 |
03/10/08, 7:34 AM
|
#51 (permalink)
|
|
The Sly
Draenei Shaman
The Venture Co (EU)
|
Haste vs. Crit, Mathematically (as I understand it)
Firstly, let's ignore all specials talents, eg. impale, flurry, deep wounds etc.
1% haste means you get 1 extra 'swing' per 100 swings. A swing can miss, be dodged, or crit as normal. It's damage is therefore (avg. swing damage) * (1 + crit - miss - dodge)
1% crit means you get 1 extra 'crit' per 100 swings. A crit turns a 100% damage hit into a 200% damage one. It's damage is therefore (avg. swing damage)
DW miss chance is 27%, I think? With 9% hit, this means 18% miss, and a normal 5.6% dodge, reduced to 3.6% (for now, let's assume 0 expertise as it's not *currently* a common stat on warrior dps gear). We'll use Hiddens example of 45% crit for now, though this is *vastly* higher than what I run with (we don't put our feral druid in the melee group, he goes with the hunters and we get the ret paladin, plus some of those stats seem a bit optimistic).
So, the damage of the extra hit from haste is therefore: (swing) * (1 + 0.45 - 0.216) = 1.234
This is for white damage only though, and haste does not affect yellow damage (bar a small amount of interaction with heroic strike), whereas crit does. Looking at my damage parses, yellow damage (excl. heroic strike) is normally around 50%, maybe slightly less, meaning that crit's effect is pretty much doubled.
This means we're comparing a crit:haste dps bonus ratio of 2:1.234
This is per 1% though, and haste *is* cheaper in terms of itembudget - 15.7 per 1% compared to 22.1 (correct me if I'm wrong) for crit.
Dividing to get per rating point, we get 0.0786 for haste and 0.0905 for crit, so Crit still comes out ahead per item budget point.
This is before we include flurry, impale, RED etc. which *all* increase the value of crit compared to haste, and this is already at very high levels of crit. At lower levels of crit, haste will decrease in effectiveness against crit.
In short, haste will pretty much *never* be better than crit, rating point for rating point. Of course, this doesn't stop particular items (eg. DST, Dragonstrike) with haste from being very good, and it's still a useful dps stat (more than some might consider). I'm looking forward to getting my JC neck in 2.4
|
|
|
|
|
|
03/10/08, 8:17 AM
|
#52 (permalink)
|
|
Von Kaiser
Orc Warrior
Stormreaver (EU)
|
The shoulder enchant "Might of the Scourge" from Sapphiron in Naxxramas is still the best.
|
|
|
|
|
|
03/10/08, 9:19 AM
|
#53 (permalink)
|
|
Von Kaiser
Human Warrior
Bladefist (EU)
|
Originally Posted by Cavein
The shoulder enchant "Might of the Scourge" from Sapphiron in Naxxramas is still the best.
|
Noted, but I really wonder how many people are in guilds that still clear to Sapphiron just for people to get mildly better shoulder enchants.
In regards to Moogul's math on Haste, that extra swing in 100 should, as far as I know, be subject to glancing blows as well unless anyone has information to the contrary.
|
|
|
|
|
|
03/10/08, 9:27 AM
|
#54 (permalink)
|
|
The Sly
Draenei Shaman
The Venture Co (EU)
|
Originally Posted by Voxx
In regards to Moogul's math on Haste, that extra swing in 100 should, as far as I know, be subject to glancing blows as well unless anyone has information to the contrary.
|
Well spotted, that's something else I'd missed out from the calculations.
|
|
|
|
|
|
03/10/08, 9:56 AM
|
#55 (permalink)
|
|
Von Kaiser
Human Warrior
Ragnaros (EU)
|
Is there anyplace i can see the underlying discussion that lead to the conclusion that deathwish would be superior to rampage for a 2-hand fury build (21/40/0 vs 20/41/0). Intuitively rampage seems to be a lot better, and there is plenty of time to keep it up in the "instant" slot in the rotation.
|
|
|
|
|
|
03/10/08, 11:16 AM
|
#56 (permalink)
|
|
Glass Joe
Orc Warrior
Kel'Thuzad (EU)
|
Originally Posted by Machinator
For the people asking about 9% and precision. Hit does lose effectiveness after 9%, but with more hit your rage generation is smoother. Missing with one or both weapons can throw off your cycle, which is something the spreadsheets cant show. How much hit you personally need is really something you need to answer yourself. If you feel like you dont have rage due to misses get more hit. Many best in slot pieces come with a decent amount of hit anyway, but in general you dont need more than you get on those pieces ie. dont gem for hit unless under 9%.
|
Well exactly what feeling I had a few months ago. I simply missed alot of rage in fights, wich resulted in mostly having barely enough rage to use BT and WW but none at all to dump rage in HS. So I respeced and took 5/5 in Unbridled Wrath, what solved that problem for me. Now, that my hitrating is again above 200 I have no rage problems at all and due to BT and MH gear giving me even more hit I think I'll put that 5 points in UW again somewhere else. Although I don't think I'll be too happy with a hitrating under 200 and as the guide here suggested that, I was really curious what you guys think.
|
|
|
|
|
|
03/10/08, 12:59 PM
|
#57 (permalink)
|
|
Von Kaiser
Draenei Warrior
Steamwheedle Cartel
|
I believe you are going about the calculations for average damage added per swing slightly wrong Moogul. In addition to the various attack table outcomes, I believe that ignoring the increase that heroic strike brings to a portion of your mainhand swings is important when considering the added value of an extra swing versus a crit.
This means that for offhand attacks, the calculation for average damage remains the same as posted above, (avg. hit damage) * (1 + crit - miss - dodge), as well as taking into account glances. I believe the glance rate is 20-25%, and reduces damage to ~70%, but someone can correct me on that if it's wrong.
However, the formula for average damage from a mainhand swing is going to be more complex. We first need to know the ratio of normal swings to Heroic Strike. Once this number is determined, the average damage per MH swing will be [ (avg. hit damage) * (1 + crit - miss - dodge - glancing_reduction) * (%normal attack) + (avg. HS damage) * (1 + crit - dodge) * (%HS) ]
Pinning down the average number of MH swings that end up as heroic strikes is a bit difficult, but I'll simply estimate one in three, as this is a conservative estimate for my own gear and group composition. With that assumption, as well as the 18% miss, 3.6% dodge and 45% crit you used in your examples, as well as the assumption that glancing will result in ~7.5% lost damage(25% of swings deal 70% damage) we get the following:
[ (avg. MH hit damage) * (1 + 0.45 - 0.291) * .66 ] + [ (avg. HS hit damage) * (1 + 0.45 - 0.036) *.34 ]
[top] avg. MH Swing
[ (avg. OH hit damage) * (1 + 0.45 - 0.291)
avg. OH Swing
That said, I'm having some difficulty fully following this through. Heroic Strike adds a static amount of damage to each swing, which means that solving the above expressions by canceling out the average hit damage term as Moogul did won't tell the whole story. Additionally, we need to know what percentage of damage is coming from the offhand, and what is coming from MH/HS.
|
|
|
|
|
|
03/10/08, 1:44 PM
|
#58 (permalink)
|
|
Von Kaiser
Human Warrior
Bladefist (EU)
|

Originally Posted by Katrael
I believe you are going about the calculations for average damage added per swing slightly wrong Moogul. In addition to the various attack table outcomes, I believe that ignoring the increase that heroic strike brings to a portion of your mainhand swings is important when considering the added value of an extra swing versus a crit.
This means that for offhand attacks, the calculation for average damage remains the same as posted above, (avg. hit damage) * (1 + crit - miss - dodge), as well as taking into account glances. I believe the glance rate is 20-25%, and reduces damage to ~70%, but someone can correct me on that if it's wrong.
However, the formula for average damage from a mainhand swing is going to be more complex. We first need to know the ratio of normal swings to Heroic Strike. Once this number is determined, the average damage per MH swing will be [ (avg. hit damage) * (1 + crit - miss - dodge - glancing_reduction) * (%normal attack) + (avg. HS damage) * (1 + crit - dodge) * (%HS) ]
Pinning down the average number of MH swings that end up as heroic strikes is a bit difficult, but I'll simply estimate one in three, as this is a conservative estimate for my own gear and group composition. With that assumption, as well as the 18% miss, 3.6% dodge and 45% crit you used in your examples, as well as the assumption that glancing will result in ~7.5% lost damage(25% of swings deal 70% damage) we get the following:
[ (avg. MH hit damage) * (1 + 0.45 - 0.291) * .66 ] + [ (avg. HS hit damage) * (1 + 0.45 - 0.036) *.34 ]
[top] avg. MH Swing
[ (avg. OH hit damage) * (1 + 0.45 - 0.291)
avg. OH Swing
That said, I'm having some difficulty fully following this through. Heroic Strike adds a static amount of damage to each swing, which means that solving the above expressions by canceling out the average hit damage term as Moogul did won't tell the whole story. Additionally, we need to know what percentage of damage is coming from the offhand, and what is coming from MH/HS.
|
I made a post about this earlier on in the thread. While more Haste increases your swings, if you're not already at the point where every single Main-hand swing is converted to Heroic Strike, then in the long run Haste won't be affected by Heroic Strike. The idea is that if you have the rage to add a Heroic Strike in somewhere in those 101 hits, it'll be added in. Seeing as most people don't have that rage to convert every Main-hand swing to a Heroic Strike, the extra hit will be a normal attack, not a Heroic Strike.
In short, if there's room in the "normal" 100 swings to add a Heroic Strike, we can't count the fact that the extra swing gained from 1% Haste will be converted to a Heroic Strike, because that could be converted somewhere else in the 101 swings. I hope this makes sense.
Originally Posted by mulinexman
Well exactly what feeling I had a few months ago. I simply missed alot of rage in fights, wich resulted in mostly having barely enough rage to use BT and WW but none at all to dump rage in HS. So I respeced and took 5/5 in Unbridled Wrath, what solved that problem for me. Now, that my hitrating is again above 200 I have no rage problems at all and due to BT and MH gear giving me even more hit I think I'll put that 5 points in UW again somewhere else. Although I don't think I'll be too happy with a hitrating under 200 and as the guide here suggested that, I was really curious what you guys think.
|
What I've stated in the first post is that the general "minimum" requirement for Hit is 9%. That's because after 9% additional +hit will not affect your yellow damage. How much Hit Rating you actually run with is entirely dependent upon your feel for it. If you run with 119 Hit Rating (my personal number) and are happy with it, then fine. If you feel like you're rage starved without 200 Hit Rating, go for 200. I personally have no problems with rage starvation, but I'm currently using a fast offhand so that's a factor. Ideally I'd like to get a bit more Hit Rating for when 20% comes, however that's not the top priority on my list of things to do.
|
|
|
|
|
|
03/10/08, 1:47 PM
|
#59 (permalink)
|
|
The Sly
Draenei Shaman
The Venture Co (EU)
|
Whilst much of what you say is true Katrael, it is only on 'infinite rage' fights (I use the term loosely) where heroic strike will have an affect on the value of haste. For me, currently, I am rarely in the position to have more than, anecdotally, around 70% of my MH hits Heroic Strikes. Adding in more swings due to haste will not directly allow me to use more heroic strikes, as I will not have the rage for it.
In an infinite rage scenario, haste *will* allow for more heroic strikes, but also crit will still increase the damage of those heroic strikes too.
Modelling everything 100% accurately is a gargantuan task, and I don't think it can be fully accomplished short of an actual simulator (something I have contemplated creating), but I'm confident that those elements I have missed out are not enough to change the balance of crit vs haste.
For me, my priorities are mostly: Expertise > Crit/Str/ArP > haste > hit
Unfortunately there aren't many good items with Expertise currently (that's changing a bit next patch), for Crit/Str/ArP I value them fairly closely, though I'm trying to push my ArP up right now, and I generally socket Crit rather than Str (I'm running lowish at 32-33% unbuffed in Zerker, and as I mentioned earlier we don't run with a feral druid in the melee group).
|
|
|
|
|
|
03/10/08, 1:52 PM
|
#60 (permalink)
|
|
Piston Honda
Night Elf Warrior
Maelstrom
|
Glancing blows: its 25% chance for your damage to be reduced to 55-75% damage, for an average reduction of 35%
About Crit vs haste: Think about it. Impale. Flurry Uptime. R.E.D. Affects yellow and white damage. Haste only inc. white damage and grants RPS over cycle to use HS.
|
|
|
|
|
|
03/10/08, 1:55 PM
|
#61 (permalink)
|
|
Von Kaiser
|
So are we saying that if your DW'ing and you have achieved the 9% soft hit cap, that you should stop aiming for gear with Hit% if you can get gear with more Crit/AP/ArP etc?
|
|
|
|
|
|
03/10/08, 2:45 PM
|
#62 (permalink)
|
|
The Sly
Draenei Shaman
The Venture Co (EU)
|
Originally Posted by Axodry
So are we saying that if your DW'ing and you have achieved the 9% soft hit cap, that you should stop aiming for gear with Hit% if you can get gear with more Crit/AP/ArP etc?
|
In simple terms, generally speaking, yes.
That doesn't mean you should completely avoid it, many of the best DPS pieces around have hit on them, and taking the optimal gear will often give you more than 9% hit. Hit also offers smoother rage generation, which affects dps in a way that's hard to model. That said, I believe you should not be aiming to get more hit, and hit should not be the deciding factor in which gear you choose. You should almost never be socketing hit (it's quite easy to get 9% without gems), and Crit, Str, ArP (in some order, which will vary depending on your current gear) are the stats to aim for. Haste is also better than hit when you have +9% hit.
|
|
|
|
|
|
03/10/08, 3:01 PM
|
#63 (permalink)
|
|
Von Kaiser
|
Thanks..that was what I was thinking..sometimes you can't avoid going over the Hit cap due to the fact the gear is an upgrade and it happens to have hit on it. I always find myself wanting to HS when i'm doing my dps, but I guess its better to only HS when your rage is over 60 so that you dont miss a BT or WW when its only CD.
|
|
|
|
|
|
03/10/08, 3:57 PM
|
#64 (permalink)
|
|
Von Kaiser
Human Warrior
Bladefist (EU)
|
I'll add all this about +hit to the FAQ reservation since it seems to come up no matter how often we discuss it.
|
|
|
|
|
|
03/10/08, 7:40 PM
|
#65 (permalink)
|
|
Glass Joe
|
slow or fast off hand
Ive heard that a slow offhand is better dps for duel wield fury but i have also heard a fast one is also better. maybe someone will have an answer.
|
|
|
|
|
|
03/10/08, 7:55 PM
|
#66 (permalink)
|
|
Glass Joe
Orc Warrior
Kel'Thuzad (EU)
|
slow OH = more WW damage
fast OH = more rage
So regarding that what has been said to hit: If you want to keep your hit low (eg. go for other stats), take a fast OH. If have quite some hit, take a slow OH.
|
|
|
|
|
|
03/10/08, 7:56 PM
|
#67 (permalink)
|
|
Glass Joe
|
thanks for the help
|
|
|
|
|
|
03/10/08, 9:10 PM
|
#68 (permalink)
|
|
Von Kaiser
Human Warrior
Eredar (EU)
|
Originally Posted by mulinexman
slow OH = more WW damage
fast OH = more rage
So regarding that what has been said to hit: If you want to keep your hit low (eg. go for other stats), take a fast OH. If have quite some hit, take a slow OH.
|
That's simply not true. A fast offhand and slow offhand both offers the exact same rage in average. However, slow offhands are harder to handle because the rage income is in big chunks, rather than small chunks. That said, when your hit is low, you *CAN* have problems with rage sometimes, doing the BT/WW-Cycle. If a warrior can decide between a fast and slow offhand with similar stats, the warrior should always get the slower weapon.
I played with 122 Hit and slow offhand, and it was only rarely a problem in 25 man raids fully buffed with melee shaman.
|
|
|
|
|
|
03/10/08, 9:35 PM
|
#69 (permalink)
|
|
Von Kaiser
Human Warrior
Bladefist (EU)
|
Sigh. This is all discussed in the first post... look at the "Attack Power Normalization and Weapon Speeds" section.
|
|
There are three cardinal rules for raiding: Sometimes there's fire, you have to not be in it. Sometimes there's green shit on the floor, you have to not be in it. Sometimes there's falling shit from the sky, you have to not be in it.
|
|
|
|
|
03/11/08, 5:22 AM
|
#70 (permalink)
|
|
Von Kaiser
|
Couple questions. Armory's at the left if you need it.
(And yes, before it is asked, that is the best I can muster for trinkets - I've lost on every Berserker's Call I've seen, and no TT's, DST's or Madness' have dropped for our guild. Ever. EVER. Fuck this game.)
Question #1 - I usually run a 33/28 spec, but lately we've had a second Arms warrior in the raid, and of the two of us I'm better geared, so tonight I gave 2h Fury a try (not really interested in DW Fury at all, the playstyle just isn't to my liking). Unfortunately, our WWS guy wasn't logging tonight, and we only did Azgalor and Archimonde anyway (not the best DPS tests, especially since on the Arch kill I spent 2 fucking minutes running away from one unfortunate Doomfire after another and ended up in the goddamn Well of Eternity for my troubles, a unique experience I don't care to repeat).
Anyhow, without WWS to go on, my impressions were this - spottier rage generation than 33/28 because of loss of crit from Axe Spec, which led to several instances where my Flurry was down for a decent period. Our dps group usually runs warrior/warrior/shaman/rogue/retpally, with our feral in the tank group (or, in the case of Arch, out healing/decursing). The individual hits were lower as well, and even BT felt like it was landing more weakly than MS, but I'm sure 5/5 Flurry (when it was up) was making up for this. I was doing the expected rotation of 4x Slam/2x BT/1x WW, with very few HS if any, since my rage generation was so much different.
My question, after all of this shit, is... am I missing something? 2h Fury just felt like... less, which seems the opposite of what I expected. Should I just go 31/30 (5/5 flurry in exchange for Blood Frenzy, since the other warrior is almost always there) and save myself the trouble?
(again, I wish I had a WWS to better examine this, but sadly...)
And, much more simply...
Question #2 - I (finally) replaced Fel-Steel Warhelm tonight with Onslaught, which should boost my dps dramatically (thanks metagem!), but it had the unforunate side effect of dropping me below the hitcap with no clear way to get back above it.
I'm sitting at 130 hit, so I have a couple options, and I'm pondering which is the best:
1 - Surefooted, and live with being 2 under cap. Not preferrential, because I love me some run speed, especially on Council and Illidan attempts. (I use Boar's over Cat's because I occasionally PvP and fuck it, 9 more stam).
2 - Arcanite Steam-Pistol over Tuskbreaker. 18 hit would cap me plus some excess, and 19 agi would help with crit (esp if I stay 2h fury), at the cost of a little AP and some meh haste.
3 - Brooch of Deftness over C. of Serrated Blades. This is where my real question is. My (outdated) dps spreadsheet tells me the Expertise/Hit gain is a much lower dps increase than what I lose in ArP and AP from Choker, but then this thread is telling me than Hit/Exp > *. Is my spreadsheet just calculating shit wrong, and I should probably update? The answer to the second half of that is undoubtedly yes, but am I missing something besides?
4 - Redo my Fury spec to dump a point into Precision and just say fuckit. 50g is simpler than changing gear obviously, but would the added Expertise from Brooch make that a better choice?
Last edited by Corkscrew : 03/11/08 at 5:34 AM.
|
|
|
|
|
|
03/11/08, 5:25 AM
|
#71 (permalink)
|
|
Von Kaiser
|
Also, as a more general question, one that I didn't see addressed here (but could easily have skimmed over it by accident)...
Is it just considered a matter of course that being too quick on the trigger with Slam will cut out extra attacks from Windfury? I've noticed lately that I've clipped WF swings several times by Slamming too quickly - I'll see Extra Attack come up on SCT, but only one white swing actually occur. This is obviously a massive dps loss, both in straight up damage and rage gained, and I'm anxious to make sure this doesn't happen. Sometimes, though, it feels as though my Slams are landing massively late as I try to counter this... up to a 0.7 sec lag time (I usually have 150-200 latency).
Does anyone know of a trick to avoid this, something I'm just not seeing? I use Quartz, although unmodded, so fixing that would likely help things as well (since certain haste effects screw it up), but anything else?
Edit: Also, while I'm spouting questions like an idiot, does this too-fast Slam crap negate Sword Spec swings too? I haven't been Swords in a while, using Soul Cleaver at the moment, but I'd like to know before Cat's Edge (FINALLY) drops and I switch back, assuming I don't stay 2h Fury.
|
|
|
|
|
|
03/11/08, 8:44 AM
|
#72 (permalink)
|
|
Von Kaiser
|
This happens to me a lot as well. Clipping windfury and sword spec attacks is just something that happens occasionally, and you have to get used to it.
Its because although technically they are "instant" attacks, they still take place a short time after the initial swing happens. I find that if you constantly delay your slam just in case you were to get a wf/ss proc, you'd probably end up losing a similar amount of damage overall. Just a design flaw that we have to live with unfortunately.
|
|
|
|
|
|
03/11/08, 8:50 AM
|
#73 (permalink)
|
|
The Sly
Draenei Shaman
The Venture Co (EU)
|
Originally Posted by Corkscrew
3 - Brooch of Deftness over C. of Serrated Blades. This is where my real question is. My (outdated) dps spreadsheet tells me the Expertise/Hit gain is a much lower dps increase than what I lose in ArP and AP from Choker, but then this thread is telling me than Hit/Exp > *. Is my spreadsheet just calculating shit wrong, and I should probably update? The answer to the second half of that is undoubtedly yes, but am I missing something besides??
|
Remember, it's not just the type of stat that's important, but the amount of it. Expertise is a great stat, and Hit also when you're under the special hit cap (for 2h, the complete hitcap), but you have to compare the amount of them against others.
The [Brooch of Deftness] has 43 item points worth of offensive stats (1 rating = 1 item point)
The [Choker of Serrated Blades] has 60 item points worth of offensive stats (2 AP = 1 item point, 7 ArP = 1 item point, so 35 points in AP and 25 in ArP).
So, whilst hit and expertise are very nice, the Choker just has a lot more going for it due to a) being higher level and b) not wasting quite so many points on stamina.
Stats should never be considered in a Vacuum, you need to look at the items available. I (generally) socket crit over str, since gems have an equal amount, but I'd never take, say, 10 crit over 14str, because 'crit is better'.
To answer your question, I would probably choose to take a point in precision, but unfortunately I can't see your talent spec properly (Armory is crashing Firefox at work).
As for your effectiveness with 2H fury as a whole, I can't speak with any authority on the subject as I haven't been 2H fury since 60, but your attacks *should* be hitting harder due to Imp Zerker Stance, and your bloodthirsts should be hitting for more than your Mortal Strike. As much as it might be an obvious response, and one you're not looking for, I would suggest looking at a WWS, or failing that, just get a nice real time damage meter like Assessment or Recount - something that can show you some details on your performance.
I've had situations in the past where I've 'felt' I was doing well, only to find out afterwards that I wasn't, or the opposite.
One last thing, if you do feel you really were hitting weaker, it might be worth checking that the bosses were fully debuffed? Maybe the warlocks were slack and weren't CoRing, or you didnt have a Survival Hunter, etc. Things like this can make a big difference that you might not notice straight away.
|
|
|
|
|
|
|