galzohar

It's not sometimes, it's just depending on what you're looking at. If you're looking at burst, then yes I'm ignoring it. If you're looking for efficiency, I'm not ignoring it, but it's not the most powerful stat.

The reason it's definitely less than 50% is that obviously you will overheal more with crit than you'd overheal without crit, while doing 1.5x healing every time you crit (mana returns from illumination is calculated already, it's not being ignored either). So with say 20% crit, your heal will land on average for 10% more than it would with 0% crit, but will also overheal more than your average heal overheals for. This is because you *try* to heal people based on how much HP they're missing based on your non crit heal (wether you succeed in practice or not doesn't matter), and the fact you try that is already enough to force some level of bias against crit, as in assuming it heals for less than +50% on average on every crit. Of course sometimes you heal someone who needs more than your actual heal heals for, so you'll actually get effective healing from it, therefore crit doesn't go all into overhealing.

Where exactly it's at between that 0% and 50% I'm not sure, but even assuming it's close to 50% doesn't show crit as an amazing stat. And considering the way I play and the way people posting here claim they play, seems like at least a significant portion of my and their crit healing goes to waste. For the reasons that were already brought up (such as the fact your normal heals also overheal, and the difference between fights/players/situations), finding the exact or even a decent estimation of how much of the crit actually goes to overhealing is very hard. But I think I brought enough evidence for why it should definitely be noticeably under 50% and noticeably above 0%, for efficiency calculations.

The claim I make of crit giving nothing for emergency burst healing is a completely different argument.

 

Stolidus

I'd like to chime in on the effect of spell haste really being noticeable because of the controlled burst capability it adds and simply for the fact that it makes your errors more forgivable. Throwing a FoL to a random dpser on kale feels so much safer when the round trip back to the tank with an HL takes 3.2 seconds instead of 3.5.

While I haven't gone beyond gemming yellow sockets with it so far, I'm going to toss quick lionseyes in blues as well if the socket bonus is bad, and I'll probably use them in any future red sockets as well (partially due to the spinel scarcity). Spell haste is, in my opinion, just an amazing stat because in my experience tank deaths do not occur due to insufficient +heal or mana problems, but the heals just not landing (yes, this is presumably because of healer error, but you wear extra stamina to reduce dps death error too, so I don't see any intrinsic failure with the logic of reducing the consequences of healer error). Those deaths where you felt like you just couldn't react fast enough due to ui lag and the tank died after your heal completed can potentially be averted with spell haste, but not with +heal, mp5, or crit.

The way I see it, the stats do the following things to keep a tank up:
mp5: Not running oom=heals
crit: Tops off tanks much quicker than haste could occasionally, reducing the chance of the next spike happening. Also, helps with efficiency less dramatically than mp5 (this I think has been proven enough)
+heal: Can certainly make a difference in the raw throughput of a single heal, similar to stam stacking for tanks
haste: Can make your heals hit on time, with less of a period between for spikes to kill your tank

From there I draw my conclusions as to which will help me reduce the most tank deaths, and for me that's +haste.

 

Fielding

Definitely not trying to derail the thread, because I'm still quite uncertain of how to value crit, but this is something I've posted on my guild forums and something I'm trying to figure out for myself.

Thanks to arena points/personal ratings, I don't really have any money concerns, so the money spent leveling alchemy, etc is completely irrelevant.

Here is the post:




Trying to decide what 2 professions to use while we progress and clear through sunwell.

Currently I'm Enchanting and Jewelcrafting, but I've been considering Alchemy.

Here are the benefits of each over status quo/non-tradeskill best of the best.

Enchanting:
-> +40 healing (enchanting your rings versus scrub non enchanting rings)

Jewelcrafting:
-> +4heal, +2crit (BoP Epic gems vs BT/SW/lolbadge Epic gems)
-> +16heal, +2mp5, +3int, +5haste ([Amulet of Flowing Life] vs [Brooch of the Highborne])

Total Jewelcrafting Benefits:
-> +20heal, +2crit, +2mp5, +3int, +5haste

Alchemy:
-> +7mp5 (taken from following excerpt)


WHAT I'M TRYING TO FIGURE OUT:

Before I can really compare the 40 healing(Enchanting) to the 7mp5(Alchemy) to the 20heal, 2crit, 2mp5, 3int, 5haste, 100lawl(Jewelcrafting), I need to determine first that the Glimmering Naaru Sliver actually exists and second how the effect being channeled effects it.

Since it is channeled you lose 8 seconds of doing anything else... what does this cost you and more importantly when can you fit it in.... I'm just not sure how I feel about it yet. I'm a lot more partial to passive abilities, and I consider alchemy ability passive, because potting isn't an option sometimes.

I'm trying to think of encounters that I would be able to think to myself, "oh here I am with a 2000 mana deficit... time to pop my SLIVER MAN" and drop healing and channel for 8 seconds. Illidan moving in to phase 2 maybe and of course it would be pretty amazing for arena, but I dunno how useful it is otherwise for a Paladin.

If the fact that it is channeled makes the mp5 gain higher than 7, closer to like 20 or even more, then I think alchemy will pull ahead of jewelcrafting for min/maxing.

*Note: I'm not actually considering replacing Enchanting, because the gains are independent of gear, whereas the Alchemy and Jewelcrafting gains are strictly this patch, sunwell top end loot dependent.



Sorry if some of that seems a bit silly, odd, but it was for my guild forums/didn't original type it to post here lol. If anybody has any insight please let me know =))

Thanks

 

Endoscient

One thing to note about alchemey trinket is that most people won't get the Muru trinket until after kil'jaeden is daed. Which is ultimately what we are gearing for. You also don't need to channel it which can be a huge deal on some fights. Highly unlikely on Brutallus I can user Sliver anywhere near enough for 33mp5.

You also forgot about Leatherworking with [Drums of Restoration] 25mp5 for each person in your party (and they stack).

 

Silabiss

Ok, someone help me out here. I'm trying to make sure I'm not missing something simple. This goes back a few pages, but I've seen it referenced a few times since then...

The idea that they would only be casting ~100 ~2 second cast time spells in a fight that lasted 360 seconds (so only 200 seconds of cast activity) on one of the highest burst fights currently in game seemed very odd to me, so I went and took a closer look at the WWS.

Looking at the WWS, it's clear that the two bears (Sammy and Azzie) are the tanks. Now let's look at the report for abilities used for one of the 3 pallies, Fenira for example. According to this list, Fen cast only 98 HL, as Palados had mentioned.

However, if you look at the Breakdown for Fen under "Heals Out", it's stated that Fen hit Sammy 91 times and Azzie 80 times. If we take these two together (91+80=171) and assume a cast time of ~2 seconds, then we come out with a total cast time of ~342 seconds for these casts alone (Fen actually had a heal or two cast on himself in addition). This ~342 out of 360 seconds seems far more likely to me than having only spent 200 seconds casting, and also makes more sense in covering the damage. It really is far more likely that none of the pallys had no down time than that they had 40+% down.

Am I misreading the WWS somehow, or is there really an issue with what's stated on the abilities page? Or is there potentially something else here that I'm overlooking...

 

◊ Shalas

The "hits" column doesn't include crits, which makes it mostly useless and very misleading. If you expand the Holy Light row, you'll see that Fenira had 98 hits and 73 crits, for a total of 171 HLs cast (342.5 seconds with zero spell haste and lag).

 

Silabiss

Yup, that explains it then. This makes far more sense, encounter wise.

 

Endoscient

I don't agree with that logic at all. If I always cast heals in estimation of fulling there health up exactly, then yes crit would be quite crappy for increasing healing done. But how often when does that happen in tank healing. If my tank is down around 5000 health (the situation where you say crit overhealing is the most noticeable) I will not use a Holy Light, except if maybe I am the only healer on the tank and only then if he can take a lot of burst. When tanks have 25k health, being at 80% health is not cause for a Holy Light. Especially when my average FoL hits for 2600 (with crits, 2300 without). Most of time I use Holy Light on the tank is when he gets closer to 8k health down. In that situation a Holy Light crit won't really overheal at all. If I overheal then it is mainly wether or not someone healed the tank before I did.

Sure, non crit Holy Lights overheal as well. But you also get a lot of crits that do 8500 pure effective healing. So you need to empirically analyze data on how many of your non crit Holy Light spells overheal versus how many crit Holy Lights don't overheal before you can argue that point.

This isn't really true for raid healing at all. But the almost all of the time I am tank healing, so that is what I gear for.

 

galzohar

I never claimed I know exactly how many of the spell crits effectively add healing. I just said that the total increase in healing done is less than crit/2 and more than 0, for the exact reason of what you just said - you don't always get heals wasted when you heal nor do you always use them fully. On average it would be somewhere in between, and for efficiency what matters is the average.
Anyway remember I did the "max possible" value of crit assuming you get most of the crits healing as effectively as non-crits and crit still wasn't a stat worth stacking mostly (although at that max value it gets more comparable to argue that it's dependent on some factors). However I don't think it's anywhere near the full 50%. Even when FoLing a tank you're usually just topping him off along with the other healers, and while it's hard to say how much overhealing it adds versus effective healing, it definitely adds more "overhealing per effective healing added" than +healing/mp5/haste add. How much I'm not sure, but it's most likely not something neglicible.
If you find a way to actually give a good estimation of crit overhealing I'd love to see it.

Regarding the WWS, if he casted 98 hits and 73 crits for a total of 171 HLs, he spent 143640 mana. While he did have every mana support possible (which generally would mean you don't care about efficiency much so all efficiency calculations are nearly moot anyway), I still don't see how he had mana to cast all of this.

Illumination 29,772
Vampiric Touch 21,545
Spiritual Attunement 8,107
Restore Mana 8,116
Drums of Restoration 4,560
Mana Spring 3,925
Mana Restore 3,900
Mana Tide Totem (Mana) 2,226

Totals to 82151 mana, which means he started the fight with 61489. So either he regened mana from sources WWS doesn't record, in which case he had amounts of mana that are more rediculessly high than what you'd normally expect from having SP+shaman+pots which would mean he doesn't really care about efficiency so efficiency calculations don't hold, or he simply didn't cast that many holy lights. Either way there's a problem with WWS, and either way what actually happened in the fight doesn't contradict anything I've said so far.


I'd also love to see a way to calculate how faster heals (while keeping HPS constant) help your efficiency as well as your burst HPS - as they obviously help but the amount in which they help is quite hard to put into numbers for niether efficiency nor burst healing. That is how much burst HPS would you give up for having your cast time shortened by X (assuming that's an HPS loss after reducing cast time)? How much efficiency are you willing to give up for reducing your cast time by X (ignoring burst HPS benefits of it)?

 

• goss

Generic mana regeneration (mp/5 or spirit) is not recorded in the combat log and thus will not show up on a WWS.

 

galzohar

With 15k starting mana then he'd need 645 mp5 over the 6 minutes. The only way to explain is if he wasn't casting max ranks, which means the HPS->mp5 conversion works just fine except you need to calculate using the minimum HL rank instead of FoL, and that's assuming you really can't cast a single FoL during the fight and he wasn't just using downranked HLs instead of FoLs as a way to burn extra mana for some extra safety (which would again mean you don't really care about efficiency but more about HPS).

Bottom line is no matter what situation that paladin from the WWS was in, that data doesn't contradict any theorycrafting I've explained so far.

 

• goss

Paladins on Brutallus generally vary between R7/R9/R11 HL, and perhaps R1 for LG on occasion, so thats definitely a possibility.

 

galzohar

If you can afford a R7 HL are you really unable to ever afford a FoL (resulting with more mana to sparingly HL11 when needed)? The HPS difference isn't really big while the efficiency loss isn't small.

Anyway the bottom line is you're never spamming HL11s therefore my efficiency calculation system has use when efficiency matters.

 

• goss

Keeps LG active while maintaining a similar hp/s to FoL.

 

Endoscient

Sorry, the first time I read your post I thought you meant 50% of the gained healing from crits (ie 1250 of a 7500 crit). Yes, the value of crit overheals is between the min and max, that is all you proved.

Why do you say that when spamming FoL on the tank that +heal adds a lot more then +crit that it is worth more the Illumination mana no matter what? With 2300 average Flash of Lights (which I can get with 2750 buffed healing and ToL in tank group) 1% crit adds 12 per FoL on average, while 45 healing (aproximately equal in item points) adds 22. The crit also gives you 3.6mp5 on average. While this isn't good enough to stack it (ie still not gem for crit), it can still effect gearing choices. Like trying to decide if you should replace [Implacable Guardian Sabatons] with [Pearl Inlaid Boots].

I am not saying I know exactly how much it is worth. But that you shouldn't just write it off for burst healing and it as at healing throughput as you say. Yes I know it isn't efficient to stack, but there are lot a gearing choices besides what stat to stack.

 

galzohar

Again, let's keep efficiency and burst discussions seperate, as they're 2 different things in terms of how they're calculated.

I can't say for 100% certainity on your gear (you're welcome to read the other thread and calculate yourself and posting your results with your gear), but with my gear 1 crit rating would've been worth 1.9 +healing if I'd assume full effect, which is way overestimating crit, which makes crit a pretty inefficient use of itemization. If you consider crits actually don't have full effect crit is even worse, for example if I assue crit bonus to be 1.2X (20% extra healing on crits on average), I get 1 crit = 1.31 healing. Note that this is taking into account both the HPS->mp5 benefit of the HPS added by +healing and the HPS->mp5 benefit added by the crit (although that benefit from the crit is quite smaller due to the much lower average heal size increase compared to +healing even if you give crit heals full value).

The illumination bonus from crit is easy to calculate and that's why I didn't mention it. Obviously it's taken into account in the final calculations but the discussion was on the effective healing increase portion of crit, since we (I think) can agree on how much illumination helps your mana (by using (available mana without illumination)/(1-0.6*crit) ).

Again using my gear as a baseline, the badge boots clearly win on non-stamina fights if you calculate it, regardless if you use them with 2 healing/mp5 or pure mp5 gems for efficiency with no shadow priest or 22 healing for max burst and/or effciency with a shadow priest (prefered choice). It's not a significant difference for efficiency though, but for burst the +healing is just that much higher and at the end +healing/haste is the only thing that really affects your reliable burst.


For burst I'm ruling crit out for the simple reasons brought many times before of wanting a reliable heal land that will be as big and fast as possible which will reduce the situations in which the tank has any chance of dying at all thus giving you more ability to heal with less help/backup. Again remember it's small differences it's not like the +healing stacker will do so much better that he'll get all raid invites, but at the end he'll perform better emergency healing than the one using equal itemization on crit.

 

Endoscient

I don't think using your gear as a baseline is really the best idea anyway because it is far from endgame. When talking about stats like crit that effect both burst and efficiency you can't just talk about one because that isn't the whole picture. Crit will look crappy compared to healing because it doesn't increase healing done by as much, and same for mana regen and mp5. Have to consider how it effects both.

You will never see any reasonable value in crit if you just totally rule it out for burst healing and only look at it for only burst or efficiency healing . That is one reason why pallies can be great MT healers is 8.5k 2 second crit heals. Sure you can't count on it, but you can't ignore what happens on 30-40% of our heal spells.

As I said before if you are using high rank Holy Lights on the tank and he still dies it extremely unlikely (you will have to get really lucky with numbers) that any amount of +healing you can reasonably influence would have saved him. It is much more likely that if one of those Holy Lights crit that he would have survived.

How often do tanks you are max rank Holy Lighting, and the heals get to them, die? Almost never for me, something else is most likely wrong if they do. I stack +healing for efficiency reasons (2.3k non-crit FoLs are op) not for worst case burst.

 

galzohar

The fact is even if you did decide your burst is "average" and took crit as a part of it, you'd still get very low value for crit when it comes to burst, since the mana returns do nothing for burst.

Feel free to calculate using your own gear and post results, I'm not going to take every person's armory and do it for them, obviously. The other thread that was linked a couple times has a full explenation on how to do the calculations using your gear.

However what you say is far from what I'm doing. When calculating efficiency, I calculate how much healing I can put out on an X minute fight if I use all my mana by the time the fight runs out, while forcing myself to cast a certain amount of high rank holy lights. When calculating how much crit increases it, I add up the mana retruns from crit resulting in more spells casted, extra effective healing from crits and mana saved due to having higher average HPS. I'm not ignoring any of those factors. I just calculate them seperately and then add them up.

I do keep these efficiency calculations seperated from burst, though, since the way to calculated is completely different and the fights/raid setups where you'd need one much moer than the other are also different, so I'd rather know an item will increase by burst by X and efficiency by Y rather than give it some averaged score which will be pointless if I need burst more than I need efficiency or the other way around.

Bringing up the "a little extra healing doesn't matter, but a crit can save a tank" I'll bring up that "0.5% crit from a crit gem won't cause a crit". That latter shows how the former argument is irrelevent. Just the fact that a stat gives a very small increase obviously does not mean that increase isn't bigger than what other stats give. And in the case of crit since no matter how much crit you stack you will still never be able to rely on crits, you always have to play as if you had 0 crit and therefore your ability to reliably save a tank from dying does not go up with crit, but does go up with +healing. I agree it doesn't go up much with any stat but +healing at least increases it by *something*, while crit does not. Remember it's reliable ability to save the tank not random chance to save the tank. If we wanted random chances to be saved from dying we'd all carry darkmoon cards with 10% chance to self-ressurect.

 

Palados

Well, Gal, you loose one point. Probability to save your tank from death is always a probablity. Adding +heal will increase it in reliable way, while adding +crit will still increase it in non reliable way. Imagine an intensive simulations, that will show you that 20% crit will increase it by 5%, while +500 healing will increase it by 3%. Which one will you choose? I would choose 5% unreliable over 3% reliable, since due to the law of large numbers it will mean after year of raiding I will save my tank more with 20% extra crit than 500 extra healing.

Imagine an analogy: what would you choose, extra 3% to your salary, or 50% probability of getting 10% more salary each month? First one increase your salary in a reliable way by 3% and second - by 5% in a non reliable way. I am sure you would choose second, since averaged it should bring you more money. And you won't choose first option only because 'I have to consider worst case scenario when coin always will be head' in which you will get no increase of salary few times in a row.


P.S. Remember, that casinos build their business (and it's much more serious than 'having a tank die' stuff) on a non reliable numbers. If they would assume worst sequence, no one would ever start a casino to be honest. Therefore I don't think that assuming the worst case scenario is a good way to consider burst healing.

 

Joasuf

To respond to Fieldings question further up is there a reason you have ruled out BS? The plans which drop from the trash give you the best in game chest piece. To compare it to T6 you gain:

+ 153 Armour
+18 Intelligence
-2 Stamina
1 red, 1 yellow, 1 blue socket (instead of 1 blue and 2 yellows)
+6 Spell Crit
- 3 Healing
+3 Mp5
Socket bonus is 2Mp5 (instead of +9 healing on the T6)

There is a minor drop in healing there but that is negated by the higher int (35% of 18 = approx +6 healing) and the fact you can socket a +22 heealing gem and still get the socket bonus if you so wish.

I would also be interested to know where you get an alchamists stone is only +7 Mp5. If I were to use that it would replace a trinket with no Mp5 on it at all so I get the full benefit of +40 Mp5 (thats assuming you dont already have Memento of Tyrande or other trinkets of that level. If you already have such great trinkets why go alch and then have to choose between them when you can have your already great trinkets AND something else?).

I have in the last week finally leveled up a toon to do my mining for me and so have just dropped it in favour of enchanting. The reason behind this is on the whole I am satisfied with my mana regen (yes at times it would be useful to have more but I have always managed to keep healing) while you can never have enough + healing. The other consideration is I also respec prot for MH and occaisionally BT and so the extra enchants I can do on my Prot rings will also be nice.

 

Palados

Let us look at simple model and you tell if you agree with it:


Imagine, that your HL 11 heals for 5k non critted and in a fight that we analyze tank getting less than 10k HP is 'in danger of dying'. Let us assume that probability density for having X HP under 10k is the same and we assume that tank is saved, when heal that lands will bring him over 10k. We also assume that +healing and crit are 0 at start for the sake of simplicity.

At start probability of saving a tank is 50% (it's a probablity of him having more than 5k HP, that for constant distribution is 50%). By adding 20% more crit, probability to save tank is 50%+25%*20%=0.5+0.25*0.2=0.55, thus 55%. Adding +694 to HL (that is equivalent for 20% crit=971 healing) probability to save tank is 56,94%. Thus we see, that for this model probability to save a tank with extra crit or extra healing is about the same. You can say that crit is less consistant, but I can assure you, from probability theory point of view, as soon as boss damage is random/spiky/inconsistant - it doesn't matter. Since all we get is a probability. And all we can say in prev. example - tank will die with 5% less probablity in case of having 20% crit and with 6.94% probability in case we have extra 971 healing. Thats all you can say. When calculation probability you can not make assumptions about best/worst case scenarios. You just treat all scenarious as probable with certain probabilities.

 

galzohar

You're assuming that minimizing the average tank deaths per raid is the goal and you're assuming it's impossible to eliminate tank death in all scenarios that are likely to happen. You're also assuming all bust scenarios are of similar chance. This is not the case, though. In reality you need to be able to handle every single burst - if you fail one you fail them all. So increasing your chance to not fail one and succeeding because of that just means that you will die next time.

If you're in a position where you cannot add enough +healing to make the tank saveable with 100% (assuming mana is not a limiting factor) chance you need to:
-Add more healing/haste gear
-Assign more healers
-Focus

Adding crit while increase your chance to save the tank will not save him and he's still going to die. All fights I've done and I bet all the fights I hadn't done as well, are doable with a neglicible chance to ever running into an unhealable burst in the firstplace.

For example let's say you have enough +healing to handle a block-block-parryhaste-crush combo every time (which is by itself not a very common event). Now you're only vulnerable to something like a block-parryhaste-block-parryhaste-crush scenario or worsel which is a lot lot less likely to happen, the difference in chances for all the burst events equal or worse than the first is a lot bigger than the chance for all scenarios that are worse than the first. If instead I stacked so much crit to even give me 50% chance to stop every possible burst, if that burst happens 3 times in a fight (let's say for the sake of the example it happens on average 3 times a fight), you only increased your survival chance by 12.5% while stacking +healing to the point of being able to handle that "next worst possible burst I can't handle" would actually help a lot more.

The more often the bursts are happening the less useful it is to increase your chance to survive them opposed to eliminate the ones that are most likely to happen. The more healing/haste you have the less fights there are that can throw a burst you can't handle reliably, and you can't say the same about crit.

The rule of big numbers works for you when you have big numbers. Any efficiency calculation can safely use it for making averages and assuming they work every time, since not getting mana now becuase something didn't proc can be made up for later, and over a long time you would get close to the average calculated value.
When it comes to saving a tank, when there are 2 events where 1 is a lot more likely to happen than the other (becuase let's face it - small bursts are a lot more common than big bursts), eliminating the common one is a lot more useful than having a chance to even survive all of them, and this becomes more and more true the more often those events are happening as if you even have 99% chance to crit but lose the tank when you don't crit, if that burst happens a lot of times you're very likely to not make it through the fight and start over. Instead having a slightly (or moderately) bigger heal, even if it's much less than 1.5x size can be enough to live through it every time, or at least when the dice don't roll *really* bad and you get that burst that is not possible to handle without a crit.

Chances and averages are mostly good, but when rolling 1 bad roll out of many rolls means you lose evertyhing it stops being a useful way to calculate things.

 

Palados

Also, one thing to add about efficiency: in 4min teron fight mp5 is much less efficient stat than in 12min council fight and 20min Illidan fight. While crit efficiency stays the same. So for shorter fights (and those are more healing intensive usually) crit gem can outperform mp4 gem in efficiency.

It can be seen from spreadsheet too, when you look which shield is better - Magtheridon one or Felstone Bulwark. For short fights 27 crit rating and 11 heal outperform 11 mp5.

When gear progress, fights are becoming shorter thus increasing value of crit and decreasing value of mp5. That is one more reason why you can see high-end paladins with quite some crit, but never with huge mp5.

 

galzohar

(speaking only of efficiency in this post)
Even on a 5 min fight 4 mp5 is still winning on crit with a shadowprist (without a shadowpriest mp5 gets even way better), but anyway:

1. You can calculate yourself and see exactly how much it is worth with your gear for the given fight duration (although keep in mind I'm not 100% sure how HPS->mp5 value changes when fight length changes).

2. When comparing everything against +healing, while mp5 drops in value (yet stays high in fight durations where mana matters) the more mana you have, crit really stays pretty much the same thing no matter what the fight duration is or what kind of mana support you have. So if for efficiency 1 crit = 1.3 healing for you it'll be pretty much true for any fight duration, while mp5 value fluctuates (but is pretty much always quite the best efficiency stat at least in fights where you would need efficiency).

3. Very short fights will be the fights where you don't really need efficiency, so calculating efficiency for short fights isn't the most relevent calculation. You probably just want max burst ability for those fights anyway and almost neglect efficiency (excpet for obvious cases where an item choice would give a *lot* of efficiency losing very little burst ability).

 

Endoscient

Its a decent chest piece, you gain some regen/crit over t6 while keeping at same healing. If you had blacksmithing sure make it, but not worth picking it up over. I would probably even drop BS to pick up like enchanting or leatherworking or alchemy.

He was saying the Alch stone was only 7 mp5 more the the Naaru Sliver. I highly doubt though in most fight, especially Sunwell fights anyway will be able to channel the Sliver long enough to get anywhere close to 33mp5, I would expect something more in range of 10-15.