Continuing my discussion on projecting Spirit and mana growth in Warlords

The last time we checked on Warlords mana and projecting how far it would go was after the general raid trinket buffs and the reveal of the BRF raid loot (including two high-Spirit trinkets).  Those two changes had severely accelerated potential Spirit growth from what I initially projected.  Because of that, I went from saying that Spirit was looking to be fine (at launch) to saying that the growth had to be arrested, and that an easy and effective approach would be to simply put less Spirit on trinkets.

Since all that, there have been two more developments:

• Blizzard, going directly counter to my opinion, further increased the Spirit on BRF trinkets.  Not only through the global 5-ilvl buff, but through a significant Spirit-only buff to the two trinkets I linked above.  The reasons for that have to do with the trinket itemization dilemma I touched on here; I won’t go into them now.
• Completely reversing course, the datamined 6.2 healer trinkets now have low amounts of Spirit again.  379 at ilvl 695 (or equivalent) on all of them, far less than Autoclave and Talisman.  This is now in line with exactly what I recommended.

So of course the question is, after all this, where did we land?

A Somewhat Convenient Truth

By now you’re probably familiar with the bar graphs from the previous versions of this analysis.  If you’re not, the framework I use is to compute how much total mana a healer will have available to spend, in an encounter of fixed length (I use 6 minutes).  This is to give a more practical and in-context comparison than simply looking at Spirit numbers, although in the Warlords system, Spirit on gear is the only part that varies.  Without further ado:

To explain the different columns:

• 695 (expected): This was my projection, before any of the post-launch changes, of where we’d be at ilvl 695.  More on this below.
• 695 (BRF 6.0): Where we were in the last version of this post, after the first round of trinket buffs, when I started to be a lot more concerned about this.
• 706 (BRF 6.1): The most Spirit you could get from BRF, wearing the two highest-Spirit trinkets after all the buffs.
• 731 (projected): where we’d be after another tier if nothing more was changed.  Take note of the number–about 950 (i.e. 950,000 mana in a 6 minute fight).  We’ll put that in context later.
• 695 (HFC): how things look at ilvl 695 wearing one of the new Spirit trinkets and a 0-spirit class trinket.
• 731 (HFC): how the previous setup scales up to the best gear we know of in the next tier.

Notes:

• The final two columns conservatively assume one Spirit trinket.  If you use two Spirit trinkets instead of your special Archimonde class trinket, the final bar would wind up at 818.  Both setups will likely be seen in practice, although hopefully the class trinkets wind up appealing.  Because they’re unique and interesting, and also because it helps keep this problem in check.
• This total ignores set bonuses, and in certain cases like Shaman, a big part of the current problem is the tremendous mana value of their set bonuses.  The currently datamined set bonuses, though incomplete, don’t seem to be focused on mana, which is hopefully intentional (all of them are linked here).  So the difference between BRF and HFC is likely to be larger than it looks.

Putting it all together, the graph should speak for itself.  Spirit growth has not only slowed down, but been given a huge one-time step backwards, so that it has room to comfortably grow again.  This outcome isn’t perfect, but it’s probably is the best that could have been done after BRF.

Historical Context

Let’s look back at the first of these graphs I posted, a few weeks before Warlords launch:

Remember all those different sources of mana we had in Mists?

Note that the WoD 695 projection is the first bar in the graph at the start of this post.  That should help put these together in context.

One important bar to note is the L90 6.0 bar–showing the weird month in between 6.0 and Warlords launch, when people probably a remember a very silly mana game where you simply couldn’t spend it down.  In fact, the main point of this chart, which I posted during that time period, was that it wouldn’t be that way once we got to actual Warlords.

Well, the repeated trinket buffs almost did make a liar of me.  The L90 bar topped out at about 1,050 (1.05 million mana to spend in 6 minutes, adjusted for inflation to be in terms of L100 mana).  And you can actually get to 868 in BRF trinkets (with the potential of going to 950 next tier if nothing was changed).  While not quite at L90 levels, it is uncomfortably close.  Close enough that many people are definitely at the “can’t find a use for more mana” level, which is exactly what we wanted to avoid.  To have the feel of early Warlords, you have to be down closer to 600.

It’s still unideal that you can get to around 818 with (known) BIS Hellfire trinkets.  That is just about what a lot of people are probably playing with right now.  People generally aren’t running around with M WF Autoclave and Talisman, because it’s simply too much.  So “current mana” is probably something less than 868–something around 800 is a lot more realistic (Mythic Autoclave and Chew Toy is pretty close to this).  Despite the big cut to Spirit happening in HFC, natural ilvl growth will get us back to current levels by the end of the next tier.

Conclusion

The devs did take significant action on this issue.  After 30 more ilvls of growth, we’ll at most be back to where we are now.  That’s saying a lot.  And I said “at most”–if the class trinkets are attractive (which is reasonably likely, especially if Spirit has questionable value), then the highest possible “mana score” you can reach in HFC is around 740.  That’s around what you can reach in current BRF normal (ilvl 670) gear.

The biggest takeaway is that you’ll likely never have as much mana as you do now for the rest of this expansion.  Unless there is in fact another tier, in which case this will definitely have to be revisited.  While in some cases you might get close, the presence of a 0-Spirit trinket option is a significant factor that helps avoid even that.

It’s been more of a fight to keep mana in check than anyone wanted.  Unforeseen changes or no, it turned out to be wrong to think this wouldn’t be an issue.  And a good point for future discussion would be how to avoid this sort of instability.  Whereby, even though things were set up fine at the start, minor changes to items completely blew up the mana projections (in brief: Spirit on items should probably not go exponentially with ilvl).  For now, the growth has been reined in substantially, which is what we wanted to see going into a new tier.

Continuing my discussion on projecting Spirit and mana growth in Warlords

The last time we checked on Warlords mana and projecting how far it would go was after the general raid trinket buffs and the reveal of the BRF raid loot (including two high-Spirit trinkets).  Those two changes had severely accelerated potential Spirit growth from what I initially projected.  Because of that, I went from saying that Spirit was looking to be fine (at launch) to saying that the growth had to be arrested, and that an easy and effective approach would be to simply put less Spirit on trinkets.

Since all that, there have been two more developments:

• Blizzard, going directly counter to my opinion, further increased the Spirit on BRF trinkets.  Not only through the global 5-ilvl buff, but through a significant Spirit-only buff to the two trinkets I linked above.  The reasons for that have to do with the trinket itemization dilemma I touched on here; I won’t go into them now.
• Completely reversing course, the datamined 6.2 healer trinkets now have low amounts of Spirit again.  379 at ilvl 695 (or equivalent) on all of them, far less than Autoclave and Talisman.  This is now in line with exactly what I recommended.

So of course the question is, after all this, where did we land?

A Somewhat Convenient Truth

By now you’re probably familiar with the bar graphs from the previous versions of this analysis.  If you’re not, the framework I use is to compute how much total mana a healer will have available to spend, in an encounter of fixed length (I use 6 minutes).  This is to give a more practical and in-context comparison than simply looking at Spirit numbers, although in the Warlords system, Spirit on gear is the only part that varies.  Without further ado:

To explain the different columns:

• 695 (expected): This was my projection, before any of the post-launch changes, of where we’d be at ilvl 695.  More on this below.
• 695 (BRF 6.0): Where we were in the last version of this post, after the first round of trinket buffs, when I started to be a lot more concerned about this.
• 706 (BRF 6.1): The most Spirit you could get from BRF, wearing the two highest-Spirit trinkets after all the buffs.
• 731 (projected): where we’d be after another tier if nothing more was changed.  Take note of the number–about 950 (i.e. 950,000 mana in a 6 minute fight).  We’ll put that in context later.
• 695 (HFC): how things look at ilvl 695 wearing one of the new Spirit trinkets and a 0-spirit class trinket.
• 731 (HFC): how the previous setup scales up to the best gear we know of in the next tier.

Notes:

• The final two columns conservatively assume one Spirit trinket.  If you use two Spirit trinkets instead of your special Archimonde class trinket, the final bar would wind up at 818.  Both setups will likely be seen in practice, although hopefully the class trinkets wind up appealing.  Because they’re unique and interesting, and also because it helps keep this problem in check.
• This total ignores set bonuses, and in certain cases like Shaman, a big part of the current problem is the tremendous mana value of their set bonuses.  The currently datamined set bonuses, though incomplete, don’t seem to be focused on mana, which is hopefully intentional (all of them are linked here).  So the difference between BRF and HFC is likely to be larger than it looks.

Putting it all together, the graph should speak for itself.  Spirit growth has not only slowed down, but been given a huge one-time step backwards, so that it has room to comfortably grow again.  This outcome isn’t perfect, but it’s probably is the best that could have been done after BRF.

Historical Context

Let’s look back at the first of these graphs I posted, a few weeks before Warlords launch:

Remember all those different sources of mana we had in Mists?

Note that the WoD 695 projection is the first bar in the graph at the start of this post.  That should help put these together in context.

One important bar to note is the L90 6.0 bar–showing the weird month in between 6.0 and Warlords launch, when people probably a remember a very silly mana game where you simply couldn’t spend it down.  In fact, the main point of this chart, which I posted during that time period, was that it wouldn’t be that way once we got to actual Warlords.

Well, the repeated trinket buffs almost did make a liar of me.  The L90 bar topped out at about 1,050 (1.05 million mana to spend in 6 minutes, adjusted for inflation to be in terms of L100 mana).  And you can actually get to 868 in BRF trinkets (with the potential of going to 950 next tier if nothing was changed).  While not quite at L90 levels, it is uncomfortably close.  Close enough that many people are definitely at the “can’t find a use for more mana” level, which is exactly what we wanted to avoid.  To have the feel of early Warlords, you have to be down closer to 600.

It’s still unideal that you can get to around 818 with (known) BIS Hellfire trinkets.  That is just about what a lot of people are probably playing with right now.  People generally aren’t running around with M WF Autoclave and Talisman, because it’s simply too much.  So “current mana” is probably something less than 868–something around 800 is a lot more realistic (Mythic Autoclave and Chew Toy is pretty close to this).  Despite the big cut to Spirit happening in HFC, natural ilvl growth will get us back to current levels by the end of the next tier.

Conclusion

The devs did take significant action on this issue.  After 30 more ilvls of growth, we’ll at most be back to where we are now.  That’s saying a lot.  And I said “at most”–if the class trinkets are attractive (which is reasonably likely, especially if Spirit has questionable value), then the highest possible “mana score” you can reach in HFC is around 740.  That’s around what you can reach in current BRF normal (ilvl 670) gear.

The biggest takeaway is that you’ll likely never have as much mana as you do now for the rest of this expansion.  Unless there is in fact another tier, in which case this will definitely have to be revisited.  While in some cases you might get close, the presence of a 0-Spirit trinket option is a significant factor that helps avoid even that.

It’s been more of a fight to keep mana in check than anyone wanted.  Unforeseen changes or no, it turned out to be wrong to think this wouldn’t be an issue.  And a good point for future discussion would be how to avoid this sort of instability.  Whereby, even though things were set up fine at the start, minor changes to items completely blew up the mana projections (in brief: Spirit on items should probably not go exponentially with ilvl).  For now, the growth has been reined in substantially, which is what we wanted to see going into a new tier.

All posts in this series can be found here.

Before the launch of the expansion, I made this post outlining, among other things, how I projected healer mana availability to increase over the course of Warlords.  The conclusion was that, due to the fact that much of our mana comes from constant sources and comparatively little on gear (due to the limited slots on which Spirit can appear), the growth would be slow.  It appeared that there would not be an explosion of mana that eliminated significant mana constraints on healer gameplay until ilvls beyond 750.  Now seemed like a good time to revisit that analysis with any new information we’ve gained since launch.

The previous post’s projection of mana resources in Warlords.  x-axis is ilvl. y-axis is thousands of mana available in a 6-minute encounter.

New Information and Assumptions

To review, the prior post’s analysis was done by examining the amount of Spirit available at each ilvl character with Spirit on two rings, a neck, a cloak, and one trinket, assuming all slots grew according to the standard ilvl budget formula.  The framework used in the post is to look at the total mana a heal has available to spend during an encounter of a certain length (I used 6 minutes) from all sources: starting mana, base regen, Spirit, and so on.  By and large the analysis is still correct.  There are a few things that have either changed or were not taken into account the prior post:

• You have a legendary ring.  This made starting Spirit a little higher than projected, since many ilvl 615 characters had a 680 Spirit ring.  However, it appears that you will keep the 680 ring at least into Foundry, and the next step is 690.  The highest one currently datamined is 710.  Finally, the proc doesn’t give Spirit, but rather Int (a lesson they probably learned from the Mists meta gem).  So in the end, the ring doesn’t significantly affect the analysis.
• You (probably should) have a Spirit enchant.  It gives 500 Spirit, with a 15s second duration and a 40 second ICD.  I’ll use 15/45 uptime below.  In any case, it is constant, so it doesn’t affect growth.
• Most or all healer-intended raid trinkets have Spirit.  Wearing two Spirit trinkets will probably not be unusual; it’s worth considering.
• Everburning Candle, when the dust settled, gives twice as much mana as its tooltip indicates.  This results in it having an equivalent of 211 Spirit.  That is actually less than normal for a Spirit trinket of its ilvl, so we can ignore it (the reason it continues to be so good is the extremely high Int).
• Finally and most importantly, raid trinkets were all buffed to account for the trinket itemization problems that were noticed after launch.  Because some trinkets were overbudget, Blizzard buffed all raid drop trinkets to ensure they were still strong relative to the others.  This results in Highmaul and Foundry trinkets having stats that are higher that would be expected for their ilvl.  One might guess that Blizzard will have to keep this up in future tiers to continue the trinket progression; in effect, they’ve been forced to increase the expected budget on trinkets.  This is the main change we should focus on now.

With a passive ilvl 630 trinket having 159 Spirit, you would expect that a level 695 trinket would have 291 Spirit.  That is in fact what Elementalist’s Shielding Talisman had when the previous post was written.  But now it has 476.  And Autoclave has 565 due to its weaker proc.  So where the last analysis imagined that a healer at ilvl 695 would have 291 from one Spirit trinket, in reality a healer at ilvl 695 might have 1041 from two Spirit trinkets.  This definitely should cause us to redo the projection.

I think, though, that this is an overestimate of the Spirit budget that will be used going forward, for two reasons.  First, a constraint Blizzard had in buffing trinkets is that they could not change proc tooltips in a hotfix, so they mostly resorted to changing passives (at least, I think that’s why they did so).  So when raid trinkets with passive Spirit and a secondary stat proc needed a large buff in that hotfix, all they could do was inflate the Spirit.  I doubt we will see future raid trinkets with a very lopsided portion of their itemization in Spirit like these two have.  Second, they might try to ease total trinket budgeting back down towards the standard track, while still ensuring future ones are upgrades.

I’ll consider a projection based on 1041 Spirit from trinkets at 695 to be the worst case.  It’s a little hard to give an expected case, because the patchwork of trinket buffs and nerfs produced some inconsistencies that prevent there from being a clear formula for trinket itemization like there was in the past.  My best guess is that the two 685 Highmaul trinkets with Spirit, one of which is fully passive, reflect an intended amount.  398 at ilvl 685, which would be 437 Spirit at ilvl 695, or 874 from two trinkets.

Updated Results

Let’s modify our previous accounting of the total mana available in a 6 minute encounter, at ilvl 695, for 1) the addition of 166.7 Spirit from an enchant (folded into “base Spirit” since it’s constant) and 2) 1041 Spirit from two trinkets rather than 291 from one trinket, or a total of 1501 Spirit from gear.

For more context I also included the real situation from ilvl 630 (with no enchant and one trinket, to match the “old” ilvl 695 model), as well as an ilvl 735 projection based on ilvl 695 as it actually is now.  Even the actual ilvl 695 behavior is a significant jump upward from what was expected.  Part of is due to using a second Spirit trinket, a possibility I should have included in the previous Spirit post.  But part of it is actual ilvl 695 trinkets having almost twice as much Spirit as expected.

788,000 mana in 6 minutes is a lot more than we were expecting to have in the first tier.  In the prior post’s model, that did not happen until ilvl 778.  And in the worst-case model I described, where trinkets continue forward with Spirit int he same proportion, at ilvl 735 we would have 889,000 mana (which the old model did not have us reaching until ilvl 815).

x-axis is ilvl. y-axis is thousands of mana available in a 6-minute encounter.

The yellow line is the reference of where we started before raiding (ilvl 630). The blue line is where we would have gone if things continued proportionately from there.  The red line is where we would go if Blackrock Foundry trinkets became the new standard.  Slightly below the red line is where we’ll be if Blizzard carves back amount of Spirit on trinkets.  Note that graph goes to unrealistically high ilvls to see the overall shape of the curve, but even below 750 the difference is pronounced.

An Inconvenient Truth

Is there cause for concern?  I think there might be, but it depends largely on some unknowns.  In the last post, I said that things looked very flat in the expected ilvl range for this expansion.  Starting from around 600,000 mana available in 6 minutes at 630, we only reached 700,000 mana available by 730 (it would have been around 750,000 if I included double Spirit trinkets).  That, to me, was no cause for concern, as I said in that post.  Healers are supposed to feel some increased mana availability as Spirit increases over the expansion, and a modest increase like that felt like something that would be noticeable but not significantly change the need to pay attention to mana.

We now know that we will surpass that amount in the first tier, with existing ilvl 695 trinkets.  In a very conservative estimate, that there’s one more raid tier capping out at 720, we’d have 847,000.  That’s getting close to a 50% increase in how much mana you can spend in an encounter, compared to what we were playing with a few weeks ago.  If there are multiple planned raid tiers, then we risk getting to ilvl 750 where the exponential growth in the graph is starting to take off.

In short, whether there’s a problem depends on how high ilvls get in remaining tiers, and how much Blizzard backs off of Blackrock Foundry trinket itemization patterns.  We don’t know either of those things.  But even if the growth in this first tier is not a disaster, it seems in any case to be more than is ideal.  I would be happy if Blizzard:

• Reduces the Spirit on Blackrock Foundry trinkets before they get released.  Not essential, but helps smooth out future growth as needed for the next step.
• In future tiers, itemizes a smaller portion of trinkets towards Spirit.  In particular, if ilvls go beyond 715 or 720, treat Spirit differently from other stats, and itemize an unusually low portion to it in the case of trinkets.

Things are probably going to be fine, but we shouldn’t take risks with our future.  To ensure that we preserve a healthy Spirit environment, we should take action now.  It will only get harder as Spirit levels continue to rise.

All posts in this series can be found here.  Good background for this post can be found in this entry.

Many of the big questions about the new healing gameplay in Warlords center around mana management.  While that will probably be a complex topic throughout the expansion, one place we can start is by looking at exactly what’s changed regarding Spirit, regen, and the amount of mana you have to spend.  I’ve been discussing recently how we have a temporary period of even more mana abundance than we had in 5.0, but that this not reflect Warlords healing at all.  Here I’ll explore that in more detail as well as other questions about mana in Warlords.

Overview of Changes

To start with the facts–this how the basic parameters are changing:

A few other relevant points:

• Spirit from will now only be available on certain slots (ring/neck/cloak/trinket).  This affects how much you expect to have, which is discussed below.
• Many significant sources of mana went away or will go away: most importantly, the meta gem and Innervate and similar spells.

Finally, spell costs as a % of base mana often went down.  Not in every case, especially for spells that got design changes which increased costs (Chain Heal).  AoE spells also tended to relatively increase.  But many typical spells that were not changed, such as Rejuvenation, Renew, and Regrowth cost around 2/3 of what they did before (as a % of base mana).  I’m not going to explore this point in too much detail, because it would require making some kind of complicated Consumer Price Index for spell costs.  But keep it in mind when comparing 5.0 numbers to 6.0 numbers.

Even if I did some elaborate normalization of spell costs, that wouldn’t mean much.  Damage comes in in different amounts now, and you use your spells in different quantities.  Basically, we can’t compute a precise watershed amount of mana or regen that will make 6.0 healing feel exactly the same as 5.0 healing, since too many other things changed.  We’ll look at the comparison as well as we can though.

We do have to adjust for deflation somehow though, since 1 mana means something very different in each of the 3 time periods in the above chart.  For the most part I’m simply going to use total mana pool size as the reference.  That is, since it’s gone from 300,000 to 160,000, assume that 1 mana now is equivalent to roughly 300/160 = 1.875 pre-squish mana.  The point above about changing spell costs is mostly to say that this is likely an underestimate of the difference.

Total Mana Availability

Let’s revisit the discussion from the first half of this post, where we computed how much total mana there was to spend in a 6-minute encounter.  There we used a character with a somewhat typical 12,000 Spirit in 5.0.  For 6.0 we’ll use a character with Spirit in the standard 4 slots (2 rings, neck, cloak), and 1 passive Spirit trinket (so you could of course have more or less than this based on trinket choice).

Things are a little different at L90 since we’re still in MoP.  My character, immediately after the squish, now has 481 Spirit.  That’s with Spirit in the standard 4 slots, plus a few leftover gems and an Amp trinket.  That’s probably pretty typical, so we’ll use that.  It also turns out it’s about the same as what you’d get from 4 Spirit items + 1 passive Spirit trinket at ilvl 580, so the comparison to L100 is still good.

At L90 post-squish, my mana in a 6-minute encounter, again in thousands of mana, is:

• 37 (starting mana)
• 107 (base regen (1480 MP5) over 6m)
• 71 (6 minutes of Spirit regen, at 481 Spirit)
• 23 (meta gem, evaluated @ 3 Rejuvs (2097 mana) per proc)
• 6 (potion)
• Total: 244

What about at L100–let’s say at ilvl 615, corresponding to Normal dungeons and Proving Grounds.  Looking at the relevant items, I’ll have roughly 54 from each of my 4 non-armor slots, and 138 from one trinket, for a total of 354 from gear.  With 784 base Spirit as a Tauren, that’s 1138 Spirit.  So now we’re looking at:

• 160 (starting mana)
• 230 (base regen (3200 MP5) over 6m)
• 169 (6 minutes of Spirit regen, at 1138 Spirit)
• 34 (potion)
• Total: 593

In order to add these new numbers to the previous chart, we have to scale them to account for changes in mana pool size, as discussed above.  Here I’ve scaled them all to correspond to a mana pool size of 160,000:

Looking at this, we can see a number of things.

First, remember that despite the disclaimers about comparing Mists to Warlords, the comparison of L90 6.0 to WoD is a perfect apples-to-apples comparison.  You’re using the new spell costs in both cases, and scaling to mana pool size exactly accounts for the changes in spell costs as you go to level 100.  So this is a clear way of seeing how L90 6.0 is not like level 100 Warlords healing (in particular, you can clearly see the effect of the temporary doubling of base regen, the red part of the bar).  Your gross mana availability is nearly double what it would be if it accurately reflected the level 100 experience.  This is the most important factor driving the sheer weirdness of healing between now and the launch of Warlords.

If you want to think about how things have changed since 5.0, this is a good start.  Even ignoring changes in spell costs, you have more mana than you did two weeks ago.  The subjective play experience right now is that you have much more mana than you did two weeks ago.  In late Mists you could throw heals rather freely, but now you can throw expensive filler heals completely at will, and your mana doesn’t even sink much.  This supports the notion that, in the aggregate, spell costs have come down.  Your total mana is only slightly higher in real terms, but in practice you’re far more flush with mana.  This strongly suggests that if we did construct a Consumer Price Index of spell costs and take it into account, the two leftmost bars in the above graph would be compressed further down.

Spirit Increases in Warlords

A major concern many healers have is whether increases in stats from gear, which tend to increase by a huge amount over the course of the expansion, will result in an overly large amount of regen a few tiers from now.  The Mists post I’ve been linking concluded that increase in Mists was more due to the meta gem and other factors than it was to straight Spirit increases.  This suggests that, even if no changes were made from Mists, the situation would be largely under control unless some ill-advised item were added in Warlords that gave healers huge amounts of mana.

It turns out that, even beyond that, other changes in Warlords are further damping the impact of Spirit increases.  Spirit from gear is an even smaller part of the overall mana picture than it was in Mists.  The chart above pools all “Spirit” together, but remember that of the 1138 Spirit at ilvl 615, 784 was the passive Spirit from being a level 100 Tauren.  Only 354 of it was from gear and will increase with ilvl.  That 354 Spirit accounts for only 52,500 mana over 6 minutes.  This is less than 10% of the total mana we had available at that ilvl (593,000).  The entire remainder of your mana supply will be constant.

Spirit from Gear (purple) will increase with ilvl. All other sources will remain constant.

However, the fact remains that stats, including Spirit, increase exponentially with ilvl (at the rate of 15% every 15 ilvls, or doubling every 74.4 ilvls).  And with exponential growth in the picture, we definitely cannot be satisfied with the mere fact that regen looks safe right now; we need to project the future increases in more detail.

I started doing that when I made the bar graph above, and included a column for Warlords at ilvl 695 (corresponding to Mythic gear at the end of the first raid tier).  Consulting Wowhead again, at that point we expect to have 115 Spirit on a non-armor item, and 291 from a trinket.  This results in 751 Spirit from gear, just over twice what we’ll have at ilvl 615 (as expected after an increase of 80 ilvls).  Total mana supply will have gone from 593,000 to 655,000, as was shown in the graph.  So far, nothing concerning.

Extrapolating forward, with the constant sources of mana staying constant and the gear term increasing exponentially, we have something that looks like this:

x-axis is ilvl. y-axis is thousands of mana available in a 6-minute period.  Red line is constant term only, blue line is constant + gear.

The exponential growth is pronounced, but the very small starting size of the gear term means that it doesn’t start to get out of hand until around ilvl 800, hopefully higher than we’ll ever see.  Between 600 and around 750, the growth is small next to the constant term.

Without getting too deep into speculating on Warlords, let’s remember that ilvls in Mists increased by 63 from the end of the first tier (509 from H Terrace/Heart, ignoring Elite Protectors) to the end of the third and final tier (572 from H WF Siege), ignoring upgrades.  There should be a reasonable hope of going from 695 in M Furnace to something that’s not too far above 750.  If that hope isn’t borne out and items get stronger than is currently foreseeable, then we will very rapidly have to revisit this discussion.  To match the end of Mists in squish-adjusted terms (the 1.01M mana shown in the bar graph), we’d need to reach ilvl 846, so we’re probably safe in that sense.  However, again, if you were to adjust for reduced spell costs, that number will get smaller.  For example, if you guess that spells only cost 80% of what they used to, so you only need to reach 806,000 mana to have something similar to the Mists feel, that will happen at ilvl 783 (hopefully, still out of reach).

One final interesting aside, which I’m going to leave open for now, is whether the very small impact of the “gear Spirit” term will make Spirit less attractive at low ilvls.  It’s not clear that it will (all other stats follow a similar growth pattern as well), but it’s worth trying to look into further.  A related question would be whether Spirit should have diminishing returns so that it grows linearly rather than exponentially with ilvl (this would be accomplished by having regen be O(log(Spirit))).  That seems like a much more stable growth pattern–in fact, if Spirit worked that way, they would probably not have had to squish it down so hard at the start to keep it under control for the whole expansion.  That might make it difficult to balance against other stats, however.

Conclusion

The most immediate thing you should take away from all this is cementing the notion that the current month of healing is a complete oddity, now with more numbers to help explain it.  Keeping our sights set on level 100 though, this is a fuller explication of what I surmised in the “mana economy” post from Mists, namely, that Spirit growth would not be nearly as pronounced in Warlords.  Absent new items on the order of the Mists meta gem, or unanticipated ilvl increases, we should see only a modest increase in mana supply over the course of the expansion.

All posts in this series can be found here.  Good background for this post can be found in this entry.

Many of the big questions about the new healing gameplay in Warlords center around mana management.  While that will probably be a complex topic throughout the expansion, one place we can start is by looking at exactly what’s changed regarding Spirit, regen, and the amount of mana you have to spend.  I’ve been discussing recently how we have a temporary period of even more mana abundance than we had in 5.0, but that this not reflect Warlords healing at all.  Here I’ll explore that in more detail as well as other questions about mana in Warlords.

Overview of Changes

To start with the facts–this how the basic parameters are changing:

A few other relevant points:

• Spirit from will now only be available on certain slots (ring/neck/cloak/trinket).  This affects how much you expect to have, which is discussed below.
• Many significant sources of mana went away or will go away: most importantly, the meta gem and Innervate and similar spells.

Finally, spell costs as a % of base mana often went down.  Not in every case, especially for spells that got design changes which increased costs (Chain Heal).  AoE spells also tended to relatively increase.  But many typical spells that were not changed, such as Rejuvenation, Renew, and Regrowth cost around 2/3 of what they did before (as a % of base mana).  I’m not going to explore this point in too much detail, because it would require making some kind of complicated Consumer Price Index for spell costs.  But keep it in mind when comparing 5.0 numbers to 6.0 numbers.

Even if I did some elaborate normalization of spell costs, that wouldn’t mean much.  Damage comes in in different amounts now, and you use your spells in different quantities.  Basically, we can’t compute a precise watershed amount of mana or regen that will make 6.0 healing feel exactly the same as 5.0 healing, since too many other things changed.  We’ll look at the comparison as well as we can though.

We do have to adjust for deflation somehow though, since 1 mana means something very different in each of the 3 time periods in the above chart.  For the most part I’m simply going to use total mana pool size as the reference.  That is, since it’s gone from 300,000 to 160,000, assume that 1 mana now is equivalent to roughly 300/160 = 1.875 pre-squish mana.  The point above about changing spell costs is mostly to say that this is likely an underestimate of the difference.

Total Mana Availability

Let’s revisit the discussion from the first half of this post, where we computed how much total mana there was to spend in a 6-minute encounter.  There we used a character with a somewhat typical 12,000 Spirit in 5.0.  For 6.0 we’ll use a character with Spirit in the standard 4 slots (2 rings, neck, cloak), and 1 passive Spirit trinket (so you could of course have more or less than this based on trinket choice).

Things are a little different at L90 since we’re still in MoP.  My character, immediately after the squish, now has 481 Spirit.  That’s with Spirit in the standard 4 slots, plus a few leftover gems and an Amp trinket.  That’s probably pretty typical, so we’ll use that.  It also turns out it’s about the same as what you’d get from 4 Spirit items + 1 passive Spirit trinket at ilvl 580, so the comparison to L100 is still good.

At L90 post-squish, my mana in a 6-minute encounter, again in thousands of mana, is:

• 37 (starting mana)
• 107 (base regen (1480 MP5) over 6m)
• 71 (6 minutes of Spirit regen, at 481 Spirit)
• 23 (meta gem, evaluated @ 3 Rejuvs (2097 mana) per proc)
• 6 (potion)
• Total: 244

What about at L100–let’s say at ilvl 615, corresponding to Normal dungeons and Proving Grounds.  Looking at the relevant items, I’ll have roughly 54 from each of my 4 non-armor slots, and 138 from one trinket, for a total of 354 from gear.  With 784 base Spirit as a Tauren, that’s 1138 Spirit.  So now we’re looking at:

• 160 (starting mana)
• 230 (base regen (3200 MP5) over 6m)
• 169 (6 minutes of Spirit regen, at 1138 Spirit)
• 34 (potion)
• Total: 593

In order to add these new numbers to the previous chart, we have to scale them to account for changes in mana pool size, as discussed above.  Here I’ve scaled them all to correspond to a mana pool size of 160,000:

Looking at this, we can see a number of things.

First, remember that despite the disclaimers about comparing Mists to Warlords, the comparison of L90 6.0 to WoD is a perfect apples-to-apples comparison.  You’re using the new spell costs in both cases, and scaling to mana pool size exactly accounts for the changes in spell costs as you go to level 100.  So this is a clear way of seeing how L90 6.0 is not like level 100 Warlords healing (in particular, you can clearly see the effect of the temporary doubling of base regen, the red part of the bar).  Your gross mana availability is nearly double what it would be if it accurately reflected the level 100 experience.  This is the most important factor driving the sheer weirdness of healing between now and the launch of Warlords.

If you want to think about how things have changed since 5.0, this is a good start.  Even ignoring changes in spell costs, you have more mana than you did two weeks ago.  The subjective play experience right now is that you have much more mana than you did two weeks ago.  In late Mists you could throw heals rather freely, but now you can throw expensive filler heals completely at will, and your mana doesn’t even sink much.  This supports the notion that, in the aggregate, spell costs have come down.  Your total mana is only slightly higher in real terms, but in practice you’re far more flush with mana.  This strongly suggests that if we did construct a Consumer Price Index of spell costs and take it into account, the two leftmost bars in the above graph would be compressed further down.

Spirit Increases in Warlords

A major concern many healers have is whether increases in stats from gear, which tend to increase by a huge amount over the course of the expansion, will result in an overly large amount of regen a few tiers from now.  The Mists post I’ve been linking concluded that increase in Mists was more due to the meta gem and other factors than it was to straight Spirit increases.  This suggests that, even if no changes were made from Mists, the situation would be largely under control unless some ill-advised item were added in Warlords that gave healers huge amounts of mana.

It turns out that, even beyond that, other changes in Warlords are further damping the impact of Spirit increases.  Spirit from gear is an even smaller part of the overall mana picture than it was in Mists.  The chart above pools all “Spirit” together, but remember that of the 1138 Spirit at ilvl 615, 784 was the passive Spirit from being a level 100 Tauren.  Only 354 of it was from gear and will increase with ilvl.  That 354 Spirit accounts for only 52,500 mana over 6 minutes.  This is less than 10% of the total mana we had available at that ilvl (593,000).  The entire remainder of your mana supply will be constant.

Spirit from Gear (purple) will increase with ilvl. All other sources will remain constant.

However, the fact remains that stats, including Spirit, increase exponentially with ilvl (at the rate of 15% every 15 ilvls, or doubling every 74.4 ilvls).  And with exponential growth in the picture, we definitely cannot be satisfied with the mere fact that regen looks safe right now; we need to project the future increases in more detail.

I started doing that when I made the bar graph above, and included a column for Warlords at ilvl 695 (corresponding to Mythic gear at the end of the first raid tier).  Consulting Wowhead again, at that point we expect to have 115 Spirit on a non-armor item, and 291 from a trinket.  This results in 751 Spirit from gear, just over twice what we’ll have at ilvl 615 (as expected after an increase of 80 ilvls).  Total mana supply will have gone from 593,000 to 655,000, as was shown in the graph.  So far, nothing concerning.

Extrapolating forward, with the constant sources of mana staying constant and the gear term increasing exponentially, we have something that looks like this:

x-axis is ilvl. y-axis is thousands of mana available in a 6-minute period.  Red line is constant term only, blue line is constant + gear.

The exponential growth is pronounced, but the very small starting size of the gear term means that it doesn’t start to get out of hand until around ilvl 800, hopefully higher than we’ll ever see.  Between 600 and around 750, the growth is small next to the constant term.

Without getting too deep into speculating on Warlords, let’s remember that ilvls in Mists increased by 63 from the end of the first tier (509 from H Terrace/Heart, ignoring Elite Protectors) to the end of the third and final tier (572 from H WF Siege), ignoring upgrades.  There should be a reasonable hope of going from 695 in M Furnace to something that’s not too far above 750.  If that hope isn’t borne out and items get stronger than is currently foreseeable, then we will very rapidly have to revisit this discussion.  To match the end of Mists in squish-adjusted terms (the 1.01M mana shown in the bar graph), we’d need to reach ilvl 846, so we’re probably safe in that sense.  However, again, if you were to adjust for reduced spell costs, that number will get smaller.  For example, if you guess that spells only cost 80% of what they used to, so you only need to reach 806,000 mana to have something similar to the Mists feel, that will happen at ilvl 783 (hopefully, still out of reach).

One final interesting aside, which I’m going to leave open for now, is whether the very small impact of the “gear Spirit” term will make Spirit less attractive at low ilvls.  It’s not clear that it will (all other stats follow a similar growth pattern as well), but it’s worth trying to look into further.  A related question would be whether Spirit should have diminishing returns so that it grows linearly rather than exponentially with ilvl (this would be accomplished by having regen be O(log(Spirit))).  That seems like a much more stable growth pattern–in fact, if Spirit worked that way, they would probably not have had to squish it down so hard at the start to keep it under control for the whole expansion.  That might make it difficult to balance against other stats, however.

Conclusion

The most immediate thing you should take away from all this is cementing the notion that the current month of healing is a complete oddity, now with more numbers to help explain it.  Keeping our sights set on level 100 though, this is a fuller explication of what I surmised in the “mana economy” post from Mists, namely, that Spirit growth would not be nearly as pronounced in Warlords.  Absent new items on the order of the Mists meta gem, or unanticipated ilvl increases, we should see only a modest increase in mana supply over the course of the expansion.