Monday, February 28, 2011

Virtual Rewards and Mortality

Tobold has a post about the "trap" of virtual rewards which I posted a medium-sized comment on. Of course these rewards could be achievements, gear, titles, gold, high scores or whatever else a game gives you, and the topic is obviously quite relevant - if not specifically directed at - WoW.

I commented because I wasn't quite sure what he was trying to say. I don't put much stock in the idea that virtual rewards are impermanent, because we are impermanent and everything we do is impermanent. If that disqualified things from having value then nothing would have any value.1 I think, though, the phenomenon that he is concerned with is the following. Please bear in mind this that I don't mean to put words into Tobold's mouth, but just that this is what I understand the problem with virtual rewards to be:

When we first start playing a game, we get really excited by the gameplay, the graphics, the world, or whatever it is we like about the game. Most games quickly start giving us some kind of reward for playing. Our character levels up or we get a new weapon or we unlock an achievement, for example. The reason we like the game is because we actually like playing it, but the rewards we get serve as a point to crystalize our enjoyment around. In order to easily attach value to an activity as a whole, we like to be able to give that activity a beginning, a middle and an end and the end is what gets top billing. This, of course, all has something to do with dopamine, apparently, but whatever the chemical explanation, we do seem to have a fascination with conclusions.

So though we liked exploring the game, we affix all of that like to hitting level 2 or getting a new weapon or whatever. Then we affix a bunch more like to hitting level 3, and so on. We create for ourselves the mythology that we are enjoying the game because of the rewards it gives us. At some point we will have ceased enjoying the actual gameplay, but we will have long since convinced ourselves that the rewards are good, and so we will continue to trudge through gameplay in order to get to the next reward.

Humans, being terrible at knowing what makes them happy, will not realize that this is not making them happy, and will continue doing it, expecting it to make them happy despite the fact that reality has stopped supporting that expectation.

The fact that there are undoubtedly tons of people playing WoW who fall into this category, and tons of people playing other games as well doing this same thing, certainly seems like a problem. Perhaps less of a problem than marrying someone you really don't like or spending your entire life addicted to a damaging substance - two other things that can be caused by the same phenomenon - but a problem nonetheless. People are doing something for fun, and not having fun doing it.

Of course I think Tobold is also concerned with the other effect that this has. The longer a game like WoW exists, the more people are playing it who don't actually like playing it. Those people think that they'd be happier if they got more stuff, so they tell the developers they want more stuff. The developers put in more stuff. The game ends up being all about getting this stuff and it loses the reasons why the people liked it in the first place.

But here's where I think things get a little weird. Yes, the developers eventually start taking out the things that made those games popular in the first place. But no one actually likes those things anymore. People may have come to the game because it was hard as hell, because there were insane grinds, because there was massive travel time and forced grouping. Those are the things I was talking about at the beginning - the things they liked that ended up getting lumped in with the virtual rewards. The whole reason, however, that the virtual rewards became the desired thing was because none of that stuff that originally attracted people to the game is enjoyable for them anymore. If they still loved all that stuff then they'd be getting the same satisfaction out of the game they were in the first place. The reason they are on the virtual rewards treadmill is because the game isn't fun for them anymore. So what value are all those things that have been taken away to make the virtual rewards grind faster? Who wants those things anyway? Just because people liked a game for reason X when they first started playing it doesn't mean they will ever want to play any game for reason X again.

Virtual rewards trapping people into games is more of a function of how people are than how games or virtual rewards are. The real problem here is that our love for games, like everything else, dies.

And this is why the reams of blog posts saying that WoW sucks or that it has gone completely wrong or that it is dying are so depressing. It's like looking in on a eulogy - an angry eulogy all about blaming the person who died. Of course this has nothing really to do with the Tobold post I started off talking about, since it isn't a "look at where WoW went wrong" post.

A lot of people are still genuinely enjoying playing WoW. It is quite possible to do so. Lots of people like getting virtual rewards, not because they've fallen into a trap, but because they are fun to get. If you don't like doing these things anymore, then maybe it's time to give up the game, or at least cut back on it. As long as you actually like doing those things, there is no problem.

Game designers can make better decisions that will help games have more longevity and that will make them better games, but they can't do anything about the fact that the individuals who play their game will stop wanting to play one day. They also can't make a new good game by just resurrecting all the things that made an old game good.

1. The fact that I am personally troubled by the idea of affixing value to impermanent things is a major contributor to me needing drugs to be able to function in society. I don't recommend this, and if you can manage it, I'd go with accepting that impermanence is okay instead.

Friday, February 25, 2011

10 / 25 Player Differences

For most heroic bosses, it seems that 10-player is currently harder than 25-player. Interestingly, the two bosses where it is clearly much harder on 25-player both have chain lightning abilities.

Which brings us to the obvious question of what abilities actually make fights harder for one size of raid than the other. Here is a quick compilation of things that the developers should be keeping in mind when balancing the next round of raid encounters for 10 and 25 players.

Spread Out
When there are more people, there is less room for each person to stand. This is pretty obvious, but it doesn't seem to translate into ability balancing. If a chain lightning jumps 8 yards on both 10 and 25 player mode then it is pretty easy to stand in appropriate positions to avoid jumps on 10 and quite difficult on 25. Since chain lightning often does more damage with each jump, 25-player makes it is very harsh mechanic.
Solution: Vary the jump distance for 10 and 25 player. With a shorter jump distance on 25 player or a longer distance on 10 the mechanic won't favour one size or the other.

Melee Hate
A typical 10 player raid might bring two melee characters, while a typical 25 player raid may have five. Many melee hate mechanics much more difficulty with more melee characters. Think of Kel'Thuzad's Frost Blast. In 10-player, there was no need for more than one person to get hit by it at a time. In 25-player, brining more than three melee meant it was always possible.
Solution: Significantly reassess melee hate - this needs to be done anywway.

Limited Access Abilities
This is a wide range of different abilities that are only available to certain classes and specs. I'm not talking about the fact that it's harder to get all the buffs you need when player with 10 people than it is with 25 people. I'm talking about interrupts, damage reduction cooldowns, Leap of Faith, AoE slows and other abilities that make boss fights much easier. In a 10-player raid, you may have no AoE slow, making the fester blood part of Cho'gall extraordinarily difficult. On 25-player you should have four or five interrupters available to stop a mob that needs to be interrupted regulary, in 10 you might have just one or two. The less specs that can use an ability, and the more uesfull the ability, the more likely it is than 25's have the edge.
Solution: Consider which abilities have a strong influence on the ease of the encounter and whether those are widely available or open only to a few specs. Especially do this with Hand of Protection.

In 25 player you have a pretty easy option for a third tank, in 10-player a third tank is not very viable for the majority of fights. Putting in a large number of adds to be tanked dramatically shifts the difficulty from 25-player to 10-player.
Solution: Vary the number of adds, not just the add health, between the difficulties.

There are plenty of other asymmetric abilities between 10 and 25 player. Most 10-player encounters still feel like afterthoughts based almost precisely after the 25 player version but with lower numbers. Blizzard still has a way to go in terms of understanding 10 player raiding and the same mechanics feel so different when there are different numbers of players up against them.

Thursday, February 24, 2011

Strength of Soul Revisited

I posted about Strength of Soul in light of the 4.0.6 changes to Power Word: Shield. Now we have the subsequent hotfixes to increase the cost of Power Word: Shield and to increase the value of Rapture. Is Strength of Soul still a useful talent?

I had concluded that against a single target, Strength of Soul provided a 10% increase in throughput and a 4% increase in efficiency. Assuming you still cast Power Word: Shield whenever weakened soul is done, Strength of Soul now provides the same 10% increase in throughput but actually ends up costing you some mana for that extra throughput, decreasing efficiency by 2%.

Of course Strength of Soul is also what lets us do heavy single target throughput when the fight calls for it. The Shield -> Flash Heal x3 rotation just doesn't work without it.

Strength of Soul is now a talent that lets you burn mana for throughput when you need to. Given the state of Discipline Priest healing, I think this makes it a worthwhile investment. But if you are not in a moment where high throughput is required, Power Word: Shield should be timed to maximize Rapture procs instead of being cast whenever it is available.

Tuesday, February 22, 2011


1 week ago - Power Word: Shield cost in increased by 36%

Now - Rapture is increased from 6% to 7%.

Power Word: Shield cost too little, so they made it more expensive. Apparently then discipline priests couldn't afford it so they gave us more mana when we cast it. At this rate by mid summer Power Word: Shield will cost 50k but Rapture will return 70k when it breaks.

Wednesday, February 16, 2011

Cast More Spells

On Monday I looked at one way to liquidate excess mana: by substituting an intense casting rotation for a more regular one. Another way to liquidate mana is to simple cast more spells. If we simply took our excess mana and converted it into Power Word: Shields then the rankings for secondary stats would be Spirit (41.5%) > Mastery (28%) > Crit (12%).

If we instead used the extra mana to cast half shields and half renews - thinking that we are probably squeezing these extra spells in while on the move - then the values wouldn't be much different with Spirit (41.4%) > Mastery (23.5%) > Crit (14.7%).

But I feel that this is probably less representative of what is really going to happen in a battle, since encounters are designed around the idea that you will cast spells pretty much as often as you can, but in reality we don't have 100% active time on our log reports, so we can definitely squeeze some more casts in there. I was using a figure of 130k excess mana to spend, which would be enough for 28 shields, or about 38 seconds of extra casting over a 5 minute fight. This is surely an unattainable figure. 38 seconds would mean increasing your active time in a fight by 12.6%. Since my active time is usually around 87% this is getting a little too close to 100% to be believable. In addition to this, the reason we use intense healing rotations is because there are periods that warrant intense healing - even if these periods are just the result of a long time with dodges or critical heals. We can't make up for that loss by simply casting more shields during the lighter damage times.

Another option is to take a half-and-half approach, where we spend half of our excess mana on intense single target healing and half on extra shields and renews.

This result looks a lot like the result from Monday but with the numbers even closer together: Mastery (31%) > Spirit (28%) > Crit (23%).

Of course this doesn't mean that it's time to reforge spirit into mastery. In the case where we are simply running out of mana by the end of the fight, more mana would allow us to continue our normal 6.92 mana efficiency rotation. In this case Spirit (40.6%) > Mastery (18.7%) > Crit (16.5%). So it's a real blowout in favour of Spirit. If such a situation where actually winnable, it would mean damage output was low enough that the full rotation wouldn't be necessary, and mana efficiency would go even higher, making the benefit of spirit relative to the other secondary stats even more extreme.

Spirit has a kind of diminishing returns. Once you have enough mana to cast a baseline mana efficient rotation over all available times during an encounter, the value of spirit drops off. Similarly because an important part of what Intellect does is provide you with mana, the value of Mastery and Crit relative to intellect increase when you pass that point.

But of course determining where that drop off point happens is pretty difficult. Fights are not simple enough to be easily modeled all the way through. Also, I haven't talked about haste at all. Haste increases throughput at the cost of mana. But if I am trying to compare spirit to mastery through estimations of what extra mana can do for us, then surely I can try to fit haste in there somewhere.

When I have some time in the next week I'm going to build myself a spreadsheet dragon and discipline priest and have them face off in a tug-of-war over tank health. That will provide a better picture of how haste fits into the overall scheme of things.

Monday, February 14, 2011

Mana Liquidity

In a previous post I said that Discipline Priests were in the position of usually having enough mana to do the healing they needed to do over the course of the fight, but not having the throughput to actually apply that mana to remedy bad situations. Pach 4.0.6 changed that. Discipline priests now have the ability to have an all out mana fire sale, shipping our mana off in massive quantities for whatever healing we can get.

Before I get into that, I want to look at the question of how to understand spending more mana to do more healing. Given a fixed fight length you generally find yourself with a certain amount of mana to spend. In the case of a discipline priest and a 5 minute fight, that amount of mana might be in the neighborhood of 450k to 490k (plus about 40k per resto shaman in the raid). In a fight of that length you will probably be called on to do something like 2.5M healing, meaning you'll need to average between 5 and 5.5 healing per mana to get there.

A simple single target rotation of Shield -> Penance -> Renew (when Borrowed Time is up) -> Inner Focus -> Greater Heal gives 6.92 mana effiency, more than enough to do the job. With a rotation like this, you'll spend only 361k mana doing 2.5M healing, leaving you with something like 130k mana to play with. The question is, how quickly can you liquidate your mana into healing when you need to?

There are two ways to spend more mana: cast more often or cast more expensive things when you are casting. If we are sticking with a simple single target healing model then casting more often probably isn't a legitimate option. While no one is casting all the time, we are generally casting at all times when we are able and when healing is needed. Running out of a fire might cost us healing time, and a phase transition where no damage is incoming might also cost us healing time.

So today I am going to talk about replacing our existing casting time with more mana intense casting time. I hope to come back to this subject tomorrow or the next day with comments about casting more during downtime.

Discipline now has the Shield -> Flash Heal x3 rotation available. The ability to cast a Power Word: Shield every 5.3 seconds makes for some pretty extreme throughput. Using just Flash of Heal and Shield in 359s you can achieve a throughput of 19.7k, but at the cost of around 4360 mana per second, 2100 more healing per mana than the baseline rotation. For that extra mana you get 4130 more healing per second. You are buying extra healing at a rate of 1.97 healing per mana.

If you use the slightly less insane rotation of incorporating Penance when it is off cooldown, you can buy 3670 healing per second for around 2.23 healing per mana. If you incorporate Renew you get only 3250 extra healing per second but actually only get 2.19 healing per mana, so this is not recommended.

Given the high cost of these rotations, it stands to ask how long we actually afford them. With the 130k unused mana, Shield/Flash could replace the baseline rotation for up to 61 seconds. That actually sounds like an awful lot. If you are casting 90% of the time then you could afford to be on all out mana destruction mode for almost a quarter of the time that you are casting.

So can this tell us anything about how to value Spirit compared to throughput stats?

Using this model, we can calculate how much more healing we can expect to get out of various stats. Of course there are two ways to approach it. One is to fix the amount of time you have to heal and calculate how much more healing you can squeeze out in that time with more of a stat. Another is to fix the amount you need to heal and see how much mana you have left over.

Using either of these methods generates fairly similar results. With fixed healing time, we get Mastery > Crit > Spirit at 36%, 28%, 20% of an Intellect respectively. With fixed healing to be done we get the same order but closer together in the raw amounts at 33%, 27%, 26% respectively.

This model has a lot of deficiencies, and should not be applied as a stat weighting. My hope is that by doing calculations like this I can get a better sense of how Spirit interacts with the other secondary stats. These results can only be narrowly applied, but could suggest that if you find you have enough mana to get through fights, you should consider moving some of your spirit into mastery to increase your ability to deal with difficult situations. They also make me fairly comfortable that mastery is the better choice than crit for improving my throughput, but that could change as I go through more examples.

Friday, February 11, 2011

Almost 50%. Really!

In July I wrote about why the developer idea that discipline priests would get 50% of their healing from shields was ridiculous.

This week Blizzard multiplied the strength of Power Word: Shield by more than three. Of course this didn't just mean three times as much shielding from shields, it meant casting more shields since Power Word: Shield became such a good spell with the change. It also meant stacking mastery over crit, further increasing the power of the shields. I now have one night of raiding data, and my total shields as a fraction of heals was 49.4%. Very near the 50% mark that the developers thought we might have.

So they were basically right, assuming you are willing to overlook the mistake of making our signature spell about 30% as good as it should be.

Thursday, February 10, 2011

Mana Tide Madness

I wrote before about Mana Tide Totem and how it seems like a large and meaningful anomaly in the current mana regen landscape.

Patch 4.0.6 significantly changed Mana Tide by basing the bonus off the Shaman's spirit rather than the recipient's spirit. I understand this was to prevent Mana Tide from being overpowered when used with temporary spirit buffs like Core of Ripeness. Preventing temporary spirit buffs from generating more than three times as much mana as they were intended to is probably a good thing.

Elevating the value of Spirit to obscene levels for Shamans was not.

If you are a Discipline Priest in 359 gear, gaining 100 spirit would give you about 5k more mana to spend over a 6 minute fight. For a Paladin its slightly less, only around 4.6k mana. Druids get 4.8k. Holy Priests, thanks to Holy Concentration get a whopping 6.5k. What about Shamans?

Ignoring Mana Tide, Shamans get around 4.6k just like Paladins. Mana Tide essentially multiples that extra 100 spirit by five for 16.8 seconds out of every 180 seconds, or 9.33% of the time. That's the equivalent of a 37.3% increase to your spirit. If spirit was increasing your mana by 4.6k without Mana Tide then it will be increasing it by 6.4k with Mana Tide.

No big deal, around the same as the Holy Priest. Except for the fact that the other people in your raid get the same bonus. If you are in a 10-player raid, your 100 spirit generates not only an additional 1.8k mana for you, but also for the two other healers. So the total healer mana from your 100 spirit is 9k. That's if the other two healers are also paladins or shamans. If one is a holy priest and the other is a druid then your 100 spirit generates 11k extra healer mana for your raid.

If you are in a 25-player raid with 2 paladins, 2 holy priests, 1 discipline priest and 1 druid, then your 100 spirit generates 19.1k mana.

So in 10 player raids a shaman might get around 1.7 to 2.4 times the benefit from spirit that another healer would get. In 25 player raiders, the number is more like 3 to 4 times the benefit in terms of total mana restored to the raid. Of course shaman are also widely thought to be one of the weaker healers in terms of throughput. That means that giving other people mana is probably better than giving yourself mana.

If other healers stack spirit as high as they can, reforging other stats to it and matching blue sockets with int/spirit gems, then shaman should be taking this a step further. Gem pure spirit, use spirit enchants in every slot (such as 40 spirit to chest rather than 20 stats), eat spirit food and drink a spirit flask. Tell your other raid members they can let a little bit of their spirit go in favour of throughput because you are the mana guy and they don't have to worry about it.

It may not be possible to directly compare throughput to mana stats, but when a mana stat that is desirable to other healers is four times as good for you as it is for them, there is no possible way that at throughput stat can keep up.

And still there is the problem that there is no analog to Mana Tide anywhere. More resto shamans means more mana for your raid. Less shamans means less. You cannot substitute anything, and you can stack as many as you want.

Wednesday, February 9, 2011

Are Glyphs Better?

Nearly a year ago I wrote about Glyphs and how I thought they were a failed game element because they existed to offer choice but didn't offer any real choice and because they interacted very strangely with the skill that made them. Glyphs have been changed substantially since then, so I thought I'd take a look back and compare the state of glyphs then and now.

One complaint on my part was that inscription skill has nothing to do with glyph power. This has only gotten worse since there are no new glyphs at higher skill level in Cataclysm. Inscription is the profession that lets you make glyphs, but you cap out on your ability to do that at 450 skill. The glyphs still come in a seemingly random order and can still be learned well before or after they become useful.

The introduction of major glyphs, however, solved many of the other problems that I had. We no longer have a slew of useless glyphs sitting around that no one will ever use. There are meaningful choices to make between major glyphs that will be made on personal preference as well as on fight mechanics.

But while the major glyphs work well, the prime glyphs and minor glyphs are in the same sorry state they were. Prime glyphs basically offer no choice at all - they are determined entirely by your spec. Their purpose, apparently, is just that players would find it weird that they didn't have glyphs for their most important abilities. I could grudgingly suggest that if this were true there would be a glyph for Greater Heal. But why do I even want them to make a prime glyph for Greater Heal? It's because if it is better than what I'm using I'll use it, and if it's not then it's no skin of my back. Prime glyphs don't add anything to the game. Major glyphs that affected these abilities that didn't get them a percentage boost would be way more interesting.

There is also still remarkable homogeneity in the choice of minor glyphs, when these should be the place where there is the most flexibility. Some choices are just much more attractive than the others, and this should be fixed.

I don't understand the system of getting to choose three minor glyphs of out of five or six - it doesn't seem like much of a choice at all. I truly don't understand the system of having players choose three prime glyphs out of the three that affect their spec - this is definitely not a choice.

Was I right about Glyph of Feint? I don't know, I could actually imagine a rogue taking that for certain fights. Since you aren't GCD locked you could use that on cooldown to lower AoE damage. 50% less AoE damage 60% of the time is actually a very good glyph.

Tuesday, February 8, 2011

Saving the World

Last night my wife asked me how my raid went and I said we beat Al'Akir. She said, "Yay! You're a hero!"

Later on, my wife asked me what the ultimate goal of WoW is.1 My first response was "There is no ultimate goal." My second try, after further pressing was, "To beat all the bosses on hard mode, but in a few months there will be new bosses to beat."

Then, in no small part because me talking about things helps her go to sleep, I told her something that started with, "I'm not sure whether the titans created or found Azeroth," journeyed through old gods, the corruption of Deathwing, his recent escape and the ensuing cataclysm and ended with, "So we need to go to the throne of the four winds and defeat Al'Akir, the lord of air; break into the bastion of twilight, defeat the guards and kill Cho'gall to cut off the head of the twilight hammer cult; and stop Nefarian in Blackwing Descent." We're out to save the world.

I really try to avoid writing nice things about WoW in my blog, but recently I've been feeling kind of down from the number of posts I've been reading about problems with the game. I write posts about problems with the game, of course, but it's the doomsaying that really bothers me. Whether it's outlandish claims that Cataclysm is the worst expansion, not only for WoW, but for any MMO ever, or it's people's concluding blog posts because they just aren't having fun playing anymore, it makes me feel like the sorts of problems that I constantly see with the game are somehow adding up to a bigger picture of decline.

But I genuinely like choosing which heal to cast and on whom. The game really is more fun because I'm playing it with other people. I even like running heroics on easy mode2 to get my guild experience. I also really want Sthenno to save Azeroth from Deathwing, even if I need my wife to remind me of that.

WoW is great. Now if they could only do something about the numbers.

1. She had absolutely no idea who Al'Akir was.

2. In 359 gear.

Monday, February 7, 2011

Strength of Soul

Strength of Soul is a really rotten talent. It lets you cast a bad spell more often at the cost of casting another bad spell. There was really no reason to take it, and if you did take it, you were best off treating those talent points as a sunk cost and never making use of the bonus it gave you.

But of course I am writing this on the eve of 4.0.6, and given the coming changes, this talent deserves some reevaluation. As of tomorrow, instead of letting you cast a spell you don't want to cast more often by casting another spell you don't want to cast, Strength of Soul is going to let you cast a spell you desperately want to cast more often at the expense of casting a spell you want to cast anyway. That sounds like it has a lot more potential.

So I modified one of my old simulators to see what benefit Strength of Soul would give if you had the following casting priority on a single target:

Power Word: Shield
Renew (if Borrowed Time is up)
Greater Heal

With this priority, Strength of Soul results in a 10% increase in throughput with a 4% increase in efficiency for the two points. Of course this isn't a very realistic situation, so I modified the order to the following:

Power Word: Shield
Prayer of Mending
Renew (if Borrowed Time is up)
Prayer of Healing*
Greater Heal

In order to simulate sometimes using Prayer of Healing I gave it a random cooldown with each cast from 0-30 seconds, which roughly corresponds with how often I cast Prayer of Healing against bosses on my last raid. Since Strength of Soul isn't doing as much with this set up, the two points give a 5.4% throughput and 1.4% efficiency increase. That's still a good amount of work to be getting out of two talent points.

With a top end of 5% more healing per point and a midrange of 2.5% healing per point, I think these talent points are probably better than some of the others I currently have, but I don't have a lot of flexibility to move talent points around. I think my weakest talent in my current build that I can afford to get rid of is Surge of Light.

As a side note, for now I am going to be moving two points of Mental Agility to Soul Warding as a nod to the new strength of Power Word: Shield. Both talents get better with Power Word: Shield getting better, but I think Mental Agility goes from somewhat weak to mediocre while Strength of Soul goes from downright useless to situationally great. After I've actually raided with the changes I'll be able to reevaluate Mental Agility to see how much I'm getting out of those points.

Friday, February 4, 2011

Plans Get More Complex

Matticus recently put up a really simple healing guide for Discipline for path 4.0.6. I was asked once in the past if I could put together such a thing, and I've really struggled with the idea because my concept of how to heal isn't very readily summed up.

I'm not saying that healing advice can't fit into short tidbits and neat packages. I actually think Matticus' advice is a pretty good starting point for people who are wondering how 4.0.6 will affect them, or people who are starting to heal as discipline priests after the patch comes out. For example, his AoE healing advice (5 players) is to cast Prayer of Healing. Good advice!

But as I discussed on Wednesday, you can actually increase your throughput by alternative Power Word: Shield and Prayer of Healing. In fact, on Wednesday I forgot to include that using a Shield between Prayers would allow you to get a tick off of the Glyph of Prayer of Healing, increasing Prayer's effect by 6.5%1 and meaning that this alternating cast heals for almost 13% more than prayer alone for only about 7.5% more mana per healing - again assuming that your Rapture proc time will decrease to once every 14 seconds from once every 15.

There are other crazy interactions, however. If you are in for a long period of AoE healing and you cast Penance on one of your targets then while Penance has clearly worse numbers than Prayer, it stacks Grace on that person, giving them 24% more healing from Prayer for 15 seconds. That's an increase of 4.8% to the effect of the prayer as a whole, or more if you were only intending to heal three or four people. You can also haste the Penance with a Borrowed Time without consuming the Borrowed Time so it still gets used on the subsequent Prayer. By squeezing in Penance and Prayer of Mending in the same Borrowed Time window, you get an additional tick of Prayer of Healing. These shenanigans have the potential to get you around 0.8% less healing for around 1.6% less mana, which is hardly any different unless you have your 4-piece set bonus in which case you are spending around 28% less mana by doing this instead of just alternating Shield and Prayer.

Of course if the damage is being dealt equally to all people and there is no additional damage being thrown around, then you won't want to use tactics like these because there will create differences in the health of the people you are healing which will cost you efficiency later. That is, unless you know damage will be coming in for a long enough time that you will be able to give everyone a turn getting the boost. Or if the tank is if the group you are AoE healing then you can do this on the tank while AoE healing everyone else.

So, my AoE healing strategy for discipline priests would be something like this:

1. This is 6.5% rather than 10% because this glyph does not help the Divine Aegis portion of the Prayer.

Wednesday, February 2, 2011

Power Word: Shield is Awesome, but it May Not Be Overpowered, but it Probably Is

Obviously Power Word: Shield is going to be awesome come patch 4.0.6, if it doesn't get nerfed into the ground before then. I uploaded a fragment of a spreadsheet I have to google docs so you can see some of the numbers yourself. Check out the bottom of the "Spells" sheet here. The values on the sheet are based on my current gear.

For those who don't like looking at spreadsheets, the key facts are that Power Word: Shield has 97.2% of the throughput of Prayer of Healing and 73.3% of it's efficiency. Those are some pretty strong figures. The only single target competitor is Penance, which is 7% more mana efficient than Shield, but Shield has 72% better throughput than Penance. Nothing else in the single target department is ever close to Shield in either category.

The question is, is shield really a single target heal? If there was a healer who had an instant cast heal-to-full spell that didn't cost mana, but couldn't do anything else and that spell had a 15 second cooldown, you probably couldn't afford to use one of your three healer spots in a 10-player raid on them. 15 seconds is really a very long time to wait between being effective. Of course Discipline priests can do things when they are not casting shield, but the question is, is it right to consider a heal that can only affect a single target once every 15 seconds a single target heal, or is it more like an AoE heal?

As an AoE heal Shield is right about where it should be. Worse than Prayer of Healing in raw numbers, but probably better than Prayer of Healing because of targetting restrictions and overheal. Prayer has a tendency to overheal in excess of 20% because it will hit a target who doesn't need the healing. Shield doesn't overheal much if at all when used correctly.

Shield is both better an worse than other AoE heals. It's better because you can stop the damage from happening at all and react quickly and powerfully to individuals who are at low health. It's worse because you have to anticipate damage rather than being allowed to react to it, and because it doesn't heal evenly over multiple targets.

The second part is a minus because AoE heals are much better than single target heals. If a big AoE is coming at five people and I shield two of them before the AoE hits then those two people don't end up down health, or at least not much. If I then choose to use Prayer of Healing, my best spell for healing three people up after an AoE then it is working at 60% effectiveness. My pre-shielding was acutally just pre-overhealing. If I can shield all five people, or shield four and use Penance on the last one, or if the damage is dangerous enough that waiting on Prayer to heal them up after is too slow, Or shield everyone but the shadow priest and the feral druid who will heal themselves eventurally, or shield the one person who isn't in the same group as the others then Shield was great. If I just shield a couple of people at random, Shield may end up being a total waste of mana.

This is also why just casting shield every three or four seconds probably won't contribute enough to the group healing. By keeping a larger number of people at full health through the AoE you are denying your co-healers maximum mana efficiency in order to maximize your own. With smart shielding this shouldn't happen, but there are some damage distributions where it's simple not possible to contribute beneifically with a single-target effect that cannot be repeated within 15 seconds.

Back in Wrath I read all about how Discipline priests did nothing but spam Power Word: Shield. Some players countered that they cast all kinds of heals and that people who just spam one spell aren't playing right. The argument from Blizzard was that while spamming one spell may not have been the best way to play, it was close enough to the best way to play that it was a problem.

This never entirely clicked with me, but I realized at some point it was because I raided with 10 players. With a 1 second GCD on shield spam and only 10 people to target every 15 seconds, I was forced to do other things. Not that I kept a shield on everyone at all times, but even in a fight like Freya + 3 where putting shields on everyone is the norm, I still had to spend some of my time on something else. Now that the majority of raiders are raiding 10 player we may see shield "spam" go down dramatically even if Shield is substantially overpowered.

If Power Word: Shield is an AoE heal, then it probably balanced. If it is a single-target heal then it is grossly overpowered. In reality it is neither of these things, and I think the experiment of making it extremely powerful will be an interesting one. We'll see how it works out over the next month or so.