Boevis

Correct me if I'm wrong, but this isn't supposed to be a discussion thread is it? As far as I know this thread is essentially a "Factual Reference Thread", debates on theory/mechanics really have no place in this thread and belong in one of the many threads that are likely already devoted to that subject (especially healing which is so circumstantial that almost every boss gets their own "how to heal on X" thread)

 

Smasher

I agree with the above, but following conclusion is wrong because it assumes deterministic
arrival of damage events and healing events. In practice this is not going to be the case
precisely because of lags, network latency, human reaction time, special attacks and
random damage events etc... What it means there will be some average number of the
outstanding damage events. Also there will be a non zero probability of getting streaks of
unhealed damage events which are longer then average.

This is where extra effective HP or avoidance is needed. What I shown in my post is that
having high avoidance benefits more to surviving unlucky event then HP. Of course tank
must have enough HP to allow healing in the steady state (about 2-3 hits).

 

Nite_Moogle

Steady damage doesn't kill tanks, spike damage kills tanks. Having enough HP to survive spike damage with a buffer of HP on reserve is the best way to gear once you're capped, which means either more HP or more AC. If AC weren't so insanely expensive in terms of item budget versus effectiveness at high levels of itemization and nearly impossible to attain outside of what's already on your gear it might be a different story. Whenever it's your choice (i.e. sockets), stamina is a better choice.

Yes. Someone else will need to carry this on at some point in the future since I'm not playing anymore.

 

Zurgat

New link for rogue spreadsheet : [Rogue] DPS Spreadsheet
The old link can be deleted and the original thread locked.

-= Random Ravings - RSS Feed =- Rogue and Hunter stuff here. As well as guides to get you trough your spare time.

 

• Wodin

As was kindly pointed out to me on SA, there's an error in the rogue section. It currently reads:
This statement wildly overstates the viability of Mutilate, and was probably coined before the changes to melee mitigation on raid bosses. I'd suggest replacing the word "slightly" with "well" to fix it.

 

Rinced

For Rogues :

0.1 Speed on Offhand is worth about 10 OffhandDPS (with Combat Potency)

 

DecimusGarona

Not to nitpick, but if you are going to make a generalized statement like that you might want to make it as a percentage of your dps - since obviously a rogue with 2000 AP, 30% crit and an epic weapon is going to make a lot more out of the energy from a CP proc than a rogue with 1k AP, 20% crit, and a green weapon. You might also want to calculate separate numbers for both backstab and sinister strike assuming standard combat raiding builds for both.

 

galzohar

It's even more complicated than that, as with more hit CP will proc more but your energy is a smaller % of your DPS due to it only affecting specials and your white DPS is higher. Not easy to generalize, although if someone came up with it being worth 10 OH DPS I doubt anyone would get results far off of that, but definitely not the same (assuming he doesn't have errors obviously), as the value of stats changes slightly but not a lot when you upgrade gear.

 

DecimusGarona

Keeping in mind that I'm not an expert on rogues, the way I would go about calculating something like this is:

First of all, calculate how much energy per minute you gain on a 0.1 speed change on your offhand.

Calculate how many extra backstabs or sinister strikes you get per minute as a result

The gained DPS is now (# of extra attacks per minute) / 60 * (insert average attack damage)

This gives you a general formula where the user can insert their average backstab or sinister strike, which can be obtained from either a WWS report or by theorycrafting. This gives more exact results that can be easily obtained for anyone given their current gear level. For the sake of simplicity, and I have no idea if it is actually true, I'll assume that extra combo point generation balances out with any lost CP procs due to misses/dodges/parries on your offhand. I'm sure somebody somewhere would have a better formula or spreadsheet to calculate the exact effects.

In any case I doubt it really matters, since there really isn't a whole lot of choice for offhand daggers in end game TBC. My guess would be that the list pretty much consists of Messenger of Fate > Merciless Gladiator Shiv > Searing Sunblade, the last 2 being fairly easily obtainable by anyone with some minimal dedication to the game, either through arena point farming or heroic badge farming.

 

Evolve

Could you add the Paladin TPS spreadsheet to the list of spreadsheets please ?

[Paladin] Tank TPS spreadsheet

 

Dontmindme

I believe it has been well established in the Rogue threads (over 40000 attacks without a miss) that 308 is sufficient to not miss with Precision and Weapon Expertise dual-wielding. Part of the discrepancy is below. 308 hit = 19.531707+5+1=25.531707% to hit.

The exact formulas for the rating system (not rounded to .1%) are the ratings for 60*(82/52) with the exception of Haste which was reduced to 2/3 it's original value in a patch (i.e 10*(2/3)*82/52)

 

Deelit70

- There is no "cast rotation" for Warlocks due to variance in DoT duration.

This is clearly not true imo. Just because the cast rotation is complicated does not mean there isn't one. There is one and there is a most efficient one, it just can't be modelled easily.

An example:

Suppose you want to use UA, Corr, and CoD. In between these castings you want to fit as many 2.5s Shadowbolts as you can. Unfortunately it is impossible to fit everything in perfectly. You will lose half-seconds and whole seconds as you wait for you dots to expire, or you will cast the Shadowbolt and go without 1.5-2 seconds of the dot being up.

You could choose to split it up stupidly, for example part of your cast could be UA, CoD, Corr. UA and Corr will run exactly 1.5 seconds from each other, which means if you want to maximize your dot damage and keep UA and Corr up, you have wasted a whole 1.5 seconds as you continue to refresh UA/Corr. During that 1.5 seconds you can't do anything else except cast Searing Pain, which nobody in a raid would do anyway, or Lifetap, which you may not need (Say on the first round). That means the 1.5s is essentially dead time. On the other hand, you could cast CoD, UA, Corr. This may result in the occassional half second or full second of loss, when they fail to synch up with your SB spam, but it won't result in the SYSTEMATIC loss that the first pattern would (losing 1.5 seconds every 18 seconds FOR SURE).

As you can see, not all cast rotations are created equal, which right there ought to mean the removal of that sentence, but I would go further to say that not only are they not created equal, but there is a determinable "perfect" cast rotation that minimizes systematic loss. Since, as you note, there is a variance in dot duration, as well as lag, being forced to move and whatnot, there is no perfect cast rotation all the time, but there is a cast rotation that minimizes systematic loss, albeit a complicated one (unlike rogues, we can't just go 1,2,1,1,1,1,1,3, etc. ours would be more like 1,2,3,4,5,2,3,5,4,2,3,1,5,4, etc etc) but it *is* out there, and it can be determined, both empirically and theoretically.

 

◊ Shalas

The existance of non-ideal cast sequences does not mean that the ideal cast sequence involves a rotation. If the ideal cast sequence is 3,1,4,1,5,9,2,6,5,3,5,8,9,7,9,3,2,3,8,4,6,2,6,4,3,3,8,3,2,7,9,5,0,2,8,8,4,1,9,7,1,6,9 ,3,9,9,3,7,5,1,0,5,8,2,0,9,7,4,9,4,4,5,9,2,3,0,7,8,1,6,4... then it is in no way a "rotation". There's only a rotation if at some point the ideal casting sequence forms a loop. For shadow priests, that was not the case pre-2.1 (with 6 second SW: Smiley) for fights under 30 minutes, and unless you have evidence otherwise, it is fairly safe to assume that is true for warlocks as well. It's certainly possible to use a rotation anyways -- if you want to alternate Shadow Bolt and Corruption, nothing is stopping you other than that it'll do terrible damage.

 

Deelit70

I guess the hangup is on the term "rotation" then. We shouldn't let terminology get in the way here - I mean to say that there is a superior initial cast order, that minimizes systematic time loss. I don't understand what relevance it is if something forms a repeating loop versus if something is determinable. The fact is that a superior dot cast sequence is within the realm of theorycraft.

The dot "rotation" naturally flows from an initial cast order. It doesn't create a loop, but it's explainable by a short set of rules, a mundane set of these rules would go like this:

1) The best initial cast sequence is A,B,C,D,E
2) If a dot has ended or is about to end, that dot should be recast before any other spell.
3) If you have more than 2.5s before a dot runs, you should cast Shadowbolt.
4) If at any time you only have 1.5s before a dot runs, you should cast lifetap, unless the dot running is immolate or UA.
5) ...

These rules could get as complex and refined as you wish (applying rules for 1s or .5s before a dot runs, adding more conditionals, etc), but I'm willing to bet that the first 4 right there cover 95% of the cases for all of warlock casting if you are trying to max dps. Referring specifically to dots now: taking rules 1 and 2 together, you can map out your entire dot cast order, and while it may not be a loop, it will give you the rotation of key presses for most (except for the extreme end) of the fight.

The reason blizzard prevented addons from casting for you was because you could take 20 rules like the ones I had above, let the addon choose your spell, and just spam one button without thinking for max dps. Just because we have to do real-time thinking ourselves now doesn't invalidate the rest of the rules...

I don't see what the difference between needing a "loop" is as opposed to being determinable, unless you are attempting to macro everything to just one button. Even then it is partially solvable, because there is almost no time ever where you cast UA and shouldn't follow it with Corr (unless the fight is ending in under 30s, you are better off Lifetapping and being without the dots for 1.5s than splitting them up with a lifetap inbetween).

 

Rasputin

The difference is that a cast sequence is not an algorithm. It is a repeatable series with no variation. You describe a few logical rules which are the framework for affliction warlock DPS. Following these rules does not mean you have a cast sequence on the order of 5xScorch, then 7xfireball, 1xScorch and repeat(I think those are the numbers, I've never looked into the length of the Fire Vulnerability debuff). One is a simple invariant series which you can actually macro if you so choose. The other is a set of rules into which you must input various pieces of real-time data in order to determine the next cast. You can determine a priority order such as: If a DoT has expired or will expire within the cast time for that DoT then reapply, else shadowbolt. But you do not have a system which says: Corruption, CoD, Siphon Life, UA, Immolate, Shadowboltx3, repeat.

Out of curiosity, has anyoen tried to math out the ideal, no-lag, perfect cast synchornization cast sequence? I remember drawing one up for SM/Ruin using CoS in vanilla WoW and even that was something like 90 seconds long before it repeated. I can only imagine the heinous sequence before you saw a repeating pattern if you included CoA/CoD and UA.

 

◊ Culok

The reason the difference is relevant is that a cast sequence will always have one and only one answer to "what should I cast next" whereas the priority algorithm can suggest multiple solutions at any given point in time. Given a long enough fight, an Affliction warlock will hit multiple points where there are 2 or more recommended actions.

Another reason that it's relevant is that for a cast sequence, the best way to do dps for a long fight is to repeat your opening sequence regardless of fight length. For a priority algorithm, the solution to "what should I cast to max dps in this fight" requires you to know in advance how long the fight will be since the non-repeating cast order means that dps level will never be a constant. So you can't even say "this is the best opening order."

Given that it is a priority algorithm and not a random event, there are some actions that are always non-optimal but, given the nature of the problem, being able to determine that a particular solution is wrong doesn't always mean that you've moved closer to the best solution.

 

kronchev

Having a bit of a scuffle with an up-and-coming mage over this

I tried looking at the thread but it seems to be really back and forth, depending on personal preference.

He says that arcane is the best for raids right off the bat. He is 63 and will be jumping into kara as soon as he is 70. I am his raid leader and I am telling him I won't accept anything any 10/48/3, based on this obviously. He counters that the thread says arcane is fine, and we are just going back and forth.

So let me rephrase it a little clearer: Does this hold for new raiders in Kara, those who do not have T5 or even any epics? I told him once he gets geared, I will allow him to start messing with specs, but right off the bat, I feel that 10/48/3 is a tried-and-true DPS build he can rely on.

Elitist Jerks forum: Now work safe, because you don't want to get caught wasting your company's time!

It was a question of how the abilities of the fight are handled. I did not know the answer so I come to the place where I expect to see well formulated, concise and correct answers. Not snotty comments. - eclectic778

Vent is only necessary because of bad players. - ebbv

 

tetracycloide

Let him spec however he wants and when his performance comes up subpar stop inviting him to raids until he respecs. I'm not 100% on what you meant by 'or even any epics' but if it means he will not have access to the tailoring BoP epics or the BoE epics I would say that's a far greater priority than spec discrepancy. If the spec is a problem that can be fixed rapidly and cheeply compared to grinding tailoring to 375.

My vanity is justified.

 

Nite_Moogle

This is better directed to a mage thread.

 

Vrangsinn

I'm editing this out until I find the proof I'm looking for.
Moderators, feel free to delete

 

Machinator

I am looking for information on the theory that special melee attacks use a 2 roll system. Its referenced in the Working Theories of Theorycrafting thread and a couple other places but no links. I tried the search but couldnt find anything. I believe it was proven for rogues but I cant find the evidence for it or other classes.

"Information is ammunition."

 

Apate

See you, auntie.

 

Hate Monkey

I know that this is widely believed, but it's been bugging me ever since people stated it to be true. Been going through LOTS of WWS reports for various fights, mages only, and for every Fire mage with 30 penetration, I'm seeing ~3% partial resist rate. But for mages with higher than 30 penetration, mainly 35-50, I see 1% or less partial resist rate. Going to make a compiled list here soon of about 50-100 mages for each fight that doesn't "violate" the rule.

 

galzohar

I was reading theories that the miss is on a seperate roll, so you may want to test and verify it. After all it's only a few % difference and definitely a non-issue for warriors but it may make pallies crushable if they're "on the edge" and relying on misses to add to the avoidance needed to be crush immune.
If it's already tested, ignore this post

 

Nite_Moogle

There isn't anything in the original post that hasn't been tested. The theory that the miss is a seperate roll is bunk.