[Priest] State of Shadow - 4.2.0

[Arlee]

I'd also like to point out that if RNG was a huge factor in our dps then the people who get top rankings would vary a lot depending on who happened to have good RNG that day. That's not what is seen though... the people who get top rankings are often the same people repeatedly.

[Worshaka]

Doing 120k on one MB is what % of average damage in a fight? (and yea it can hit that hard I am giving you a best case scenario here). Lets take my totally mediocre total damage done on shannox tonight of 7,786,682. That means if I had a MB that would have hit for 120k and missed then I would miss out on a total of 1.5% of my dps there.... you are right that is huge. Also, that is completely not taking into account that if you are stacking more int/haste/mastery for the hit you are dropping then your spells are all hitting harder all around which also mitigates a miss on MB.

As far as hit always sim'ing out of the ballpark compared to the other stats... no this is not always the case. It depends on the variables you select and what kind of fight style you sim with. I have seen hit/spirit be equal/slightly below mastery.

I am not sure what your point about missing a MF is. Suddenly you are out one GDC if you miss it, but that's not a huge deal either. Also you should be clipping MF for ANY other spell that comes up.

For more information about this read Priest Hit Cap FAQ.

Also last thing, current number 1 ranked shadow priest (#2 overall dps) on shannox http://worldoflogs.com/reports/rt-qo...&e=5153#Velran take a look at how many times he missed on that fight.

Haven't been on the forums in awhile but seems you didn't get the point I was making:

1. Why people try to compare Int with secondary stats bemuses me. Int is effectively a value derived from your iLvl and is mutually exclusive to that of secondary stats. You don't get items that have varying about of Int for more secondary stats, the only way you can nerf your Int is by not using Int gems. As such discussions about secondary stats should stay with other secondary stats. Making a comparison about gaining hit at the expense of Int is a pointless statement (you could only do this with gems and we can all agree that's something you should never do).

2. I was talking about clipping MF with another MF and yes I'm well aware you clip MF for all spells but you do so smartly (just after the next MF tick). The reason we clip MF with MF is to remove reaction lag, the time between a spell ending and the next one starting. As such all of us will recast MF within our lag + reaction time of the last tick. You can think of this as we clip the last tick for a fresh cast of MF. If the new MF misses what have you effectively done? You have removed the 3rd or last tick from your current cast and added a GCD's worth of time meaning you can actually lose out on 2.5 ticks of MF for every MF miss. Of course if the spell doesn't miss then the MF tick interval is not widened and this is the point I was making. Miss doesn't just hurt MB & SW, it severly hurts MF as well, but this is something priests don't seem to think about.

Just off topic, someone linked a hit cap faq where a statement made was misleading. Basically it was pointing out if a 3 orb MB was to crit that missing that spell would be bad. The combat table doesn't work that way, crit is exclusive to hit... there isn't a roll to see if a spell will land and then if it will crit. It either hits, misses or crits, so increasing your miss doesn't have any relevant to crits.

[Deleted]

This Mind Flay clipping obsession amuses me. When 99% of Shadow Priests can't keep their DoTs up properly, why even worry about clipping Mind Flay? Stop focusing on that and maybe you'll be able to keep your DoTs up better and that will be a higher DPS gain in most cases.

I'm sorry, but if people who are doing low DPS just focused on keeping their DoTs up and Mind Blasting on cooldown, they would be very competitive DPS wise. Instead they focus on breakpoints, Mind Flay clipping, hit cap or not, Mind Blast with orb or not. Sure, these things are good to know but if you are not good enough to keep your DoTs up, you shouldn't be focusing on these things.

[Deleted]

Clipping MF is easier than not clipping, so...

[Deleted]

Clipping MF is easier than not clipping, so...

Not caring about how you clip or not clip Mind Flay is easier than sticking to one thing, so...

[Deleted]

But then, not caring to use dots at all would be easier too
Some things give greater dps increase than others: clipping is very probably greater dps increase than switching couple almost but not optimal gems to proper ones or switching some reforges to different stats or other gear fine tuning. So dont see point downplaying it.

[Kilee25]

If you are bad at clipping MF, don't do it. You will lose dps if you do it poorly. The mod Gnosis can help tremendously as it models the channel very well.

There's not a lot of DPS to gain by cliping MF to cast dots, because of the large-ish window you have to refresh them. There's generally a 2 second (or more) window between the second to last and last tick to refresh dots. It's a fairly large window, so you should be able to get good up time on your dots no matter which casting method you use. If you aren't getting good up time, then you need to try to develop better planning for spell sequence. You should be planning each spell you will cast at least 1-2 spells in advance.

The best reason to learn to clip MF is to cast MB earlier. You should see roughly a 200 dps increase from clipping early for MB (more with the t12 4-piece bonus), so it is not trivial, but you could very easily get a neutral change in dps if you clip poorly.

You can continue a channel of Mind Flay by recasting/clipping a Mind Flay on top of another Mind Flay. Doing this will create a 4-tick Mind Flay. Again the mod Gnosis should model this correctly. Casting like this can reduce downtime between mind flays by taking out the latency between re-casting. It also takes out the negative effects of being spell-interrupted during casting. (Any interruption during a MF results in losing the last tick, but you can continue the channel and not lose the tick if you re-cast MF quickly enough.)

Anyway, I would suggest going to a target dummy and practising the casting techniques to get the hang of them. Good luck.

[Deleted]

In fact if I remember correctly, doing clipping randomly is still dps increase. Personally I see biggest advantage being how easy it is to chain MF's by clipping them after each others. Before MF was changed to dot like spell, it was always pain to try cast new MF just right time, including latency and trying to not lose last tick. Wouldnt be surprised if in real boss fights MF chaining being biggest advantage clipping gives.

[Felade]

1. Why people try to compare Int with secondary stats bemuses me. Int is effectively a value derived from your iLvl and is mutually exclusive to that of secondary stats. You don't get items that have varying about of Int for more secondary stats, the only way you can nerf your Int is by not using Int gems. As such discussions about secondary stats should stay with other secondary stats. Making a comparison about gaining hit at the expense of Int is a pointless statement (you could only do this with gems and we can all agree that's something you should never do).

Gems sir, Gems. Int directly competes with secondary stats while gemming. It may seem obvious to us grizzled veterans, but it is important (especially for players new to DPS) to note that gemming for secondary stats comes at the cost of Int, which isn't true for reforging. Thus, knowing Int is better than all the secondary stats is very useful to know (as well as just HOW much better it is) when you are choosing your gems.

[Deleted]

Gems sir, Gems. Int directly competes with secondary stats while gemming. It may seem obvious to us grizzled veterans, but it is important (especially for players new to DPS) to note that gemming for secondary stats comes at the cost of Int, which isn't true for reforging. Thus, knowing Int is better than all the secondary stats is very useful to know (as well as just HOW much better it is) when you are choosing your gems.

Not only this. I'm tired of telling my raid mates who skip rolling for a dropped higher item lvl item because "I don't like their secondary stats". Man, if anyone else needs this item because it is in his/her BiS lists, I understand that you don't roll for this item. But if it's gonna be disenchanted, a 391 item is always better than your 365 and even 378. Because the difference between haste, mastery and crit is nothing compared between your secondary stats and your main stat, in our case, Intellect. So, the little dps or hps you lose because of the undesired secondary stats of that 391 item is nothing compared to the dps or hps you gain with the extra Intellect you get.

[Arlee]

Most often the conflict between int and secondary stats comes into play when someone is looking at upgrading to a higher level item that doesn't have hit/spirit on it and would drop them below hit cap.

@Worshaka- Your point about missing a MF is that you miss a whole MF... which yea hurts but is honestly barely noticeable especially when you consider that often full MF's aren't even a cast. I am not saying you shouldn't be concerned about getting to hit cap and keeping it, but worry about missing a MF (or any one spell actually) is pretty darn'ed moot, especially if you are really close to the cap in the first place. Also, that part about how there is one roll for hits, crits, and misses is completely wrong. That is how it is done for melee; for spells first the game rolls to figure out if you hit, then it rolls to see if you crit.

[Worshaka]

Discussions blow out of proportion quickly... let me try to be a bit more concise.

The discussion I was entering into was the value of Hit and the reasons why it is valuable. In the context of that discussion, any talk about how hit will some how be a deteriment to the amount of Int you can get is invalid. This is largely because Int is derived by iLvl and the only way to gain hit at the expense of hit is to use non Int gems.

As such my discussion relates to the make-up of secondary only stats, therefore any comparisons on the value of hit need to be relative to haste/crit/mastery. Trying to derail the discussion to anything else than this is pointless.

There is a misconception that hit doesn't hurt us that much due to the fact that DoT's can be recast instantly, they only become subject to hit on cast and that our nuke damage is a relatively small percentage of our damage makeup. They all sound like compelling arguments but I was trying to illustrate the flaw in that argument.

Most people will use the MB/SW/Imp DP as the only sources of nuke damage with MF being lumped into the DoT category. However MF really is a hybrid DoT/Nuke. One of the good things about MF is that if you do miss with it you only lose a GCD's worth of time with MF actually taking 2 GCD's to complete. This is much better than say shadowbolt because if it misses you used the full amount of cast time to find out that fact.

Now the thing that is ignored with MF as I tried to explain is the purposeful clipping of MF with a new MF. Now I'm amazed at some of the replies saying this is pointless, too much time is spent explaining it and don't do it. Clipping MF with your next MF is one of the best ways to increase your DPS. I could write paragraphs of info explaining how channeled spells are implemented in the client/server environment but the short of it is, you can't spell queue channeled spells without a) clipping the current cast b) adding in large amounts of server lag between casts via no-channel macros.

Just to be clear, if I clip MF with another MF I do NOT delay the tick intervals. That means if the 3rd tick of mf is due to happen 0.2 seconds from now and I recast MF the next tick still occurs 0.2 sec from now, exactly the same mechanic as our true dots.

The situation I was trying to explain was where you clip MF with a new MF but the 2nd MF misses. What is the effect? The true effect is that the previous cast which didnt miss is arbitarily finished earlier and suffers the loss of a tick. The MF you just cast prevents you from casting for a GCD. And your next MF cast has it's tick timer intervals reset to 1/3 of the cast time. So in essence instead of getting a MF tick every 1/3 of your MF cast time nice and regular you have prevented casting for a near a full MF channel. I'll even be nice and only call it 2 ticks but it probably is worth 3 ticks to the average player.

Therefore miss is still very impactful on your MF casts which make up 30% of your damage...

Now to suggest that missing a MF isn't a big deal, in isolation probably not. However that isn't the discussion. The discussion is relative to, if I have a 1% miss chance what do I get instead... you can get less than 1% haste, something around 0.5% crit or less than 1 mastery. Now the real question is, if I get one of these other stats will it do more damage on average than the impact of missing 1% of my spells. The answer to that is no and one of the reasons it's no is to do with understanding how miss affects MF.

---------- Post added 2011-10-13 at 02:09 PM ----------

The game doesn't 2 roll for spells, it does 2 roll for some special melee attacks as was seen with sinister strike and the old hit debuff that Attumen gave players back in Kara. Spells are most definitely single roll, otherwise if you had a 30% crit rate but missed 10% of the time your crit rate would really be 0.3 * 0.9 = 0.27, and that would increase the value of hit quite a bit.

[CaseyTheRetard]

If you are going to contradict the conventional wisdom that spells use a 2-roll system, the burden of proof lies upon you to demonstrate otherwise.

[Arlee]

Worshka... if you have some proof that there is not a two roll system for spells please by all mean present it. However, currently most evidence points to a two roll system unlike the one roll for melee. Your example is complete conjecture. If you really believe this then do some testing and prove it. Until that time, don't tell people that it has changed with out anything to back it up.

[Kilee25]

In general, classes with direct target (nuke) spells have a much higher hit weight than classes with dot spells.

In simulationcraft, you can affect the value of hit rating by making simcraft refresh dots earlier. At current, simcraft refreshes dots at the "last possible moment". It does this because DPS goes up in the sim when you make the action list do this. However, if you change the action list and tell it to refresh dots at the "earliest possible moment", the value of hit goes down significantly, far below that of our other stats.

This to me basically means that the biggest deciding factor in hits value becomes our up-time on dots. If we refresh dots early, and then quickly re-cast the spell if it misses, hit becomes less important. Even though we do have direct target spells like Mind Blast, SW: D, and Mind Flay (if you count it), they don't impact hit as fiercly as it would a completely direct-target class like a mage.

If you are trying to approach this from a purely scientific angle, as you appear to be, you're fighting an up-hill battle. In general I tend to feel simcraft is fairly accurate when it comes to stat weights. Sure there is a margin for error, but I trust how it behaves. If it's showing hit as a mediocre weight (which it frequently does for shadowpriests), then I believe that to be true and the way that it behaves for our class.

If you want to convince me that hit is more important than simcraft says it is, then I think you should go about this from a different angle. For example, you could argue that a human being simply lacks the reaction skills to ever achieve the level of expert refreshing that simcraft is capable of, and thus hit is in actuality far more valuable than simcraft says that it is. I could certain get behind an argument like this.

[Worshaka]

Worshka... if you have some proof that there is not a two roll system for spells please by all mean present it. However, currently most evidence points to a two roll system unlike the one roll for melee. Your example is complete conjecture. If you really believe this then do some testing and prove it. Until that time, don't tell people that it has changed with out anything to back it up.

I have been playing WoW for a very long time, pretty much since the gates of AQ first opened. When researching and learning mechanics long ago I was under the impression that WoW used a 1 roll table and that this was the 'coventional or accepted' belief. Doing some very quick reading it appears that belief has either changed or I got the wrong impression long ago... however I really didn't see anything conclusive but it's generally accepted by most people that it's a 2 roll system.

I do remember reading a proof about a 2 roll system back in TBC with rogue special attacks and the Attumen hit debuff. I can't remember the exact figure but if attumen gave a 75% hit debuff and your crit rate was 30% then every special attack would have 2 outcomes on a single roll system, crit or miss. However this wasn't what was happening. As it turns out melee white hits are on a 1 roll table but special attacks are on a 2 roll table and I was under the impression at the time that melee special attacks were the only exception to the 1 roll table.

In any case I retract my statement... if spriests had higher crit rates I would do an experiment myself but to get anything reliable you'd need quite a large amount of crit to see a genuine drop in crit rate when hit capped and not hit capped. I don't have a mage alt either otherwise i'd do that... but basically the experiment would be to go with the hit cap at say 50% crit and do 10,000 casts then do the same with 0 hit (17% miss) and if it's a 2 roll system your crit rate should drop from 0.5 to 0.435.

[Arlee]

It's actually sort of a common misconception especially from people who switch from a melee class to a spell dps class (looks at self in the mirror) because melee classes do use a one roll system (with the notable exceptions that you mention). It's a bit odd I know. The other thing that would happen is that on a 1 roll system adding enough crit would make your miss chance lower because more of the "roll chart" would be taken up by crit. I remember back in BC some warriors were stacking more crit instead of hit because of this.

[Worshaka]

In general, classes with direct target (nuke) spells have a much higher hit weight than classes with dot spells.

In simulationcraft, you can affect the value of hit rating by making simcraft refresh dots earlier. At current, simcraft refreshes dots at the "last possible moment". It does this because DPS goes up in the sim when you make the action list do this. However, if you change the action list and tell it to refresh dots at the "earliest possible moment", the value of hit goes down significantly, far below that of our other stats.

This to me basically means that the biggest deciding factor in hits value becomes our up-time on dots. If we refresh dots early, and then quickly re-cast the spell if it misses, hit becomes less important. Even though we do have direct target spells like Mind Blast, SW: D, and Mind Flay (if you count it), they don't impact hit as fiercly as it would a completely direct-target class like a mage.

If you are trying to approach this from a purely scientific angle, as you appear to be, you're fighting an up-hill battle. In general I tend to feel simcraft is fairly accurate when it comes to stat weights. Sure there is a margin for error, but I trust how it behaves. If it's showing hit as a mediocre weight (which it frequently does for shadowpriests), then I believe that to be true and the way that it behaves for our class.

If you want to convince me that hit is more important than simcraft says it is, then I think you should go about this from a different angle. For example, you could argue that a human being simply lacks the reaction skills to ever achieve the level of expert refreshing that simcraft is capable of, and thus hit is in actuality far more valuable than simcraft says that it is. I could certain get behind an argument like this.

I think simcraft is a pretty reliable tool too but it's not perfect and as any researcher would know you don't trust simulations implicitly. They help you make informed decisions about things but when a simulation gives you a result you almost always try to understand what is causing that result.

As for simcraft, you only need to read a few forums to see the vast array of differing stat weights that all conflict with each other. I've seen stat weights that show hit to be very high, very low, the same as haste, below haste but greater than crit/mastery. I've seen mastery as the highest secondary, i've seen haste as the lowest.

So when you say you trust simcraft my question is which result? Changing it from patchwerk to helter skelter and not changing any other variables can dramatically change the stat weights. Does helter skelter truly measure the encounters that are currently available? Do the stat weights work for bosses 1 3 & 5 but for the other encounters the stat weights are wildly different?

To me just trusting these figures without putting some sort of human intiution and experience is a bit silly. From all the simcraft data I've seen I see hit close enough haste as our best secondary stat that I conclude it's worthwhile reaching that hit cap provided a) you dont compromise int and b) you try as hard as possible not to compromise your haste. But not to hit cap in order to get more mastery/crit I think is a silly approach.

I have provided reasons as to why I think hit can impact our dps quite adversely, I don't know of simcraft factors is MF clipping for new MFs or whether it assumes you perfectly cast them back to back with no clipping. You also raise a good point about recasting... the way I play it's almost impossible for me to react to a missed dot perfectly because I spell queue as much as possible. If I cast VT during that cast I'm spamming my next spell, when I hear the warning that VT missed I'm already casting my next spell so invariably if I miss a DoT it doesn't get put back up for a GCD's worth of time.

VT is a bit of a nasty example because the spell lands after the GCD, but even with SW:P and DP (with my aussie ping) reacting to a miss is very hard, I've usually already spell queued the next spell.

The other thing I find reasonably ammusing is that with 378 gear, which most people will be in or close to, it's pretty hard not to be near the hit cap. So I can only surmise people are reforging away from hit or not reforging 100 or so rating to hit the cap. At the moment I'm 1% under the cap but I've not really been raiding and playing quite casually the past 6months. However 1 or 2 more 378 items and i'd have to start reforging out of hit not to be over. I think im in 374 iLvl gear... im sure a lot of people are a lot higher than that.

[Deleted]

I don't know why you say Simcraft says hit is a mediocre stat. In my simcrafts, hit is below spell power and above haste and I am already hit capped.

[Kilee25]

re: Eldarc. Sometimes it simes high, sometimes it sims low. The behavior I described still happens.

re: Worshaka: What I'm talking about is that for mage type casters, hit often sims between .7-.9. For shadowpriests, it sims between .5-.65. Both can sometimes be higher or lower. I then went on to say the reason for this is because we are a dot casting class instead of a direct spell casting class. And yes, I consider .5-.65 compared to .7-.9 to be "mediocre". The truth is that losing hit does not affect us nearly as much as many other classes.

If you're just trying to make a point that we should try to hit cap. I'm with you. If you're trying to say hit is much more valuable than simcraft "generally" ends up being, I'm not so much with you.