at all, but what I found when I looked at this a few months ago, if my math is right is hit > haste unless you have more than 28% (so basically if you have to choose, until you get to 28% haste, you get more landed hits via hit over the same time period).
This also doesn't take into account the possibility that hit and expertise (only?) might require more rating per % in 4.2. Last I heard this was still a rumor?
Now, if my math is correct...
Hit is a pretty simple formula:
Over 10,000 seconds, going with 3% increases (just so I don't map out all 27% of hit) - there will be ~3846 swings with a 2.6 weapon speed weapon. I chose 10,000 seconds to just make any differences visible.
So 0% hit = 27% miss = 2808 landed swings
3% = 2923
6% = 3038
9% = 3154
12% = 3269
15% = 3385
18% = 3500
21% = 3615
24% = 3731
27% = 3846
Now, for haste, I hope I have the formula set up correct - [ Weapon Speed / (100% + (haste % * buffs like Hero)) ]. Also, because I was looking at the effect opposite of each other, I use 0% hit (27% miss chance) so the rate choice is diametric, haste or hit, but only one.
What I found was:
0% = 2808 (again, this is 0% hit, 0% haste, it is the same for both)
3% = 2892
6% = 2976
9% = 3060
12% = 3145
15% = 3229
18% = 3313
21% = 3397
24% = 3482
27% = 3566
At 27% you see hit landing 3846 where 27% haste lands 3566, a difference of 280 landed swings over 10,000 seconds. A difference of 0.028 landed swings per second, or, for a 'normal' 10 minute (600 second) enrage time, 600*0.028 = 16.8 landed white attacks per boss fight. This number is high, because of movement and mechanics, but shows the difference.
This relationship also held up when I increased each side by 0.5% at the same time (so .5% hit and .5% haste, 1% hit + 1% haste) to see if there was a point that they worked better together - I didn't find one by the time I hit 13.5% hit and 13.5% haste (which would be the same rating required for 27% hit).
The effect was also linear, per %, hit seemed to resulted 10.38 more landed attacks per 10,000 seconds than haste.
****
Now, as with the caveats - I didn't model in Flurry or Icy Talons or Hero's effect. And I'm not even 100% sure my haste math is correct.
What I would guess, when you factor in those additional factors is that at some point, and probably heavily influenced by crit and flurry uptime, haste is more valuable than hit, but not at low haste/crit levels.
I'm also not convinced that haste is smoother rage generation than hit. Unless I'm missing something, haste doesn't make generation smoother, it makes it's burst potential larger.
Hit will result in fewer bad miss strings and more MH hits (OH as well, but MH miss hurts more than OH miss). Thus hit results in smoother rage generation, haste may aid in more rapid recovery about a bad miss string.
At some amount of haste, when you factor in Imp Icy Talons and Flurry, haste will generate more rage, but even then, it's not smoother, just more of it. But until then, hit will be smoother and result in more rage.
An introduction into WarTanking (no longer updated as I've retired from WoW - the concepts will still be mostly accurate but the numbers no longer will be.) - http://www.tankspot.com/showthread.p...101-The-Primer
landsoul recently touched on what you were talking about at the end of your post loganisis,
the blog post he is referring to is this oneHit makes the normal curve skinnier and taller at the large rage increments, while haste makes the curve fatter. My blog has more detail about that curve.
also I am not really concerned about hit vs haste right now as I would definitely agree that there is almost no question that hit is better, but what I am interested in is haste vs mastery.
I'm working on a more accurate model in maple13 currently, and am a bit stuck on how to handle the frequency of "gained" attacks from haste.
Basically I am using the same premise to find the time per 1 gained attack set, but the problem is that because the haste variable is in the denominator, it makes the equation non linear so I am having a mindblock on how to handle that when I start comparing it mastery on a point by point basis (which is my end goal)
EDIT: I am also attempting to keep this working on an infinite timescale, because I personally like the consistency generally provided by doing so, but I am beginning to think that I won't be able to in this case
Last edited by Thegreatme; 05-13-2011 at 04:21 PM.
|TGM UI |Fury Warrior Guide | How To Use Landsoul's Spreadsheet| The Numbers Game |
"I am an elitist; I will never accept mediocrity and I openly show no respect for stupidity"
no matter how I end up modeling it, I keep ending up with a DPS cut point of ~16k. meaning if you are under that, haste is better, and if you are above it mastery is better. Please note that all of my models were using my own values of crit, and were assuming 2x heroic lavaspine.
also because I am never fully confident in my ability to model things, I would say that the accuracy of my work is in the 80-85% range.
So basically long story short get mastery over haste.
Last edited by Thegreatme; 05-20-2011 at 09:07 AM.
|TGM UI |Fury Warrior Guide | How To Use Landsoul's Spreadsheet| The Numbers Game |
"I am an elitist; I will never accept mediocrity and I openly show no respect for stupidity"
Hey all kinda new to the site, but mad props great info source. So here is my question (as a tg fury war) should I go for 20% crit before or after b.s. or should i just continue to build crit as my best reforgeable stat. And second question is the off hand weapon exp. as important to have exactly at cap, b.c. of the human racial stats I feel like I am wasting points when then main has 29-30 exp and the off is right there at 26.
http://us.battle.net/wow/en/characte...zumaa/advanced
I actually have been wondering the same thing concerning main hand vs offhand expertise. I'm currently at 29/26. Do I need to ensure that both are a minimum of 26 for expertise or only my MH rating?
Your ego is writing checks your body can't cash
I believe crit is still worth priority reforging for over 20%... the only reason it theoretically gives slightly diminished returns after that point is because around 20% crit flurry should be seeing fairly frequent uptime, and flurry procs are less beneficial the more often they occur since it becomes less and less likely for each proc to occur at a time which allows them to grant 100% of its benefit.
On expertise, I often wondered this myself and concluded that its more important to have 26 expertise on your main-hand than your off-hand. The reason for this is that as TG the majority of your special attacks are made with your main-hand only; only your raging blow and auto-attacks get made with your off-hand weapon, and they are significantly weaker than your main-hand attacks. For SMF it seems like it would be more important to have 26 on both weapons since slam also gets made with both weapons and slam procs now properly give their full bonus for off-hand hits in patch 4.1, although; since swords are the only viable STR dps 1-handers at the moment you are unlikely to end up with an expertise imbalance due to a racial bonus.
I would imagine that reforging for CRIT &/or HIT will give greater DPS returns than the last 3 points of expertise on an off-hand weapon, if the main-hand is already 26, simply because CRIT will buff all attacks, and HIT (over 8%) will buff main-hand and off-hand white attacks equally which is significant because white attacks are also quite strong as of patch 4.1
Last edited by Beefius; 05-25-2011 at 06:35 PM.
Guys, as for a wrist enchant, wich is better? 65 crit OR 50 strenght??
50 str. Str flucates but is roughly 1.5-2x the value of crit.
An introduction into WarTanking (no longer updated as I've retired from WoW - the concepts will still be mostly accurate but the numbers no longer will be.) - http://www.tankspot.com/showthread.p...101-The-Primer
OK let me ask this of ya'll should I, as fury use 2 souldblades or 2 worm-breakers? Basically I am asking which is the current top dps spec. TG or SMF
Last I heard, SMF > TG for maximum potential damage - however I made the switch and I'm not sure I should have. I was pretty good at TG and I just don't feel I have the same proficiency as SMF (read I'm an idiot that hits HS too much because I just can't stand spamming only one button, what I am, an Arcane or Frost mage?).
If you have both, try them both out and the one you are most proficient with is likely to be the one to go with.
An introduction into WarTanking (no longer updated as I've retired from WoW - the concepts will still be mostly accurate but the numbers no longer will be.) - http://www.tankspot.com/showthread.p...101-The-Primer
SMF is more dps overall, however, the dps increase is relatively small. Also, because the play styles are so different, if you are already doing good dps as a TG warrior stick with it. IMO its more fun now anyway because you actualy get to use RB...
Bookmarks