This wiki contains inaccurate and out-of-date information. Please head over to for more accurate and up-to-date game information.




Under Periodic Energize Effects it states that BoW happens to regenerate in ticks of 2 seconds as of a certain patch. Does that imply that the spell is bugged and rather than supplying 16 mana every tick, it gives 40 mana a tick? Or prior to this, it never gave the initial 16 mana ticks? (Disregarding talent buff). --Furinaux 13 Oct 2007

Ok, I added my priest mana regeneration work and made an excel graph out of it. Let me know if you guys think it is easy to read or if I should post the tables as well. I'll probably add equations and graphs for the other mana using classes later. --Jundicator 12:56, 28 March 2007 (EDT)

Would be cool if someone fixed the tables so that they are visible to all of us (ie, not black on black) crash

I have given this article a complete overhaul so it better represents how mana regeneration is actually handled internally in the game. Hope it is ok with everybody. Please check that the information is still correct. --Zootsko 01:17, 13 Mar 2006 (EST)

I Don't agree with this formula.

In theory and according to this formula:

((1 spirit) / 5 ) + 15 = 15.2 Mana/Tick (Druid consideration) Mana/Tick = Roughly 2 seconds so 15.2/2 = 7.6 Mana Per second.

Most items are 5mana per 5 seconds, or 5 mana per 8 seconds, so according to the formula, with consideration for the "mana/sec" rule..:

I conclude that +1 spirit >> 5mana / 5 sec (or 1 (mana/second)).

In order for 5 mana/5sec to be better, spirit must be 1/64th or something like that.

This value you maybe accurate at higher levels, but it doesn't seem valid if tryign to compare 1 mana/sec = "X" spirit (with consideration for the 5 second rule).

Help please,

(My first time posting, Write me Thank you.

Ah, but that's if someone only had 1 spirit. Consider the following:

((1 spirit) / 5 ) + 15 = 15.2 Mana/Tick
((25 spirit) / 5 ) + 15 = 20 Mana/Tick
((50 spirit) / 5 ) + 15 = 25 Mana/Tick
((100 spirit) / 5 ) + 15 = 35 Mana/Tick

As you can see, there is decreasing returns for more spirit, because that 15 free is always there.

Each point of spirit only adds .2 Mana/Tick, or 2mana/5sec if I am reading it right. So yes, 5mana/5sec is not worth much, but it is worth something, around 2 spirit I'd say.

--Stfrn 00:15, 30 Jan 2006 (EST)

"A system in which each additional unit of input yields less and less additional output is said to exhibit diminishing returns. "
If you add one spirit it goes up by 0.2 in your example. If you add one more spirit it goes up again by 0.2 so it is **NOT** diminishing returns as it doesn't yield "less and less" additional output; Each additional spirit point adds as much as the first.
--Byo 13:19, 9 April 2006 (EDT)

I divided the 'Mana gained per tick' column with Spirit and arrived at the third column which for all classes have the form: a+b/x, where x is Spirit and a and b are constants. From that it seems pretty clear to me that 'Mana gained per tick' exhibits diminishing returns. Am I wrong? I don't know where the 0.2 value came from?
--Zootsko 21:05, 10 April 2006 (EDT)

That shows constant returns as adding the first spirit point adds to the spirit as much as the very last one you have. ie a graph would have a constant gradient (be a straight line) as opposed to diminishing returns where it would go up quickly and flatten out.
y = mx + c
Mana Regen = 1/5 (spirit) + 15
Your extra column is not the 'extra regen' added by a specific point of spirit; to get the actual value for this you would differentiate (Spirit/5 + 15) with respect to spirit giving.... (1/5) showing that for every additional point of spirit you get 0.2 mana hence it is constant returns.
--Byo 17:22, 17 April 2006 (EDT)

You're right. Differentiating does not account for the constant however - but I guess one have to assume that it is added independently of Spirit i.e. at Spirit = 0, regen = constant. --Zootsko 23:49, 19 April 2006 (EDT)

I think you guys are missing a major point here.

The whole basis of the 5 mana over 5 secs is that it provides you mana no matter if you are chain casting or not. No healing spell casting time lasts less than 5 seconds. With the 5 second rule now put on spells this means that you will never get anything from your spirit if you are chain casting (which will happen to a priest in nearly every MC or high level instance fight).

With that being said ANYTHING which gives you mana while casting is invaluable. The effects may not be huge but even if a few of the mana/sec items only add up to 50 mana / 5 seconds then over a 5 minute battle thats (300 secs x 10 mana/sec) = 3000 mana you wouldnt have had. I dont think anyone would put their nose up at 3k mana if it were in the form of 200 int spread over a few items.

I thought the mana regen forumula for Priests was: 13 + (Spirit / 4) per tick (one tick = 2 seconds). Where did the 12.5 come from?

There is a very easy way to compare mana per second regen and spirit regen. Someone who never casts spells will benefit 100% spirit, but the more they fight, the less mana they would regen when compared to the mana per second item. At some point, those two values will be equal. That's the number you want to find.

Lets say I have two pieces of armor. One regens 8 mana per 5 seconds, the other has 20 spirit. When should I use one and when should I use the other? 8 mana per 5 sec = 1.6 mana per second. 20 spirit = 2.5 mana per second (when not "in combat", in other words having cast a spell in the last 5 seconds.)

Let x be the % of when we are in combat

1.6 = (1-x)2.5

Solve for x. You get 36%. In other words, if you spend more than 36% of the time "in combat", you should wear the mana per second gear. You can also factor in talents that give you 15% mana regen when in combat.

1.6 = (1-x)2.5 + (x)2.5(.15)

NOTE: Factoring in talents like Spirit Tap is more difficult. There is no easy way to tell how often you will proc it, and what percent of the 15 second effect is "in combat" and out of combat. However, lets assume for a moment that spirit tap is always active when you aren't in combat:

1.6 = (1-x)5

Solving for that gives us 68%. The only time you're going to be constantly using spirit tap is when grinding, and then you would hopefully spend more than 68% in combat.

The previous formulas for health/mana regeneration of warlocks didn't seem to fit my experience at all, so I've decided to measure them using a good old stopwatch, wearing different equipment to get varying attributes. The raw data was analyzed by a gnuplot script, resulting in a good fit for the formula Spirit/5 + 15. Can other warlocks verify this? --Archibalgul 12:29, 4 Mar 2006 (EST)


There was some attempts to change the description of Innervate in this article. I know the nature of Innervate is often misunderstood and that people want to correct these misunderstandings but this is not the place to do it. The section where Innervate appears in this article is about the five second rule, not about spell effects in general. As such, only Innervate's +100% modifier to mana regen when the five second rule is in effect is relevant here. The resistance effects of Mage Armor isn't listed either, for instance. --Zootsko 10:02, 5 April 2006 (EDT)

I greatly disagree. Innervate does two effects for mana regeneration- it increases your rate by 100% while casting, and multiples the final rate by 4. Out of combat this means you recover four times faster, but during the 5-second rule you actually can recover faster then when not casting since the 15% and so forth are also multiplied. However, it looks like the developers consider this a bug so it will be changed in the next patch.--Stfrn 08:56, 8 June 2006 (EDT)

What do you disagree with? And where do you see that the 15% and so forth are multiplied by Innervate?--Zootsko 09:25, 8 June 2006 (EDT)

I disagree with your opinion that the details are not needed :) I know that the extra is multiplied by using innervate for months, and from the consensious of the druid forums. Also, the fact that it is currently listed as a bug that will be fixed. But in any case, only half of the mana-regen properties of Innervate are listed currently.--Stfrn 11:12, 8 June 2006 (EDT)

Ah ok. Well, I can't agree with you as to me it would imply that mana regen would be increased by 800% during casting - but I can't stop you from being bold of course. Also, can you please provide a link to something that list it as a bug? I can't seem to find anything. Thanks. --Zootsko 11:37, 8 June 2006 (EDT)

Ah, no you misunderstood- the +100% only comes in when casting, hence the bug. There is no 800%.--Stfrn 23:42, 12 June 2006 (EDT)

Again, I understand how Innervate works (think so anyway). What I mean is that listing other details of Innervate in this table would give the reader the impression that the spell increase mana regen by 800%. --Zootsko 03:24, 13 June 2006 (EDT)

So instead you prefer it lists 200%? I don't see how being inaccurate one way is better then another :P Simplest entery in the table would be "Ignores 5second rule, and recovers four times normal rate" or some such. --Stfrn 23:36, 15 June 2006 (EDT)

I don't know what you mean by 200% but my intention with creating the table was to list the 'Mod Interrupted Mana Regen' value of each effect, as it appears on Thottbot. This is what it currently does for all effects and I don't see why Innervate should be different? --Zootsko 09:00, 16 June 2006 (EDT)

Aha, that is where the confusion is from. Innervate has the 'Mod Interrupted Mana Regen' stated here, that is true, however it also has 'Apply Aura: Mod Power Regen % (Mana)' which my argument has been, is just as important. They do not work exactly the same, however this is the Mana_Regen page, not the % of mana recovered after casting page. And so the fact that innervate that recovers 100% after casting, and multipies the regen rate are just as important to list.--Stfrn 12:28, 16 June 2006 (EDT)

The section where the table appears is only about mana regen after casting. This is why I prefer to leave Innervate at 100%, as I explained in my initial post. Please link to some documentation (such as forum discussion) showing that regen rate after casting is multiplied by Innervate - if it's true, I agree it deserves a note somewhere. If you insist that Innervate's general mana regen increase should be listed, it should at least go into another table in another section. --Zootsko 20:28, 16 June 2006 (EDT)

And as I said, I disagree :) You can insist that innervate only has 100% 'Interrupted mana regeneration ratio increase', but telling only part of relevent infromation is less then useless. As to the forums, you might want to take a look at assuming the forums are working.--Stfrn 21:56, 16 June 2006 (EDT)

It's not useless. No matter how you see it, it's a factor in calculating your interrupted mana regen. And I couldn't care less that you disagree, if you can't explain why - sorry. --Zootsko 04:33, 17 June 2006 (EDT)

I would like to add a few clarifications about Innervate and Evocation that make it necessary to modify how they are displayed in the table. First, both spells only allow 100% of mana regeneration from Spirit while casting to occur; this is highly important to note, as I will show later. Second, Innervate as a spell, increases your total mana regeneration from Spirit by 400% - a final result of 500% mana regeneration. Evocation, as a spell, increases your total mana regeneration from Spirit by 1500% - a final result of 1600% mana regeneration. The difference between the increase in mana regen percentages and the fact that a certain percentage is allowed while casting became considerably important after patch 1.12. Prior to that patch, the percentage of mana regeneration a person could have active while casting was unlimited (i.e. Blue Dragon and Innervate would stack, providing 5 x (100%+100%) = 1000% mana regeneration while casting). Indeed, Innervate would stack with Evocation and Blue Dragon to unecessarily produce 20 x (100%+100%+100%) = 6000% mana regeneration. After patch 1.12, a hard cap of 100% was placed on the amount of mana regeneration from Spirit a person could have active. As such, things like Innervate and Meditation no longer stack together while within the 5 second rule, a total of 500% is only obtainable versus 575% (500%+5x15%). The hard cap of 100% was always in game while a caster was outside the 5 second rule, and maintained that way after the patch. The table on this page is quite misleading without displaying this information.--PlaneShaper 02:07, 30 September 2006 (EDT)

Also, I would like to add that your table titles the last column as an "Interrupted mana regeneration ratio increase" when it is not an increase at all. It is a multiplier. The column title should read "Interrupted mana regeneration percent allowed." Again, clarification should be used to show that the total of the percent allowed is hard capped at 100% after 1.12.

Comparing spirit based and direct regeneration

In theory, direct regeneration is preferable over spirit-based regen, as it is not subject to the 5 second rule. If for example a priest could give away 8 spirit, and gain 5 mana/5, the latter would give the same regen as the spirit at all times, and thus be better. He would forfeit HP regeneration, but mana regen in combat is much more valuable. The factor where Spi and mana/5 give the same base amount of regen (outside the 5 second rule) is 1.6 spi for 1 mana/5 for priests and mages, and 2 spi for all others.

In the Item Values article an attempt is made to reverse-engineer the factors Blizzard uses when creating Items. According to it, 1 mana/5 has the same value as 2.4 spirit. The main insight from this is that priest and mage have no easy solution. Comparing items of otherwise similar properties, the only difference being that one has Spi and the other mana/5, mana/5 gives mages and priests (1.6/2.4 =) 66% of the regen which they would get from spirit. All other classes get (2/2.4 =) 83% of the regen from spirit out of mana/5, so for those the latter seems much more viable.

Example: Comparing Tooth of Gnarr and Seal of Rivendare. According to the Item Value analysis, the Int and Sta on the Tooth are worth a little more than the Int on the Seal. Still for the classes except Mage and Priest, the 3 mana/5 is quite close to the spirit regen (7 spi gives them 3.5 mana/5), so choosing the Tooth should be easy for them. Mage and Priest otoh gain 4.4 mana/5 from 7 spi, so they face a tougher decision.

Your assumptions are still wrong. You write that mana/5 gives 66% of the regen which they would get from spirit. This is false, plain and simple. The alleged item values from the article you refer to has nothing to do with mana regenerative effect. You are comparing apples and oranges. --Zootsko 09:57, 4 June 2006 (EDT)

No, you fail to understand what I'm saying. The premise is: "Comparing items of otherwise similar properties" What about that is wrong?? I'm not comparing apples and oranges, but Items of equal Ilvl etc. --Batox

You're right but my point is that this doesn't show from your text. "Items of otherwise similar properties" means nothing to a casual reader. Maybe we could write "items gained with similar effort" or something like that, which captures the relevance of the point your are trying to make? Also, the 5 sec rule isn't completely accounted for. --Zootsko 12:11, 4 June 2006 (EDT)

Lol, ok. Feel free to edit my text any way you like. Your approach (finding the percentages of regeneration which take place inside/outside the 5 sec rule) yields the most accurate result, no question. There are even UI mods which help finding these percentages. But in my experience, a rule of thumb like: "Take the item with the higher item level (and/or the better quality), and when comparing items of similar level and quality, have that factor of 2 (resp. 1.6) in mind" is much easier in application, and should give quite the same results as the more scientific approach. --Batox 03:08, 5 June 2006 (EDT)

Having played every class int he game to at least 25, with the exception of a paladin, i find that while its obvious the amounts of spirit effect a hunter from a warlock from a priest differently, the effects of mana/regen are always the same to a class. because mana regen is always a set #. this makes it much easier to determin on the fly, rather than the spirit, and along with the 5 second rule, is why most casters like mana/5 more. See, a casters regen, end game, is almost pointless to consider out of battle, unless you have the talents for it, as you will be in combat now in every boss fight, almost. and out of combat, you will be drinking the majority of the time. the entire idea of spirit becomes pointless for a casters mana the second they finish a cast, for 5 seconds. and you are rarely going to just sit down in combat to let your mana regen. you drink pots, get an innervate, pop a mana stone, and keep going. THAT is what makes the mana/5 so useful. 1000 spirit isnt as good as 1 regen/5 seconds if you are chain casting. --Haddon 03:15, 17 June 2006 (EDT)

Haddon, please don't EVER post again. First of all, playing classes up to level 25 is useless, I find nuances about my class at level 70. You don't even see half your abilities at level 25. You are totally ignorant about the difference between the five second rule and out of combat which is the cornerstone of mana regeneration. You probably never heard of healing rotations or meditation. Please, no more posts here.

What I don't get out of this text is, are you comparing in combat or out of combat? Cause a fellow guildy just said that during combat, if he's not casting, he does get his "While not casting" bonus. So I did some maths on my levle 61 resto druid. At that time, I got 50 Mana per 5 seconds from gear bonus. And I have 196 spirit, so that 22.95 mana per 5 seconds from that. My spell info says I gain 73 mana per 5 seconds while casting(50+23=73). While not casting I am suposed to get 197 mana per 5 seconds while NOT casting. So I brought my mana down to roughly 2000. Then, after 10 seconds I initiate combat against a low level mob. My mana regen was 73, while I was doing nothing for over 2 minutes. Which makes me comclude that only the mana per 5 second gear bonus and that gained from spirit matter in combat. Am I right or wrong on this? --TrashCan 11:37, 10 August 2007 (UTC)

I Disagree with this.

Supposedly mana regen is no different in or out of combat according to this article. However if this is the case, then why do certain items (like the Green Dragonscale set ) allow for mana regeneration during combat? It makes no sense that items would allow a certain amount of normal rengeration during combat if mana regen in and out of combat was no different. So I'm going to remove that line. It makes no sense. If I'm wrong someone take it up with Blizzard because I highly doubt they would make such a mistake.

You are wrong. The Green Dragon Mail grants mana per 5 seconds. There is no difference between mana regen in and out of combat. The difference is between mana regen while casting and while not casting. See below --Dendrius

Five-second rule

"After a character expends mana in casting a spell, the effective amount of mana gained per tick from spirit-based regeneration becomes a ratio of the normal listed above, for a period of 5 seconds. During this period mana regeneration is said to be interrupted. This is commonly referred to as the five second rule."

...Explanation, please? In plain English?

I'm interested in this aswell  D ♠ T ♣ C ♦ I ♥ WP 19:46, 15 February 2007 (EST)

Here's the simple version: 5 seconds after you successfully complete casting a spell (that costs mana), you will not receive any mana regen from spirit. The "costs mana" part is important because certain talents (like inner focus or clearcasting) allow you to cast spells that cost no mana.

The "successful" part is important because if you get interrupted, it doesn't count.
The "complete" part is important because the timer starts after you're done casting, not when you begin the cast.

Mana Regen in Druid Forms

After a few tests, I have determined that the mana regen formula in druid tree form is the same as that of biped form. Could someone find out whether or not the reduced mana regen formula applies to moonkin form, or only to bear and cat form? The article should be updated to reflect this info.

Clearcasting and five second rule

This article states that spells which cost no mana don't start the five second rule, and the only exception is listed as shaman clearcasting because it only means 60% mana cost reduction. However, mage clearcasting doesn't mean you can get out of the five second rule. If you cast arcane missiles, proc clearcasting with it, and immediately cast AM again after five seconds, you don't get full mana regeneration even though you haven't casted a spell in the last five seconds, and the spell you have casted in the last five second cost 0 mana. You can test this with regen addons like FuRegen.

CC & 5sr

I believe that your 1st AM, the one that proc'd CC, isn't considered done until you finish channeling it. So on the 2nd AM, you would be 'not casting' for mana regen purposes after it's finished (5 second cast).

It's unpredictable when you'd get your OOC mana regen however because mana/health regen is not per second, its per tick (2 seconds) on a timer independent of your spell casts. Bottom line is that it *really* does not affect mages.

Mages should almost always be casting, moving or oom (or dead :) ). It's kinda a non-factor really. If you're concerned about mana conservation, look at the following:

    - Mage armor
    - Switching to spirit wand/weapons during Evocation (pre-2.3)
    - Mana gems
    - Trinkets like Mark of Defiance or Bangle of Endless Blessing (at least pre-2.3)
    - Arcane Meditation (especially post 2.3)

Healers on the other hand, can find times they're not casting (tank mitigating a lot of damage/phase changeovers, etc..). I really notice it, not on my 70 mage, but my 70 priest when I get a clearcasting and follow it with inner focus (mana-free spell). --16:49, 5 October 2007 (UTC)~~

New formulas in BC?

Can anyone confirm or repudiate the following formulas? (These were created by re-engineering the current armory data)

PALADIN: MP5 = round(Spirit x 0.500 + 37.500);

HUNTER: MP5 = round(Spirit x 0.500 + 38.000);

PRIEST: MP5 = round(Spirit x 0.628 + 30.641);

SHAMAN: MP5 = round(Spirit x 0.5152 + 35.667);

MAGE: MP5 = round(Spirit x 0.631 + 29.877);

WARLOCK: MP5 = round(Spirit x 0.500 + 38.000);

DRUID: MP5 = round(Spirit x 0.5635 + 33.950);

Thanks Gee @ Kael'thas 08:04, 23 July 2007 (UTC)

Eviile 22:03, 26 August 2007 (UTC) (Double post, I know. But atleast you can read it)

The above formulas may not be correct (atleast for priests) (using Eviile - UD priest - Shadowsong).

Here is why...

Spirit: 306

MP5 Gear: 73 m/5

Following that formula, my mp5 should be (306 * 0.628) + 30.641 + 67 = 289.809 (290)

My tooltip listed MP5 is 304. Adding 50 SPI gives me an additional 33 MP5, or about .66 mp5/spi (on the nose). Therefore, the SPI coefficient is between 0.664 and 0.655. With that, it leaves the formula (Priests) Somewhere between

MP5 = round(Spirit x 0.664 + 27.816)

MP5 = round(Spirit x 0.655 + 30.57)

Testing for Pallys (Using PixieStix - BE priest - Shadowsong).

MP5 (Gear): 64

Spi: 105

MP5: 92 (no mp5 from gear)

Adding 50 SPI

MP5: 117

MP5/Spi: between .495 and .504

MP5 formula is between

MP5 = round(spirit x .495 + 40.275)

MP5 = round(spirit x .504 + 38.88)

Simpler Formula?

Hey, I noticed you simplified the spirit formula but not the MP5 formula:

To get mana every 5: M(s,p) = ((s/4*(1-p)+(k*(s/4)*(p)))/0.4)

is mathematically equivalent to

To get mana every 5: M(s,p) = s*(5/8)*(1-p + k*p)

What's better: (5/8) is the MP5 factor for the priest, so just replace that factor for the appropriate class and you have a universal formula that's simpler than the original. Enjoy :)

(Zusterke 15:02, 27 November 2007 (UTC))