Presently, there are 11 professions in WoW. Each profession grants unique buffs to the user. I've split the buffs into categories and listed the maximum benefit of each profession:
Gathering
Skinning: Master of Anatomy
Herbalism: Lifeblood
Mining: Toughness
Armor
Tailoring: Embroideries - Darkglow, Lightweave, Swordguard
Leatherworking: Fur Lining - Attack Power, Spell Power, Stamina, or 60 Resistance to any school.
Blacksmithing: Two Free Sockets - Bracer and Glove
Buff
Enchanting: Ring Enchants - 32 Attack Power, 19 Spell Power, 24 Stamina
Alchemy: Mixology - Effects vary. Flasks: 37 Spell Power, 64 AP, 13 MP5, or 220 HP
Expansion
Jewelcrafting: Epic Gems. Too numerous to list.
Inscription: Inscriptions (Hurrr)- Axe, Crag, Pinnacle, Storm
Hilarious
Engineering: A variety of oddities- Primarily Hand-Mounted Pyro Rocket and Hyperspeed Accelerators
Now that we have a compiled list of the effects of each profession, we can talk balance. Engineering is a very incomplete profession, and as such will be left out of basically all discussion. Several changes are coming in 3.1, and I will wait until it's out to judge it.
Flexibility
While there are three roles in the game, they're often split into four for buff purposes: Tank, AP-DPS, SP-DPS, and Healer. Many professions have four choices specifically for that reason. There are two huge balance issues in terms of flexibility. First and foremost, the gathering professions are completely and totally inflexible. They grant a buff. Enjoy it. Mining will never improve your DPS. Skinning will never improve your survivability. End of story.
On the other end of the spectrum, Jewelcrafting and Blacksmithing provide an absurd level of flexibility. Since Blacksmithing allows ANY two additional gems to be placed freely, you can stack one stat to about the level of another profession, or you can spread out the benefits between stats. You can also improve a number of stats that other professions do not, and you can modify your choices very easily. Jewelcrafting doesn't have quite the same range as Blacksmithing, but it still has far more choice and flexibility that most professions don't.
The only other flexibility issue is relatively minor. Tailoring lacks a survivability bonus and its healer effect is weak at best, making it slightly less flexible than "average" choices.
A friend of mine asked me for a more specific explanation of flexibility- For example, when would it become a real issue? A Ret Pally isn't going to swap her AP Inscription for a Healing inscription, she's simply going to use them on different piece of gear. However, there's one huge scenario where flexibility comes into play. Say your class greatly benefits from crit rating, so you picked Skinning and maxed it for some crit. Suddenly, there is a change! Now you need to stack Str, and crit isn't nearly as powerful for you anymore. Maybe Blizz hotfixed a mechanic or changed some balance isse. Suddenly, your choice of profession is weaker. However, if you picked JC or BS, you just switch some gems! HURRRRR.
Alternately, say you're sitting right at the hit cap. Then, you get a new piece of gear, and lose a tiny bit of hit. You can just replace a Str gem in your prismatic glove socket with a Str/Hit gem and you're good to go. No need to worry about socket bonuses or anything, and you can replace varying levels of Str with Hit, or vice versa.
Power
In many situations, bonuses are relatively balanced. It's difficult to quantify the effect of Lifeblood, but for example going from 3 Rare Stamina gems to 3 JC Stamina gems is an increase of 17 Stamina per (41 - 24) or 51 total. In comparison, the Mining bonus is precisely 50 Stamina, and a single-point difference is too small to worry about. However, Mixology grants only 220 additional HP. This is worth less than 20 stamina. While this MIGHT be made up for by the assumed benefit of an alchemist stone and a healing potion, the effect on survivability of an additional 30 stamina is MUCH greater than the benefit from popping a bigger potion, which you can only use once during a fight.
Jewelcrafting, however, is too powerful. It has TWO side effects. Since the JC gems are prismatic, they give you free meta activation for having three in your gear, and they give you free socket bonuses on pieces that you couldn't normally get. For example, consider this trinket: Figurine - Monarch Crab If you wanted to stack stamina, you could use two 24 Sta gems for 48 extra Sta, or a 24 Sta and a 12 Sta 12 Str, for a total of 45 Sta and 12 Str (including the socket bonus). However, with a prismatic JC gem, you get 24 Sta, 41 Sta, and the +9 bonus, totaling 74 Sta. This means the JC gem is equivalent to 26 stamina in that slot. Three gems used in such a manner would grant you 78 Sta, more than 50% higher than the other Sta bonuses. That's not counting the number of gems you DON'T need to place in your gear in order to activate metas, giving you even GREATER freedom to stack an additional stat.
The only possible manner in which JC could decrease in power is that as we move towards epic gems, the gap between a JC gem and a standard-epic gem will be much smaller. Even so, using JC gems to complete socket bonuses with an off-color gem would still likely grant an even bonus with other professions and still give you free metas. The solution here is to make JC gems non-prismatic, which would mean you still have to work to get the meta bonus and you can't use JC gems to stack a single stat significantly higher than other classes.
It's getting very difficult for me to peruse this list from my links in a text editor, so I'm going to post it and mull it over for a day or so, and then post any other significant discrepancies I can find. I mean, everyone kinda knows JC/BS is ridiculous right now, but I don't think anyone has quantified why. It's also fun looking at how bonuses to a specific stat from different professions are almost equal.
Tuesday, March 31, 2009
Thursday, March 26, 2009
Overcomeable Challenge
Everyone wants the exact same thing out of a game. Every single person. Every single game. That one thing is overcomeable challenge. I don't know if "overcomeable" is a word, firefox tells me it isn't, but it doesn't have a suggestion under a different spelling, thus I declare that it is in fact now a word.
The idea is simple. Say you're doing PvE. What you want is to get to a boss and say, "Wow, this guy is a challenge." Then, you want to kill him. That's it. Alternately, if you're a fan of PvP, you want to go into, say, a 3v3 arena, fight for a moment and exclaim, "Wow, these opponents are difficult!" Then, you want to crush their faces into the murky waters of the ruins of Lordaeron.
There are two pieces of this process, and both are mandatory. First, you MUST recognize difficulty. If you walk into a raid, pull a boss, and kill him accidentally on your first attempt by working out what you're supposed to do on the fly, you'll have no sense of accomplishment. Alternately, if two weeks down the road you're making no progress you'll feel like the content is impossible. That's why I like that phrase so much. Overcomeable. Challenge. Both parts are equally required.
But wait! There's a problem with this! If the challenge is overcomeable, you've passed it. As long as we assert that there can only ever be a finite amount of content in a game, then at some point you'll run out of things to do and things to kill. And it gets worse! Challenge means something different to EVERYONE. For example, I don't find Sartharion 10+3 to be a challenge. Thousands of people do. I don't consider Patchwerk, Sapphiron, or Kel'thuzad to be challenging, but most likely tens of thousands of people do.
And that's sort of the issue with gearing up. When you don't have gear yet, a lot of things are tuned to be a solid challenge, but still doable. My first kill of Sapphiron 10-man in a previous guild was incredibly tight on mana. However, with more gear the fight becomes less challenging and the feeling of accomplishment for having succeeded diminishes.
A good parallel is in PvP, where the matchmaking system has to be very, very well done. If I walk into an arena, obliterate my opponents, and laugh my way into 2 more arena points, then no one benefits. I don't feel like I accomplished anything, I most likely didn't learn anything from the fight, and the other players didn't feel like they ever had a chance or could have done anything better. To one side it wasn't overcomeable, to the other it isn't a challenge.
My mind is kinda floating right now and I'm starting to ramble. I have more to say on this topic, and I'll condense it into something a little more... sane... some other day. The big points are:
Everyone wants content that is challenging but doable, everyone has a different definition of challenging, and people in the very high end will always blow through challenging content and sit around bored because they've run out of things to do and everything they've done before is no longer challenging. Also, Blizzard is amazing at creating matchmaking systems which link players of equivalent skill so that rofflestomps don't occur too much, though this is shown off better in W3 and SC than in the arena.
The idea is simple. Say you're doing PvE. What you want is to get to a boss and say, "Wow, this guy is a challenge." Then, you want to kill him. That's it. Alternately, if you're a fan of PvP, you want to go into, say, a 3v3 arena, fight for a moment and exclaim, "Wow, these opponents are difficult!" Then, you want to crush their faces into the murky waters of the ruins of Lordaeron.
There are two pieces of this process, and both are mandatory. First, you MUST recognize difficulty. If you walk into a raid, pull a boss, and kill him accidentally on your first attempt by working out what you're supposed to do on the fly, you'll have no sense of accomplishment. Alternately, if two weeks down the road you're making no progress you'll feel like the content is impossible. That's why I like that phrase so much. Overcomeable. Challenge. Both parts are equally required.
But wait! There's a problem with this! If the challenge is overcomeable, you've passed it. As long as we assert that there can only ever be a finite amount of content in a game, then at some point you'll run out of things to do and things to kill. And it gets worse! Challenge means something different to EVERYONE. For example, I don't find Sartharion 10+3 to be a challenge. Thousands of people do. I don't consider Patchwerk, Sapphiron, or Kel'thuzad to be challenging, but most likely tens of thousands of people do.
And that's sort of the issue with gearing up. When you don't have gear yet, a lot of things are tuned to be a solid challenge, but still doable. My first kill of Sapphiron 10-man in a previous guild was incredibly tight on mana. However, with more gear the fight becomes less challenging and the feeling of accomplishment for having succeeded diminishes.
A good parallel is in PvP, where the matchmaking system has to be very, very well done. If I walk into an arena, obliterate my opponents, and laugh my way into 2 more arena points, then no one benefits. I don't feel like I accomplished anything, I most likely didn't learn anything from the fight, and the other players didn't feel like they ever had a chance or could have done anything better. To one side it wasn't overcomeable, to the other it isn't a challenge.
My mind is kinda floating right now and I'm starting to ramble. I have more to say on this topic, and I'll condense it into something a little more... sane... some other day. The big points are:
Everyone wants content that is challenging but doable, everyone has a different definition of challenging, and people in the very high end will always blow through challenging content and sit around bored because they've run out of things to do and everything they've done before is no longer challenging. Also, Blizzard is amazing at creating matchmaking systems which link players of equivalent skill so that rofflestomps don't occur too much, though this is shown off better in W3 and SC than in the arena.
Wednesday, March 25, 2009
Fire Healers for Being Good (Tanks too!)
DPS is awesome. You have 15 DPSers on Patchwerk. They kill him in 180 seconds. Their DPS doubles after a few months and you kill him in 90 seconds. Hooray, progress! Your raid can move faster, there's less time for a boss to gib someone randomly, dangerous abilities like Polarity Shift happen less total times during a fight, so the randomness and the chance of failure decreases significantly.
But not everyone is so lucky. Healers see it the worst. In the beginning, Patchwerk does, say, 30,000 DPS to your tanks. Each healer can sustain 5,000 HP/S, so for a good buffer you have 7-8 healers. But your tanks' gear increases. Incoming damage drops 20% across the board, and you're taking 24,000 total incoming DPS. At the same time, healers' HP/S goes up 20%, so each one is healing 6,000 HP/S. Now it only takes 5 healers to reach a solid buffer zone. The same thing can happen with tanks; In the beginning, you need multiple offtanks to help with adds or some such, and as time goes on you can drop a tank out. That's a somewhat rarer situation though.
This is an extremely annoying effect of gear increases. I mean, it's possible that you have a few people who start out enjoying healing but later want to switch to DPS, or your healers don't care about stepping on each others toes and like to be able to slack off. Who knows. I consider it a problem, though, that as gear increases, healer numbers can dwindle. The other issue is massive healing variance per fight. For example, patchwerk takes a solid amount of healing. However, many fights can be done, and done more efficiently and faster, with less healers. It's not currently feasible for healers to switch to DPS on the fly since the healer requirement changes too often and costs too much. However, with the advent of dual-specs I'm betting we'll see a lot of guilds pick 2-3 healers as swaps, who'll switch to DPS relatively consistently. Blizzard has declared repeatedly that they won't redesign fights under the assumption that you can just ask half your DPS to dual-spec to tank or anything, but I wonder if we'll see much tighter DPS constraints relevant to healing considering that dual-specs have made it easier to minmax.
I wish I had some sort of solution to this issue, but I really don't. I can't think of a way to make it so that healers and DPS see efficiency increases without eventually decreasing the healing and tanking requirements. That said, I feel healing as a whole needs a revamp. More later.
But not everyone is so lucky. Healers see it the worst. In the beginning, Patchwerk does, say, 30,000 DPS to your tanks. Each healer can sustain 5,000 HP/S, so for a good buffer you have 7-8 healers. But your tanks' gear increases. Incoming damage drops 20% across the board, and you're taking 24,000 total incoming DPS. At the same time, healers' HP/S goes up 20%, so each one is healing 6,000 HP/S. Now it only takes 5 healers to reach a solid buffer zone. The same thing can happen with tanks; In the beginning, you need multiple offtanks to help with adds or some such, and as time goes on you can drop a tank out. That's a somewhat rarer situation though.
This is an extremely annoying effect of gear increases. I mean, it's possible that you have a few people who start out enjoying healing but later want to switch to DPS, or your healers don't care about stepping on each others toes and like to be able to slack off. Who knows. I consider it a problem, though, that as gear increases, healer numbers can dwindle. The other issue is massive healing variance per fight. For example, patchwerk takes a solid amount of healing. However, many fights can be done, and done more efficiently and faster, with less healers. It's not currently feasible for healers to switch to DPS on the fly since the healer requirement changes too often and costs too much. However, with the advent of dual-specs I'm betting we'll see a lot of guilds pick 2-3 healers as swaps, who'll switch to DPS relatively consistently. Blizzard has declared repeatedly that they won't redesign fights under the assumption that you can just ask half your DPS to dual-spec to tank or anything, but I wonder if we'll see much tighter DPS constraints relevant to healing considering that dual-specs have made it easier to minmax.
I wish I had some sort of solution to this issue, but I really don't. I can't think of a way to make it so that healers and DPS see efficiency increases without eventually decreasing the healing and tanking requirements. That said, I feel healing as a whole needs a revamp. More later.
QUICKIE!
I'll post something real later, but I just felt like calling attention to this:
AMS Nerf: Ambiguous Wording (WoW Forums)
(Summary: AMS will absorb a maximum of 50% of your HP in damage, so if you have 60k HP, and you're hit for 100k, AMS will absorb 30k, you'll take 70k, and you'll still die.
Compare with:
100% Reduction: Never a Good Idea
Off to work!
AMS Nerf: Ambiguous Wording (WoW Forums)
(Summary: AMS will absorb a maximum of 50% of your HP in damage, so if you have 60k HP, and you're hit for 100k, AMS will absorb 30k, you'll take 70k, and you'll still die.
Compare with:
100% Reduction: Never a Good Idea
Off to work!
Tuesday, March 24, 2009
Raid Sizes, and an Apology
I definitely stopped writing for a few days, and definitely got yelled at by a lot of people. >_> <_< Less that, more text. HURRRRRR
So, raids. Presently, raids are 10 or 25 people. The "or" is the difficult part there. You see, fight complexity does not change between 10m and 25m versions. I realize that I'm using the same 2-3 fights for every single example here, but I'm going to have to go back to them AGAIN, simply because there's not a lot of content yet. WTB Ulduar. Anyway take... *drumroll please* Sartharion! With 3 drakes up, Sartharion requires three tanks. One tank must be on Sartharion himself, one tank must be on Drakes, and one tank must be on Adds, encompassing both fire elementals and whelplings. The issue is that 10-man Sartharion requires the SAME number of tanks, in the same roles. Yes, I know it's possible to do 10m with 2 tanks and a warlock, or in incredibly overgeared situations, there have been reports of the drake tank picking up adds as well, but in the vast majority of situations, 25+3 takes 3 tanks, and 10+3 takes 3 tanks.
Now, healing requirements fluctuate. I'm not going to lie, I don't know precisely how many healers are working various roles, but in general 25m takes 6-7 healers, and 10m takes 3 healers. Filling in the remainders, we're looking at roughly:
25m:
3 tanks (12%)
7 healers (28%)
15 DPS (60%)
10m:
3 tanks (30%)
3 healers (30%)
4 DPS (40%)
See the problem?
If you wanted to make two 10m groups out of your 25m raid, you're going to need some DPS or maybe ONE healer to swap over to tanking. Healing remains relatively static, though when you reach the point that you only need 5-6 healers (A huge issue, that I'll deal with another time...) the disparity shows up there as well.
This has been a big problem with my current raid group. We don't have a lot of spare players laying around, and while we have players with a lot of offset gear, most people don't want to spend time learning their offspec purely to attempt one of the most difficult fights in the game. Thus, each week, 25 of us can do S25+3, but only 10 of us can do S10+3, because the other 15 people, despite their skill, desperately lacks tanks.
Don't get me wrong, this situation isn't common. However, it's generally uncommon because 25-man fights require one or two tanks and 10-man fights also require one or two tanks, not because the tanking requirement is scaled down for 10m versions. I completely expect to see fights requiring 3 tanks in 25m Ulduar still require 3 tanks in 10m Ulduar, leading to the exact same situation repeating. I do feel that if a fight requires 4 tanks in 25m, it will be scaled back to allow less than 4 tanks in 10m.
The question is how to balance 10m raids to require reduced numbers of tanks without changing the difficulty of the fight. However, I feel that it's doable. For example in Sartharion, if you're attempting +Drakes, you could either remove adds from the fight but increase the HP and damage of drakes to compensate for the lack of adds, or you could have adds gain threat based off the drakes so that the drake tank is holding adds and drakes, but slightly decrease the damage output of adds (so that it's survivable). I'm sure there are other viable solutions as well.
Requiring the same number of tanks in 10m runs as 25m runs is a dangerous precedent.
So, raids. Presently, raids are 10 or 25 people. The "or" is the difficult part there. You see, fight complexity does not change between 10m and 25m versions. I realize that I'm using the same 2-3 fights for every single example here, but I'm going to have to go back to them AGAIN, simply because there's not a lot of content yet. WTB Ulduar. Anyway take... *drumroll please* Sartharion! With 3 drakes up, Sartharion requires three tanks. One tank must be on Sartharion himself, one tank must be on Drakes, and one tank must be on Adds, encompassing both fire elementals and whelplings. The issue is that 10-man Sartharion requires the SAME number of tanks, in the same roles. Yes, I know it's possible to do 10m with 2 tanks and a warlock, or in incredibly overgeared situations, there have been reports of the drake tank picking up adds as well, but in the vast majority of situations, 25+3 takes 3 tanks, and 10+3 takes 3 tanks.
Now, healing requirements fluctuate. I'm not going to lie, I don't know precisely how many healers are working various roles, but in general 25m takes 6-7 healers, and 10m takes 3 healers. Filling in the remainders, we're looking at roughly:
25m:
3 tanks (12%)
7 healers (28%)
15 DPS (60%)
10m:
3 tanks (30%)
3 healers (30%)
4 DPS (40%)
See the problem?
If you wanted to make two 10m groups out of your 25m raid, you're going to need some DPS or maybe ONE healer to swap over to tanking. Healing remains relatively static, though when you reach the point that you only need 5-6 healers (A huge issue, that I'll deal with another time...) the disparity shows up there as well.
This has been a big problem with my current raid group. We don't have a lot of spare players laying around, and while we have players with a lot of offset gear, most people don't want to spend time learning their offspec purely to attempt one of the most difficult fights in the game. Thus, each week, 25 of us can do S25+3, but only 10 of us can do S10+3, because the other 15 people, despite their skill, desperately lacks tanks.
Don't get me wrong, this situation isn't common. However, it's generally uncommon because 25-man fights require one or two tanks and 10-man fights also require one or two tanks, not because the tanking requirement is scaled down for 10m versions. I completely expect to see fights requiring 3 tanks in 25m Ulduar still require 3 tanks in 10m Ulduar, leading to the exact same situation repeating. I do feel that if a fight requires 4 tanks in 25m, it will be scaled back to allow less than 4 tanks in 10m.
The question is how to balance 10m raids to require reduced numbers of tanks without changing the difficulty of the fight. However, I feel that it's doable. For example in Sartharion, if you're attempting +Drakes, you could either remove adds from the fight but increase the HP and damage of drakes to compensate for the lack of adds, or you could have adds gain threat based off the drakes so that the drake tank is holding adds and drakes, but slightly decrease the damage output of adds (so that it's survivable). I'm sure there are other viable solutions as well.
Requiring the same number of tanks in 10m runs as 25m runs is a dangerous precedent.
Thursday, March 19, 2009
Devs, and DK Tank Threat
Quite a lot of people I know will occasionally (or, in the case of a few, very often) bash posts and ideas presented by the devs or their representatives. Primarily, one dev, who I bet you can guess, that posts very often... This always bothers me. Consider if you posted constantly regarding class mechanics, researching, discussing, changing, and commenting on every spec of every class all day long. You're not always going to get things right, and people will get angrier about the one mistake you make than about the hundred things you get right. If the game were less complex, then maybe it'd be problematic. As things stand, I really doubt that any single person could go on forever without any issues.
Anyway, on to an actual post.
All three specs of Death Knight are intended to do equal damage. We are going to assume that this is the case, and that a deep (meaning 51+ points) spec of any tree with reasonable point allocation is going to do equal damage. While this can't be perfectly dead on, they're probably (hopefully) close. A Blood DK will have Dancing Rune Weapon up for roughly 30 seconds out of every minute. An Unholy DK will have Gargyole up for one minute out of every 3 minutes. Both of these abilities have an uptime of 33%. They don't have to do equal damage, because they can be balanced around "making up" for the rest of the spec. Frost has no similar pet-based ability.
The issue, then is that if all specs are equal, and part of the damage for Unholy and Blood comes from pets, then threat for a tanking DK will be higher as Frost than as either Blood or UH. The only possible counter to this is that DRW and Gargoyle are both abilities based off of Runic Power, and RP can be funneled directly into Rune Strike instead of other RP-based abilities, hopefully for more threat. Tomorrow, I'm going to delve into the mathematics and work out whether there's an actual imbalance or not.
Yes, this is because it's late.
Anyway, on to an actual post.
All three specs of Death Knight are intended to do equal damage. We are going to assume that this is the case, and that a deep (meaning 51+ points) spec of any tree with reasonable point allocation is going to do equal damage. While this can't be perfectly dead on, they're probably (hopefully) close. A Blood DK will have Dancing Rune Weapon up for roughly 30 seconds out of every minute. An Unholy DK will have Gargyole up for one minute out of every 3 minutes. Both of these abilities have an uptime of 33%. They don't have to do equal damage, because they can be balanced around "making up" for the rest of the spec. Frost has no similar pet-based ability.
The issue, then is that if all specs are equal, and part of the damage for Unholy and Blood comes from pets, then threat for a tanking DK will be higher as Frost than as either Blood or UH. The only possible counter to this is that DRW and Gargoyle are both abilities based off of Runic Power, and RP can be funneled directly into Rune Strike instead of other RP-based abilities, hopefully for more threat. Tomorrow, I'm going to delve into the mathematics and work out whether there's an actual imbalance or not.
Yes, this is because it's late.
Wednesday, March 18, 2009
Tuesday, March 17, 2009
Tier-Based Changes
WoW, like many MMOs, has three primary roles. Tanks, Healers, and DPS. Their roles are something I'd like to look into deeply later on since it's not as easy to sum up as you would expect, but for now, we'll just go with this:
Tanks decrease total incoming raid damage.
Healers prevent death.
DPS kills targets.
I played a Warlock from early Vanilla (Magmadar attempts) until late BC (Brutallus, HOOOO). I don't recall what my damage output was like against Ragnaros, Nefarian, or C'Thun, but I definitely remember dealing ~1k+ DPS in Kara and Gruul, pushing up to 2-3k DPS in Sunwell. I watched Recount and WWS, modified gear, improved, regemmed, respecced, and tried to do slightly more DPS, both overall and on each individual boss, every single week. And I did! My DPS went up. And up, and up. Recount proved it. WWS proved it. Say I did 2400 DPS one week. I get a new piece of gear, next week I'm doing 2450 DPS. Improvement! Right?
I played a tank as well, since my guild only raided 3 days a week during BC. So I ran pugs, as a prot warrior. I didn't really have a stat to track. All I knew was that if I lived, I was doing well, and if I died, either I messed up, or healers messed up, or someone messed up. I got better at picking out when healers were terrible, or when DPS was slow, or when I'd tried to strafe move and accidentally turned my back. I started tracking avoidance, and damage mitigated. If I used to avoid 70% of attacks, and now I avoid 72%, I'm living longer! If I used to block 2,500 damage, and now I'm blocking 2,600 damage, I'm living longer! Right?
My roommate plays a druid. He's the fastest druid in existence, whipping out the Tree Dash to catch up with me as I charge headlong into pack upon pack of add. We talk mechanics all the time, often about Lifebloom, efficiency, regeneration, and throughput. Tiers go up. He has more mana, which means more heals. More spirit, which means more mana per time. More Spellpower, which means each heal does more. Thus his ability to heal constantly increases. Right?
But each of these points is wrong. That's the tricky thing. You see, you can't measure someone's increases in a vacuum. There's no denying that a Sunwell geared tank will survive better against a Karazhan boss, or that a DPSer in Sunwell gear will kill Gruul faster. The question is whether a pre-Kara tank can survive in Kara longer or shorter than a pre-Sunwell tank in Sunwell.
A DPSer will always see their damage increase. But while they'll feel warm and fuzzy, thinking "Yes! I'm doing 10% more damage!" They won't realize that bosses have 20% more HP, thus boss fights are taking longer and longer. Tanks and Healers will feel the pinch more though. If your tanks HP has gone up 50%, but your healing per cast has only gone up 30%, you'll feel like your healing power has actually decreased. If a tanks HP has gone up 50%, but a boss' damage has gone up 60%, his survivability will feel lower, even though he's got significantly better gear.
Additionally, tank survivability has to be based off both incoming damage and incoming healing. This is problematic, because avoidance does not decrease tier to tier. In other words, if you used to take 50% of attacks in Tier 1, and you take 40% of attacks in Tier 2, the attacks have to hit significantly harder, and then have to be inflated by the increase in HP/S. So while before, you were hit half the time for 40% of your HP, now you're being hit a little less than half the time for 70% of your HP, increasing the strain on healers, the danger of tank spike death, etc.
Anyway, my point is that overall, DPSers will see their damage increase from tier to tier while tanks and healers won't see their survivability or actual healing potential increase from tier to tier. Even if the DPSers understanding of the change is somewhat flawed, they at least have a number to look at, saying "it got bigger!" while tanks just say "I'm dying just like I did ever so long ago" and healers say "I can't keep you up through this" followed by "Oh wait, I can."
Tanks decrease total incoming raid damage.
Healers prevent death.
DPS kills targets.
I played a Warlock from early Vanilla (Magmadar attempts) until late BC (Brutallus, HOOOO). I don't recall what my damage output was like against Ragnaros, Nefarian, or C'Thun, but I definitely remember dealing ~1k+ DPS in Kara and Gruul, pushing up to 2-3k DPS in Sunwell. I watched Recount and WWS, modified gear, improved, regemmed, respecced, and tried to do slightly more DPS, both overall and on each individual boss, every single week. And I did! My DPS went up. And up, and up. Recount proved it. WWS proved it. Say I did 2400 DPS one week. I get a new piece of gear, next week I'm doing 2450 DPS. Improvement! Right?
I played a tank as well, since my guild only raided 3 days a week during BC. So I ran pugs, as a prot warrior. I didn't really have a stat to track. All I knew was that if I lived, I was doing well, and if I died, either I messed up, or healers messed up, or someone messed up. I got better at picking out when healers were terrible, or when DPS was slow, or when I'd tried to strafe move and accidentally turned my back. I started tracking avoidance, and damage mitigated. If I used to avoid 70% of attacks, and now I avoid 72%, I'm living longer! If I used to block 2,500 damage, and now I'm blocking 2,600 damage, I'm living longer! Right?
My roommate plays a druid. He's the fastest druid in existence, whipping out the Tree Dash to catch up with me as I charge headlong into pack upon pack of add. We talk mechanics all the time, often about Lifebloom, efficiency, regeneration, and throughput. Tiers go up. He has more mana, which means more heals. More spirit, which means more mana per time. More Spellpower, which means each heal does more. Thus his ability to heal constantly increases. Right?
But each of these points is wrong. That's the tricky thing. You see, you can't measure someone's increases in a vacuum. There's no denying that a Sunwell geared tank will survive better against a Karazhan boss, or that a DPSer in Sunwell gear will kill Gruul faster. The question is whether a pre-Kara tank can survive in Kara longer or shorter than a pre-Sunwell tank in Sunwell.
A DPSer will always see their damage increase. But while they'll feel warm and fuzzy, thinking "Yes! I'm doing 10% more damage!" They won't realize that bosses have 20% more HP, thus boss fights are taking longer and longer. Tanks and Healers will feel the pinch more though. If your tanks HP has gone up 50%, but your healing per cast has only gone up 30%, you'll feel like your healing power has actually decreased. If a tanks HP has gone up 50%, but a boss' damage has gone up 60%, his survivability will feel lower, even though he's got significantly better gear.
Additionally, tank survivability has to be based off both incoming damage and incoming healing. This is problematic, because avoidance does not decrease tier to tier. In other words, if you used to take 50% of attacks in Tier 1, and you take 40% of attacks in Tier 2, the attacks have to hit significantly harder, and then have to be inflated by the increase in HP/S. So while before, you were hit half the time for 40% of your HP, now you're being hit a little less than half the time for 70% of your HP, increasing the strain on healers, the danger of tank spike death, etc.
Anyway, my point is that overall, DPSers will see their damage increase from tier to tier while tanks and healers won't see their survivability or actual healing potential increase from tier to tier. Even if the DPSers understanding of the change is somewhat flawed, they at least have a number to look at, saying "it got bigger!" while tanks just say "I'm dying just like I did ever so long ago" and healers say "I can't keep you up through this" followed by "Oh wait, I can."
Monday, March 16, 2009
Describing Bosses
I've mentioned before that I enjoy vehicle fights. Part of that is probably a lack of new mechanics, something we've suffered from for quite a while. I've been looking back at boss fights recently, and nearly every fight can be mastered by asking these four questions:
1) Is there a place I do NOT want to stand?
Generally, this means void zones, fire walls, poison clouds, falling rocks, whirlwinds, or any other source of damage that can be avoided by moving. If there's a small area where damage might be occurring, you do not want to stand there. This is one of the first things people generally learn about bosses. However, this can also be things like spreading out; The place not to stand is "Near anyone else."
2) Is there a place I DO want to stand?
Power Sparks, Thaddius buffs, Anti-Magic Zones. If moving to a specific location will greatly increase your damage or survivability, you want to stand there.
3) Is there a time I do NOT want to DPS the boss?
If a boss has a spell reflection shield, a threat wipe, a damage shield, a bubble, or some other means to negate your damage or hurt you for attacking, you want to either stop DPS or find something else to attack during that time.
4) Is there a time I DO want to DPS the boss?
If a boss has a "weakened" phase, a phase that you want to get through quickly, a damage multiplier, or any other effect which increases your total damage or efficiency for a short period of time, you want to burn CDs and DPS as hard as possible during that time. This can be as simple as burning CDs during Heroism, if a boss has no weakened state.
I am specifically excluding the role of tanks, as their job is often expressed in very different terms, for example "Kite along this path during X ability" or "use Cooldowns to prevent instagib from X ability," which doesn't translate to my proposed system, though is equally simple to describe. Perhaps that'll be a post for another day.
If you invite a new player to a raid, that is not a tank and maybe not a healer, nearly every fight can be summed up with whichever of these four points happen to be applicable. For example, on Anub'rekhan, you do NOT want to stand near Locust Swarm, and you do NOT want to DPS him during Locust Swarm. On Malygos P1, you DO want to stand on Power Spark, and on Malygos P2, you DO want to stand in the Anti-Magic Zone. On Sartharion, you do NOT want to stand in Void Zones or Lava Walls.
A huge number of bosses can be summed up this way. Hopefully though, a few will come to mind that can't be described by this at all. Kalecgos, for example, doesn't have an easy description of where to stand and where not to stand. Positioning and DPS targets doesn't tell you how to kite striders on Vashj, or how to toss cores up to people who can use them. Also, on Sartharion, while a large portion of the fight (Everything in the standard realm) can be described in this manner, knowing when to enter the twilight realm cannot be defined so easily.
A good measure of a fight, in my opinion, is whether it can be described in the given terms. You'll find that a lot of fights in Naxxramas really can be summed up that easily, including Sapphiron and Kel'thuzad. Malygos is similarly easy to describe, though P3 is mildly unique in that just describing where to stand and who to attack doesn't teach you how to play a dragon. You'll also find that a lot of the really fun and really exciting fights that you remember enjoying learning, practicing, and executing all defy such simple description.
It looks like a majority of Ulduar will be pretty simple, sadly. You DO want to stand by the cozy fire, you DO want to stand by the glacier patch, you DO want to DPS Hodir. Hopefully the hard modes will add significant levels of depth, and not just increase the gearchecks.
1) Is there a place I do NOT want to stand?
Generally, this means void zones, fire walls, poison clouds, falling rocks, whirlwinds, or any other source of damage that can be avoided by moving. If there's a small area where damage might be occurring, you do not want to stand there. This is one of the first things people generally learn about bosses. However, this can also be things like spreading out; The place not to stand is "Near anyone else."
2) Is there a place I DO want to stand?
Power Sparks, Thaddius buffs, Anti-Magic Zones. If moving to a specific location will greatly increase your damage or survivability, you want to stand there.
3) Is there a time I do NOT want to DPS the boss?
If a boss has a spell reflection shield, a threat wipe, a damage shield, a bubble, or some other means to negate your damage or hurt you for attacking, you want to either stop DPS or find something else to attack during that time.
4) Is there a time I DO want to DPS the boss?
If a boss has a "weakened" phase, a phase that you want to get through quickly, a damage multiplier, or any other effect which increases your total damage or efficiency for a short period of time, you want to burn CDs and DPS as hard as possible during that time. This can be as simple as burning CDs during Heroism, if a boss has no weakened state.
I am specifically excluding the role of tanks, as their job is often expressed in very different terms, for example "Kite along this path during X ability" or "use Cooldowns to prevent instagib from X ability," which doesn't translate to my proposed system, though is equally simple to describe. Perhaps that'll be a post for another day.
If you invite a new player to a raid, that is not a tank and maybe not a healer, nearly every fight can be summed up with whichever of these four points happen to be applicable. For example, on Anub'rekhan, you do NOT want to stand near Locust Swarm, and you do NOT want to DPS him during Locust Swarm. On Malygos P1, you DO want to stand on Power Spark, and on Malygos P2, you DO want to stand in the Anti-Magic Zone. On Sartharion, you do NOT want to stand in Void Zones or Lava Walls.
A huge number of bosses can be summed up this way. Hopefully though, a few will come to mind that can't be described by this at all. Kalecgos, for example, doesn't have an easy description of where to stand and where not to stand. Positioning and DPS targets doesn't tell you how to kite striders on Vashj, or how to toss cores up to people who can use them. Also, on Sartharion, while a large portion of the fight (Everything in the standard realm) can be described in this manner, knowing when to enter the twilight realm cannot be defined so easily.
A good measure of a fight, in my opinion, is whether it can be described in the given terms. You'll find that a lot of fights in Naxxramas really can be summed up that easily, including Sapphiron and Kel'thuzad. Malygos is similarly easy to describe, though P3 is mildly unique in that just describing where to stand and who to attack doesn't teach you how to play a dragon. You'll also find that a lot of the really fun and really exciting fights that you remember enjoying learning, practicing, and executing all defy such simple description.
It looks like a majority of Ulduar will be pretty simple, sadly. You DO want to stand by the cozy fire, you DO want to stand by the glacier patch, you DO want to DPS Hodir. Hopefully the hard modes will add significant levels of depth, and not just increase the gearchecks.
Sunday, March 15, 2009
100% Reduction, Never a Good Idea
The concept of reducing all damage taken to 0, even if it's only a specific type, for example melee damage, spell damage, or even just shadow damage, is a terrible, scary idea.
The first time we saw this was tanks pushing 100% avoidance in Burning Crusade. A few classes could get gear sets that allowed you to avoid 100% of melee attacks. Threat became problematic, sure, but it was a very solid setup for an offtank, significantly decreasing the strain on healers, both in terms of HP/S and in terms of split focus and attention.
It was also implemented with the Warlock talent Nether Protection, which allowed a warlock to become entirely immune to Fire and Shadow spells for significant periods of time. This gave warlocks a HUGE advantage in many fights in Black Temple, Hyjal, and Sunwell, simply because raid damage tended to be either Fire or Shadow based.
Currently, 100% avoidance is doable with cooldowns and trinkets, however most of the high-avoidance cooldowns are being nerfed or removed and some of the talents that allow us to reach higher levels of avoidance are being changed as well. One thing sticks out to me, though. The talent Magic Suppression allows Anti-Magic Shell to absorb 100% of all spell damage, in addition to making you immune to negative magical effects.
Many abilities that are intended to function as "Kills you instantly" actually function as "Deals an insane amount of damage." For example, void zones on Kel'thuzad merely do insane amount of shadow damage. Some back-of-napkin calculations place the void zone as dealing 70,000 damage, a number that's definitely not survivable... Unless you have Magic Suppression and pop AMS.
Now, presently, 100% spell absorption is powerful, but it's definitely not game breaking. There's no situation where absorbing 99% of spell damage still gets you killed, but absorbing 100% won't. However, it's entirely possible that some ability will rely on killing you instantly if you make a mistake, but base itself off high damage bursts instead of a death-touch, and when we start seeing things like that, people will be crying about Magic Suppression and how DKs are ez-mode because they can mess up and still survive.
But really, do we need to make the spell a full immunity to all spell damage? It seems like the balance constraints it implies are far more dangerous than the coolness aspect of having the talent. Could just be me, we'll see what happens.
The first time we saw this was tanks pushing 100% avoidance in Burning Crusade. A few classes could get gear sets that allowed you to avoid 100% of melee attacks. Threat became problematic, sure, but it was a very solid setup for an offtank, significantly decreasing the strain on healers, both in terms of HP/S and in terms of split focus and attention.
It was also implemented with the Warlock talent Nether Protection, which allowed a warlock to become entirely immune to Fire and Shadow spells for significant periods of time. This gave warlocks a HUGE advantage in many fights in Black Temple, Hyjal, and Sunwell, simply because raid damage tended to be either Fire or Shadow based.
Currently, 100% avoidance is doable with cooldowns and trinkets, however most of the high-avoidance cooldowns are being nerfed or removed and some of the talents that allow us to reach higher levels of avoidance are being changed as well. One thing sticks out to me, though. The talent Magic Suppression allows Anti-Magic Shell to absorb 100% of all spell damage, in addition to making you immune to negative magical effects.
Many abilities that are intended to function as "Kills you instantly" actually function as "Deals an insane amount of damage." For example, void zones on Kel'thuzad merely do insane amount of shadow damage. Some back-of-napkin calculations place the void zone as dealing 70,000 damage, a number that's definitely not survivable... Unless you have Magic Suppression and pop AMS.
Now, presently, 100% spell absorption is powerful, but it's definitely not game breaking. There's no situation where absorbing 99% of spell damage still gets you killed, but absorbing 100% won't. However, it's entirely possible that some ability will rely on killing you instantly if you make a mistake, but base itself off high damage bursts instead of a death-touch, and when we start seeing things like that, people will be crying about Magic Suppression and how DKs are ez-mode because they can mess up and still survive.
But really, do we need to make the spell a full immunity to all spell damage? It seems like the balance constraints it implies are far more dangerous than the coolness aspect of having the talent. Could just be me, we'll see what happens.
Saturday, March 14, 2009
The Definition and Implementation of Skill
I often hear people say that WoW doesn't require skill, or that WoW isn't a challenging game. Most often, this is said by people that have completed all raid content without attempting a single achievement. Other times it's by high-end raiders who have completed all content months ago.
So what is skill? I would say that skill is the ability to observe a change in the environment and react accordingly. For example, Sarth+3. You see a void zone on the ground. You move. You have observed a change in the environment, and reacted. A fire wall is coming, you find the hole in the wall. On Illidan, he starts casting Shear and you Shield Block. You're dragging a fire elemental in a circle, and an eye beam is coming, so you dodge the eye beam without the elemental gibbing your raid. You can probably list thousands of examples.
Varying levels of skill are determined by the number of possible changes, the difficulty of observing a change, the time you have to react, the complexity of the necessary reaction, and the penalty for failure. There are hundreds of examples of each of these. For example, adding drakes to Sartharion does not change the difficulty of seeing a fire wall or your allowed reaction time, but it does change the penalty of failure by increasing fire damage taken and decreasing HP. It also increases the number of possible changes, in that void zones can spawn. Changing void zones from red to blue was a decrease in the difficulty of observing a change.
Mods generally help to observe changes more easily, thus increasing your allowed reaction time. For example, a mod can tell you when Fire Wall is about to spawn, so that you don't have to wait for the fire wall to exist fully in order to get away. They also help to move all information to one location, so that instead of watching a cast bar to see if a boss is casting an ability while watching a time to see when the next phase is going to happen, you can just watch one location which will say "Cast Incoming" or "Phase changing". An interesting idea for increasing the difficulty of an encounter arbitrarily is to reduce the effectiveness of boss mods. For example, using random timers instead of static timers so you can't prepare for an event, or removing spell reporting from an event so that mods can't easily tell you what's happening, and you just have to use your eyes.
Anyway, there are many tools at the designers' disposal for modifying difficulty based on skill, and with the advent of hard-modes for bosses, hopefully we'll see some clever ways these get changed up.
So what is skill? I would say that skill is the ability to observe a change in the environment and react accordingly. For example, Sarth+3. You see a void zone on the ground. You move. You have observed a change in the environment, and reacted. A fire wall is coming, you find the hole in the wall. On Illidan, he starts casting Shear and you Shield Block. You're dragging a fire elemental in a circle, and an eye beam is coming, so you dodge the eye beam without the elemental gibbing your raid. You can probably list thousands of examples.
Varying levels of skill are determined by the number of possible changes, the difficulty of observing a change, the time you have to react, the complexity of the necessary reaction, and the penalty for failure. There are hundreds of examples of each of these. For example, adding drakes to Sartharion does not change the difficulty of seeing a fire wall or your allowed reaction time, but it does change the penalty of failure by increasing fire damage taken and decreasing HP. It also increases the number of possible changes, in that void zones can spawn. Changing void zones from red to blue was a decrease in the difficulty of observing a change.
Mods generally help to observe changes more easily, thus increasing your allowed reaction time. For example, a mod can tell you when Fire Wall is about to spawn, so that you don't have to wait for the fire wall to exist fully in order to get away. They also help to move all information to one location, so that instead of watching a cast bar to see if a boss is casting an ability while watching a time to see when the next phase is going to happen, you can just watch one location which will say "Cast Incoming" or "Phase changing". An interesting idea for increasing the difficulty of an encounter arbitrarily is to reduce the effectiveness of boss mods. For example, using random timers instead of static timers so you can't prepare for an event, or removing spell reporting from an event so that mods can't easily tell you what's happening, and you just have to use your eyes.
Anyway, there are many tools at the designers' disposal for modifying difficulty based on skill, and with the advent of hard-modes for bosses, hopefully we'll see some clever ways these get changed up.
Friday, March 13, 2009
Voice is a Thin Pipe
I mentioned in Skill vs Coordination that it's much harder for a tank to use external cooldowns than internal cooldowns, primarily because keeping track of what is available and communicating clearly is harder than looking at your own skills and clicking one that's ready. However, communication is a huge aspect of many fights in WoW.
For example, when fighting Malygos, he'll spawn Power Sparks that drift in. The tank needs to pull Malygos away from the spark a tad, while the raid positions between the spark and the dragon. While it's possible for the entire raid to look for the power spark, find it, orient themselves during Vortex, and move in the correct direction, it's much easier to have one player call out where the vortex is and have every other player base their movement off those calls.
Similarly, in Sartharion+3, a Fire Wall is generally a death sentence. While you're in the Twilight realm, you cannot get hit by Fire Wall. However, if you kill the last acolyte in Twilight, everyone in the portal goes back to the normal realm. If a fire wall is coming across as you port out, you're pretty much going to eat the fire wall and die. Thus, it's common to have people outside the portals inform people inside the portal when to kill the last add.
There are many, many situations that are made significantly easier by rapid communication, which generally means voice communication. However, it's my opinion that one goal of UI mods should be to remove the necessity for that level of communication. For example, in Sartharion, mods generally announce when a Fire Wall is moving across the field, but they don't announce when the fire wall has ended. If they did, we could remove some communication and let the mods announce when we're clear to kill an add. Similarly, when fighting Rasuvious, Controllers don't actually need to communicate too much as long as they're watching enemy debuffs. Since Taunt and Bone Shield have the same duration, you can simply see when the other controllers taunt is about to wear off, and use your own taunt and bone shield. Many guilds (including mine) will of course communicate pretty much constantly during Rasuv to be safe, but it's not an actual necessity.
On other bosses, a variety of simple tricks can allow you to reduce the necessary level of communication or attention required from each individual. For example, on Malygos, sparks can only spawn from four corners, and Warlocks can exit Vortex. What this means is that our resident raid leader warlock will look for where the spark is, demonic circle out of vortex, and then place a new demonic circle in the corner the spark is coming from. When everyone lands, the main tank faces the demonic circle and starts backing up, so that he's constantly facing Malygos but pulling Malygos away from the spark. Similarly, the entire raid moves toward the circle while facing Malygos. In this manner, even if we can't call out where people should be going, people still know where to go.
In short, communication suffers from lag, isn't always clear, and can be made unnecessary via UI mods and clever tricks. While it has MANY benefits, I feel as though it's a tiny bit of a crutch, and you should still be able to play perfectly without any sort of voice chat. Pro Tip: If you can only see the fire walls because someone on vent said "Fire Wall, left side" you might be bad at WoW.
For example, when fighting Malygos, he'll spawn Power Sparks that drift in. The tank needs to pull Malygos away from the spark a tad, while the raid positions between the spark and the dragon. While it's possible for the entire raid to look for the power spark, find it, orient themselves during Vortex, and move in the correct direction, it's much easier to have one player call out where the vortex is and have every other player base their movement off those calls.
Similarly, in Sartharion+3, a Fire Wall is generally a death sentence. While you're in the Twilight realm, you cannot get hit by Fire Wall. However, if you kill the last acolyte in Twilight, everyone in the portal goes back to the normal realm. If a fire wall is coming across as you port out, you're pretty much going to eat the fire wall and die. Thus, it's common to have people outside the portals inform people inside the portal when to kill the last add.
There are many, many situations that are made significantly easier by rapid communication, which generally means voice communication. However, it's my opinion that one goal of UI mods should be to remove the necessity for that level of communication. For example, in Sartharion, mods generally announce when a Fire Wall is moving across the field, but they don't announce when the fire wall has ended. If they did, we could remove some communication and let the mods announce when we're clear to kill an add. Similarly, when fighting Rasuvious, Controllers don't actually need to communicate too much as long as they're watching enemy debuffs. Since Taunt and Bone Shield have the same duration, you can simply see when the other controllers taunt is about to wear off, and use your own taunt and bone shield. Many guilds (including mine) will of course communicate pretty much constantly during Rasuv to be safe, but it's not an actual necessity.
On other bosses, a variety of simple tricks can allow you to reduce the necessary level of communication or attention required from each individual. For example, on Malygos, sparks can only spawn from four corners, and Warlocks can exit Vortex. What this means is that our resident raid leader warlock will look for where the spark is, demonic circle out of vortex, and then place a new demonic circle in the corner the spark is coming from. When everyone lands, the main tank faces the demonic circle and starts backing up, so that he's constantly facing Malygos but pulling Malygos away from the spark. Similarly, the entire raid moves toward the circle while facing Malygos. In this manner, even if we can't call out where people should be going, people still know where to go.
In short, communication suffers from lag, isn't always clear, and can be made unnecessary via UI mods and clever tricks. While it has MANY benefits, I feel as though it's a tiny bit of a crutch, and you should still be able to play perfectly without any sort of voice chat. Pro Tip: If you can only see the fire walls because someone on vent said "Fire Wall, left side" you might be bad at WoW.
Thursday, March 12, 2009
Balancing against the upper limit
Imagine a game with three DPS classes (Ignore tanking and healing for now.) Our first DPS class, "Mcweaksauce" can pull 4,000 DPS. Our second "Balanced" can deal 5,000 DPS, and our last, "Imba" can deal 6,000 DPS. It'll take some time to gear up to these levels, so we'll assume players are doing slightly less DPS in the beginning. If we assume a raid has 25 players, of which 3 are tanks and 7 are healers, we have 15 DPS. What should the assumed damage of a raid be?
Most guilds will have a mix of all classes. Assuming a reasonable distribution, we could balance the raid around have 15 Balanced players, for a total of 75,000 DPS. If you're below 75k raid DPS, you're in trouble, either brushing up against enrage timers or getting two waves of adds on fights like Sartharion, or two teleports on Noth (lol?)
Then, we can set up progression. Fights 1, 2, and 3 only take 60K raid DPS. Fights 4-6 take 70k, and fights 7-10 take the full 75k. This means people can gear up on the initial fights, and only later manage the difficult ones. But, we have a problem. While an AVERAGE raid might have to take time progressing through easy content, a raid full of Imbas might be able to clear the whole place while undergeared. And god help any guild with too many Mcweaksauces, they'll never make it past the first sector, even with gear.
This exact problem happened early in Wrath. A few classes, primarily BM Hunters and DW Death Knights, and Sub Rogues were doing so much DPS that stacking them gave a raid a huge advantage. Some of this was due to bugs, some were mere oversights where DPS issues hadn't emerged on the PTR. The result, however, was relatively devastating; Encounters that should have been extremely difficult were instead trivialized. "Gearing up" didn't exist at all. By the time bugs and absurd-DPS issues were resolved, people had enough gear that the real versions still weren't difficult.
I'm going to go out on a limb and say that Blizzard can never, ever balance all classes completely. There will always be some unpredictable side effect of a talent combination or gear setup that causes some class to be stackable for extreme damage. That issue is something that basically has to be taken in stride and reacted to rather than prevented. However, there's another issue that's entirely predictable that has the same symptoms.
Consider a class called Supercooldowns. This bad boy can pump out 5,000 DPS. But there's a catch. Every 30 minutes he can use an ability called Spinach that triples his damage for a full minute. If a fight lasts 5 minutes he will normally do 1.5 million damage. However, every 3rd or 4th attempt, his cooldown will be up. He'll do a blistering 2.1m damage, 40% more than normal. Now assume every class is built this way. Once every 30 minutes, they can do 40% more damage over the course of a fight. Do you balance a fight against their normal damage, and trivialize the encounter when cooldowns are available? Or do you balance a fight against their high-end damage, and make fights impossible when the cooldown is unavailable?
This may seem like a nonissue. Sure, most long CD abilities were struck from the game in 3.1. Furthermore, most long CDs don't equate to triple damage or a 40% increase across a fight. However, even things that increase damage by 5, or 10% across a fight, which aren't available on repeating attempts, can cause problems. The two primary examples that come to mind are Doomguards for Warlocks and Army of the Dead for Death Knights. Doomguards are a 60 minute cooldown that can increase your DPS by around 500 across a fight, a solid 10% increase in most situations (Maybe a tad less.) Army of the Dead can add roughly 100,000 damage on top of your total during a given fight with no opportunity cost, as it can be precast right before the fight starts, thus consuming no runes and no GCDs during combat.
Doomguards are being nerfed, both damage-wise and cooldown-wise, though their damage may still be higher than other pets and their cooldown is still a full 30 minutes. Army of the Dead has no such modification in store. However, I have a suggestion. First off, make it only castable in-combat. This means that you MUST dedicate 8 seconds and 3 runes to summoning your ghouls, time and resources that could have been spent doing something else. Second, retain their threat level so they're still a viable last second save in some situations, but halve their damage. That should mean that the cost of casting them in combat is relatively equivalent to the damage you could have output in the same time, especially when you take the loss of flow into account.
Overall, long CDs are bad. Take them away.
Most guilds will have a mix of all classes. Assuming a reasonable distribution, we could balance the raid around have 15 Balanced players, for a total of 75,000 DPS. If you're below 75k raid DPS, you're in trouble, either brushing up against enrage timers or getting two waves of adds on fights like Sartharion, or two teleports on Noth (lol?)
Then, we can set up progression. Fights 1, 2, and 3 only take 60K raid DPS. Fights 4-6 take 70k, and fights 7-10 take the full 75k. This means people can gear up on the initial fights, and only later manage the difficult ones. But, we have a problem. While an AVERAGE raid might have to take time progressing through easy content, a raid full of Imbas might be able to clear the whole place while undergeared. And god help any guild with too many Mcweaksauces, they'll never make it past the first sector, even with gear.
This exact problem happened early in Wrath. A few classes, primarily BM Hunters and DW Death Knights, and Sub Rogues were doing so much DPS that stacking them gave a raid a huge advantage. Some of this was due to bugs, some were mere oversights where DPS issues hadn't emerged on the PTR. The result, however, was relatively devastating; Encounters that should have been extremely difficult were instead trivialized. "Gearing up" didn't exist at all. By the time bugs and absurd-DPS issues were resolved, people had enough gear that the real versions still weren't difficult.
I'm going to go out on a limb and say that Blizzard can never, ever balance all classes completely. There will always be some unpredictable side effect of a talent combination or gear setup that causes some class to be stackable for extreme damage. That issue is something that basically has to be taken in stride and reacted to rather than prevented. However, there's another issue that's entirely predictable that has the same symptoms.
Consider a class called Supercooldowns. This bad boy can pump out 5,000 DPS. But there's a catch. Every 30 minutes he can use an ability called Spinach that triples his damage for a full minute. If a fight lasts 5 minutes he will normally do 1.5 million damage. However, every 3rd or 4th attempt, his cooldown will be up. He'll do a blistering 2.1m damage, 40% more than normal. Now assume every class is built this way. Once every 30 minutes, they can do 40% more damage over the course of a fight. Do you balance a fight against their normal damage, and trivialize the encounter when cooldowns are available? Or do you balance a fight against their high-end damage, and make fights impossible when the cooldown is unavailable?
This may seem like a nonissue. Sure, most long CD abilities were struck from the game in 3.1. Furthermore, most long CDs don't equate to triple damage or a 40% increase across a fight. However, even things that increase damage by 5, or 10% across a fight, which aren't available on repeating attempts, can cause problems. The two primary examples that come to mind are Doomguards for Warlocks and Army of the Dead for Death Knights. Doomguards are a 60 minute cooldown that can increase your DPS by around 500 across a fight, a solid 10% increase in most situations (Maybe a tad less.) Army of the Dead can add roughly 100,000 damage on top of your total during a given fight with no opportunity cost, as it can be precast right before the fight starts, thus consuming no runes and no GCDs during combat.
Doomguards are being nerfed, both damage-wise and cooldown-wise, though their damage may still be higher than other pets and their cooldown is still a full 30 minutes. Army of the Dead has no such modification in store. However, I have a suggestion. First off, make it only castable in-combat. This means that you MUST dedicate 8 seconds and 3 runes to summoning your ghouls, time and resources that could have been spent doing something else. Second, retain their threat level so they're still a viable last second save in some situations, but halve their damage. That should mean that the cost of casting them in combat is relatively equivalent to the damage you could have output in the same time, especially when you take the loss of flow into account.
Overall, long CDs are bad. Take them away.
Wednesday, March 11, 2009
Skill vs Coordination
My good friend Zdravi brought up some interesting points about my previous post. While he had a lot to say, one major concern was the difference between skill and group coordination. The DPS method for the bronze drake relies on not just working out the rotation, but on having a buddy nearby who also knows the rotation, who can coordinate trading back and forth with you, and thus even if one player is extremely skilled, they'll still be relatively weak until they have a skilled partner.
It's my opinion that coordination is much more difficult to get than skill. My proof of this is Sarth+3. I'm going to assume you're familiar with the strategies for +3, and you've attempted it, defeated it, watched it, or will at least understand my references. If not, go youtube a video or something.
Present raid strategy involves having some overlap between the acolytes of Drakes 2 and 3. During this period, a breath from Sartharion will deal an absolutely absurd level of damage. For this reason, Bear druids and Death Knights are generally used to tank Sarth. Druid/DK cooldowns allow easy survivability. However, other classes ARE capable of tanking it, they just need external cooldowns. For example, a warrior might have to call out for Pain Suppression and Hand of Sacrifice to survive some breaths. The problem is that it takes time and thought.
As a Death Knight, I can switch on the fly. I can think, "Ok, next breath is Icebound Fortitude" and then have Anti-Magic Shield come up JUST before a breath, and use AMS instead. That doesn't cause problems. Whereas, with a warrior, you'd have to think, "That's the last of my CDs, I'll need a Sacrifice for the next breath," and call out for the CD, and pray that the player assigned to use their CD watches and casts it at JUST the right time to prevent a one-shot breath. It's much easier for an individual to manage a set of their own cooldowns than for 3 players to manage the whole groups set of cooldowns.
As time goes on, individual skill becomes more of a given. Especially in high-end guilds, players are EXPECTED to perform at the limit of their ability. At that point, I feel group coordination becomes the real test. Things like tossing Vashj's crystals across the room, positioning properly in Kael'thas, or the original 4H are all much more entertaining than a Patchwerk-style "We beat the enrage because we aren't terrible."
It's my opinion that coordination is much more difficult to get than skill. My proof of this is Sarth+3. I'm going to assume you're familiar with the strategies for +3, and you've attempted it, defeated it, watched it, or will at least understand my references. If not, go youtube a video or something.
Present raid strategy involves having some overlap between the acolytes of Drakes 2 and 3. During this period, a breath from Sartharion will deal an absolutely absurd level of damage. For this reason, Bear druids and Death Knights are generally used to tank Sarth. Druid/DK cooldowns allow easy survivability. However, other classes ARE capable of tanking it, they just need external cooldowns. For example, a warrior might have to call out for Pain Suppression and Hand of Sacrifice to survive some breaths. The problem is that it takes time and thought.
As a Death Knight, I can switch on the fly. I can think, "Ok, next breath is Icebound Fortitude" and then have Anti-Magic Shield come up JUST before a breath, and use AMS instead. That doesn't cause problems. Whereas, with a warrior, you'd have to think, "That's the last of my CDs, I'll need a Sacrifice for the next breath," and call out for the CD, and pray that the player assigned to use their CD watches and casts it at JUST the right time to prevent a one-shot breath. It's much easier for an individual to manage a set of their own cooldowns than for 3 players to manage the whole groups set of cooldowns.
As time goes on, individual skill becomes more of a given. Especially in high-end guilds, players are EXPECTED to perform at the limit of their ability. At that point, I feel group coordination becomes the real test. Things like tossing Vashj's crystals across the room, positioning properly in Kael'thas, or the original 4H are all much more entertaining than a Patchwerk-style "We beat the enrage because we aren't terrible."
Tuesday, March 10, 2009
Vehicles: Blizzard's Intelligence Test
There is an inherent limit to skill in World of Warcraft. Assuming a static game, at some point theorycraft will show the BEST way to play your character, the BEST way to set up your raid, the BEST way to react. Which is why the game isn't static. The game changes, and the way we play our classes changes. However, Blizzard tends to change small things, and rarely changes the overall direction of a class. As time goes on, we get faster and faster at returning to maximum efficiency after a change. While expansions with many new talents and a few new abilities cause us to shift models around and rerun our simulations, the actual effect on our characters tends to be minimal. As an example, Affliction was a powerful spec in early BC. It was overshadowed by Destruction later on, and with the advent of Wrath, returned to its former glory. But did the playstyle change? Ultimately it was still maintain DoTs, cast Shadow Bolt in downtime, Life Tap to avoid OOMing. The change to Drain Soul meant it became worth casting at the end of a boss fight, but that's really the only change. Even the upcoming patch won't really change the playstyle, it'll just switch up how many DoTs are on your target at once.
As time goes on, and our understanding of classes becomes more and more mature, more players will be at the limit of each class. Not everyone will become godlike masters of WoW, but each individual tends to increase in skill over time, albeit at wildly varying rates. So how do we suppress this constant increase? How do we make content difficult while assuming that players are constantly gaining skill? Well, first Blizzard introduced DKs, which was quite a challenge. DKs have basically been an arms race between Blizzard and The Community, wherein Blizzard wants DKs in a specific place in terms of tanking and survival, whereas the community wants DKs to deal more DPS and survive longer. Since the class is still very new, neither side had accurate models, and odd builds like 32/39 DW skyrocketed in effectiveness while tanking builds became largely overpowered. But that's just one arena.
By far the more interesting challenge is vehicles. Each vehicle a player can use is like playing a very tiny but entirely new class. Learning how to use a vehicle properly is Blizzard's intelligence test for the community. You can see this in the Malygos encounter. Many players cannot figure out how to properly use their dragons without outside assistance. Even with the entire skill set in front of them and time to look it over, many people cannot figure out what to do on their own.
The Oculus is an even better example. I love the instance, because you're confronted with the challenge of mastering a class that you haven't seen much and don't use all the time. You don't get to practice outside on the target dummies or solo to get the feel for it, you're in a group, with a variety of new vehicles, and a few enemies in front of you.
Because of the fact that vehicles are a nonstandard mechanic, Blizzard occasionally tries to be blunt about how to properly use each vehicle. For example, the amber (bronze) drake in Oculus has three skills:
Shock Lance
Stop Time
Temporal Rift
Can you figure out, on the fly, without being told, the optimal way to DPS a boss, assuming you're in a group? This is a relatively easy one, but I'm always surprised at the number of people that can't work it out. The answer is to talk to another player on a bronze drake. First, you use Temporal Rift, while he uses Stop Time, then Shock Lance, and immediately hit Temporal Rift. This will give your ally quite a few shock charges. As soon has he has them up, and your Rift is up, he hits Shock Lance then puts up Temporal Rift. You alternate back and forth like this until the boss is dead. This can be done rapidly, without any voice chat, as you just need to make sure another player's Rift is up when you Lance so that you do extra damage and they get charges, and if you don't have any charges, Rift until you do. Assuming everyone knows the strategy, you can do absurd levels of damage this way, including a 5-bronze kill of Eregos on Heroic.
As I mentioned before, I'm always surprised by how many players simply cannot work that out. Person after person rides a bronze drake and spams Shock Lance, or just keeps Temporal Rift up the whole time without ever attempting to DPS themselves. This isn't just the case with pugs either; I've met many players in top-level guilds on my server that couldn't work out proper strategies for the drakes in Oculus and couldn't figure out how to heal or DPS on Malygos. However, I've definitely found a correlation between skill in WoW and skill on vehicles. Generally, the people who get vehicles right are the same people that don't stand in void zones, don't get hit by lava walls, and don't get shocked to hell on Thaddius without having to be told that THIS fire is just as bad as ALL fire.
Vehicles may be a silly gimmick, but they're a really solid way of separating the different levels of player in your guild and they're a fun challenge and an interesting dynamic for boss fights. Of course, considering the number of people that hate Oculus because it's "so hard," adding more vehicles to future encounters may not be the best business strategy. Ah well.
As time goes on, and our understanding of classes becomes more and more mature, more players will be at the limit of each class. Not everyone will become godlike masters of WoW, but each individual tends to increase in skill over time, albeit at wildly varying rates. So how do we suppress this constant increase? How do we make content difficult while assuming that players are constantly gaining skill? Well, first Blizzard introduced DKs, which was quite a challenge. DKs have basically been an arms race between Blizzard and The Community, wherein Blizzard wants DKs in a specific place in terms of tanking and survival, whereas the community wants DKs to deal more DPS and survive longer. Since the class is still very new, neither side had accurate models, and odd builds like 32/39 DW skyrocketed in effectiveness while tanking builds became largely overpowered. But that's just one arena.
By far the more interesting challenge is vehicles. Each vehicle a player can use is like playing a very tiny but entirely new class. Learning how to use a vehicle properly is Blizzard's intelligence test for the community. You can see this in the Malygos encounter. Many players cannot figure out how to properly use their dragons without outside assistance. Even with the entire skill set in front of them and time to look it over, many people cannot figure out what to do on their own.
The Oculus is an even better example. I love the instance, because you're confronted with the challenge of mastering a class that you haven't seen much and don't use all the time. You don't get to practice outside on the target dummies or solo to get the feel for it, you're in a group, with a variety of new vehicles, and a few enemies in front of you.
Because of the fact that vehicles are a nonstandard mechanic, Blizzard occasionally tries to be blunt about how to properly use each vehicle. For example, the amber (bronze) drake in Oculus has three skills:
Shock Lance
Stop Time
Temporal Rift
Can you figure out, on the fly, without being told, the optimal way to DPS a boss, assuming you're in a group? This is a relatively easy one, but I'm always surprised at the number of people that can't work it out. The answer is to talk to another player on a bronze drake. First, you use Temporal Rift, while he uses Stop Time, then Shock Lance, and immediately hit Temporal Rift. This will give your ally quite a few shock charges. As soon has he has them up, and your Rift is up, he hits Shock Lance then puts up Temporal Rift. You alternate back and forth like this until the boss is dead. This can be done rapidly, without any voice chat, as you just need to make sure another player's Rift is up when you Lance so that you do extra damage and they get charges, and if you don't have any charges, Rift until you do. Assuming everyone knows the strategy, you can do absurd levels of damage this way, including a 5-bronze kill of Eregos on Heroic.
As I mentioned before, I'm always surprised by how many players simply cannot work that out. Person after person rides a bronze drake and spams Shock Lance, or just keeps Temporal Rift up the whole time without ever attempting to DPS themselves. This isn't just the case with pugs either; I've met many players in top-level guilds on my server that couldn't work out proper strategies for the drakes in Oculus and couldn't figure out how to heal or DPS on Malygos. However, I've definitely found a correlation between skill in WoW and skill on vehicles. Generally, the people who get vehicles right are the same people that don't stand in void zones, don't get hit by lava walls, and don't get shocked to hell on Thaddius without having to be told that THIS fire is just as bad as ALL fire.
Vehicles may be a silly gimmick, but they're a really solid way of separating the different levels of player in your guild and they're a fun challenge and an interesting dynamic for boss fights. Of course, considering the number of people that hate Oculus because it's "so hard," adding more vehicles to future encounters may not be the best business strategy. Ah well.
Monday, March 9, 2009
Socket Bonuses: A Tricky Line
Socket bonuses are a very interesting mechanic from a design standpoint. Sockets in general ignore a few common rules of WoW, which makes balance somewhat odd, and bonuses only serve to complicate the issue.
Stats in WoW resist stacking. For example, an item of the same level may have 100 stamina or 100 intellect or 70 stamina and 70 intellect. This provides incentive not to stack one specific stat. For example, casters are eventually looking at a trade off of 2-3 spell power for 20-30 haste. Even if other stats are much less powerful, eventually the tradeoffs are high enough to make the stat worth it. This is, of course, given gear that even allows you to make such a choice.
Sockets, however, stand outside this law. A gem may have 16 strength, 24 stamina, or 8 strength and 12 stamina. No bonus whatsoever for not stacking a stat. However, sockets are a prime example of opportunity costs. Say you're melee DPS, and you want to stack strength. A socket bonus of strength on a piece of gear with a yellow socket may convince you to spread your gems out a little bit. The two big questions are what stat we put for a bonus, and what color socket we attach to it.
If a class wants a red stat, say spell power, and a piece has a red socket, you may as well put the socket bonus into the stat weighting of the item. For example, the socket bonus on Sash of Jordan may as well be part of the armor. Oftentimes this is done on specific pieces that blizzard intends to be powerful, implying that Blizzard intends stat stacking on that item. Another perfect example is the engineering tank gun, Armor Plated Combat Shotgun.
The opposite end of the spectrum is items whose socket bonus is useless and thus is always ignored. While Blizzard tends to be good about not putting spirit on warrior gear, stamina tends to pop up on DPS gear and is largely ignored. This was much more of an issue in the BC days than it is right now.
In general, Blizzard appears to pick one stat per color, for example a Druid will have spirit for Blue, spell power for Red, and intellect for Yellow. Then, an item with one red socket will have a bonus equal to 1/4 of a yellow or blue gem. If an item has, say, a red socket and a yellow socket, there will be a bonus equal to 3/8 of a blue socket. Not all gear follows this precisely, but quite a bit does. It's rather elegant.
This formula DOES have breaking points. Helms are presently excitingly odd. Take this piece for example: Obsidian Greathelm. With an 8 strength socket bonus on a blue socket (since the Meta isn't questionable) there is no reason to ever use a single-color gem. Placing a 16 strength red gem will net you 16 strength, while a purple 8 strength + 8 stamina gem will net you 16 strength and 8 stamina. If socket bonuses increase in power, we'll see more and more people using split-color gems rather than single-color gems. This would be a very elegant way to move some of the burden off red gems and on to orange and purple gems.
Notice any oddities on the socket colors and bonuses on your gear? Feel free to comment.
Stats in WoW resist stacking. For example, an item of the same level may have 100 stamina or 100 intellect or 70 stamina and 70 intellect. This provides incentive not to stack one specific stat. For example, casters are eventually looking at a trade off of 2-3 spell power for 20-30 haste. Even if other stats are much less powerful, eventually the tradeoffs are high enough to make the stat worth it. This is, of course, given gear that even allows you to make such a choice.
Sockets, however, stand outside this law. A gem may have 16 strength, 24 stamina, or 8 strength and 12 stamina. No bonus whatsoever for not stacking a stat. However, sockets are a prime example of opportunity costs. Say you're melee DPS, and you want to stack strength. A socket bonus of strength on a piece of gear with a yellow socket may convince you to spread your gems out a little bit. The two big questions are what stat we put for a bonus, and what color socket we attach to it.
If a class wants a red stat, say spell power, and a piece has a red socket, you may as well put the socket bonus into the stat weighting of the item. For example, the socket bonus on Sash of Jordan may as well be part of the armor. Oftentimes this is done on specific pieces that blizzard intends to be powerful, implying that Blizzard intends stat stacking on that item. Another perfect example is the engineering tank gun, Armor Plated Combat Shotgun.
The opposite end of the spectrum is items whose socket bonus is useless and thus is always ignored. While Blizzard tends to be good about not putting spirit on warrior gear, stamina tends to pop up on DPS gear and is largely ignored. This was much more of an issue in the BC days than it is right now.
In general, Blizzard appears to pick one stat per color, for example a Druid will have spirit for Blue, spell power for Red, and intellect for Yellow. Then, an item with one red socket will have a bonus equal to 1/4 of a yellow or blue gem. If an item has, say, a red socket and a yellow socket, there will be a bonus equal to 3/8 of a blue socket. Not all gear follows this precisely, but quite a bit does. It's rather elegant.
This formula DOES have breaking points. Helms are presently excitingly odd. Take this piece for example: Obsidian Greathelm. With an 8 strength socket bonus on a blue socket (since the Meta isn't questionable) there is no reason to ever use a single-color gem. Placing a 16 strength red gem will net you 16 strength, while a purple 8 strength + 8 stamina gem will net you 16 strength and 8 stamina. If socket bonuses increase in power, we'll see more and more people using split-color gems rather than single-color gems. This would be a very elegant way to move some of the burden off red gems and on to orange and purple gems.
Notice any oddities on the socket colors and bonuses on your gear? Feel free to comment.
Sunday, March 8, 2009
Blog Formatting
Given the size of my posts, I'm looking for a way to take up more horizontal space, so that each paragraph appears a little smaller and you're not staring at wall upon wall of text.
That said, don't expect my post sizes to decrease at all. I'm regularly cutting large sections of posts because while there's a lot of information to convey, I simply can't get everything I want to in a post of even remotely reasonable size.
Edit: Horizontal size modified, that might be a bit better.
That said, don't expect my post sizes to decrease at all. I'm regularly cutting large sections of posts because while there's a lot of information to convey, I simply can't get everything I want to in a post of even remotely reasonable size.
Edit: Horizontal size modified, that might be a bit better.
Blockin' It.
Warriors and paladins have shields. HURRRR. A large part of this article will be review and very tiny bits of math explaining the different ways we decrease damage. If you know this already, skip to the bold title a little bit down
Lets think about this for a second. Warriors and paladins have shields. Druids and death knights do NOT have shields. This means that druids and death knights are not capable of blocking, which in turn means that other stats must be modified so that we take the proper level of damage. While druids are not capable of parrying, that's a much weaker distinction, as parry is just another form of avoidance. In fact, parry is much weaker survivability-wise than dodge as it takes significantly more parry rating for 1% parry than dodge rating for 1% dodge and the DR cap on parry is roughly half that of dodge.
So, we have three manners in which we can reduce total incoming damage. The first is avoidance. Avoidance is dodge + parry + (opposed) miss. An avoided attack does not deal any damage. While avoidance is typically thought of as a straight percentage decrease to total physical damage, its power is inflated against enemies that hit for a large portion of your HP as having multiple strikes land in a row can be deadly, especially given situations where you cannot increase your HP to absorb an additional strike. Avoidance is based entirely on the RNG, and thus is extremely unpredictable and can be volatile at standard levels.
The second is percentage-based reduction. This works on any attack that is not avoided. The most common percentage is from armor. Increasing armor is a constant decrease, percentage-wise, on damage done. If I have 50% DR from armor, and move to 60% DR from armor, all physical damage coming in will be decreased by an additional 20%. Other forms of Percentage are many cooldowns. Shield Wall, for example, is a solid percentage-based DR, as is Icebound Fortitude, etc.
The last form of damage reduction is block. After an attack has landed, and has been decreased by all percentage modifiers, block is applied. Not all attacks will be blocked of course, but on a blocked attack, damage is reduced by a set amount; Namely your block value, which is [str/2 + shield block value (on gear) + the block number on your shield]. Of course, there are modifiers to block from talents etc etc etc. The point is, block is a static damage reduction.
So, this is why block is an issue. Death knights and druids only have avoidance and percentage DR, both of which apply as percentages to total incoming DPS. If we said that a boss does 50,000 DPS before armor and avoidance, and we know that a DK has 60% avoidance and 60% damage reduction from armor, we can prove that the total incoming damage is 50,000 * .4 * .4 = 8,000 DPS. Similarly, if a bear druid has 52% avoidance and 67% DR from armor, his total incoming DPS is 50,000 * .48 * .33 = 8,000 DPS. I'm excluding talents and cooldowns and things to illustrate a point; Including them is beyond the scope of this article. Anyway, by varying armor and avoidance but keeping the final result intact, we can balance classes to be equally powerful but with a unique feel.
Or can we? In the same situation, that boss is attacking a warrior. What's his final DPS? Well, we'll say he has 50% DR from armor and 50% avoidance. That's 12,500 DPS, but we haven't included block. But... we can't illustrate the effect of block without knowing the boss' attack speed! Let's say our trusty warrior has 2,000 block value. If the boss is attacking once a second, then we're removing 2000 every 1 second, thus we're down to 10,500 DPS. If the boss is attacking every 4 seconds, then our poor warrior isn't feeling that block much at all, he's still suffering 12,000 DPS. And if the boss is a ninja that attacks every 1/4 of a second, his block is inflated absurdly, decreasing our total to 4,500 DPS. Since Block is a static reduction rather than a percentage reduction, it is worth significantly more in situations with lower damage per strike, given equal damage per second.
I, personally, feel that the Avoid/Decrease%/DecreaseStatic set of damage reduction abilities is EXTREMELY solid. It provides us with three distinct and unique ways of mitigating damage, and allows us to reallocate based on the fight. Avoidance is much easier to stack than armor, and a percent of avoidance decreases damage just as much as a percent of armor, but it's spikier, causing healer issues, making the difference justified. Static damage reduction fluctuates based on what you're tanking, and you can have a high block set for trash or Heroics, and a high avoidance/armor set for tanking bosses.
What This Means
Everything up until now should have been review. But, it tells us some dangerous things. First off, DK and druid tanks will almost always have significantly higher DR from avoidance and armor than warriors and paladins, since they need to have enough DR to equal the effect of block. Since block is a static reduction instead of a percentage reduction, this means that bosses that hit ridiculously friggin' hard will almost always strike DKs and Druids for much less than they'll strike Warriors and Paladins. In the absurd other end of the spectrum, opponents whose strikes are very weak will almost always be tanked easier by Warriors or Paladins. This is part of the problem we're seeing with tanks right now. The other portion is the fact that magical damage doesn't affect all classes the same way. I'll get to that another day.
Anyway, regarding heavy hitting bosses, we can fix this by looking at warriors' block value. For example, in the explanation I gave before, a boss is dealing 50k DPS. A DK with 60% avoidance and 60% armor decreases that to 8k. A druid with 52% avoidance and 66% armor also decreases that to 8k. A warrior with 50% Avoidance and 50% armor decreases that to 12.5k. Given that his block value is 2,000, this means that for balance to be achieved, we can modify the opponents attack speed until damage is equalized. In this case, we need to remove 4500 damage, and we're blocking for 2k, which means that we need to block 2.25 attacks per second, or the boss needs to strike ever .44~ seconds. This is an extremely easy way to change the damage warriors and paladins take without affecting DKs and druids, and I believe it has already been stated that we'll be seeing this effect in Ulduar.
The Effect On Low-End Threat
But, one interesting note that doesn't get picked up a lot is that over time, block will become so powerful that low end content is decimated by it. Furthermore, since block is directly related to strength, warriors and paladins can increase their survivability in low-end, trivialized fights while increasing their threat. Forceful Deflection was intended to be a similar mechanic for Death Knights; Ditch some survivability for threat via strength and you'll get at least a little bit back. Unfortunately, Forceful Deflection works off a percentage decrease instead of a static decrease, thus its effect on survivability on low-end content is nowhere near as powerful as block. I don't know if druids have a mechanic that is at all similar to this. I'm not nearly as well versed with bear tanks as I ought to be.
I predict that you'll still see druids and DKs slightly overpowered tank-wise in Ulduar, but that bringing either one to tank older, farm-content is going to get more and more questionable. When we see druids and DKs tanking new content and warriors and paladins tanking old content, a small change will be made to normalize threat levels across varying content and increase warrior and paladin survivability. I'd really like to see the addition of static reductions for every class in unique, class-fitting manners, but I also feel like that won't be implemented as it's too much homogenization.
Lets think about this for a second. Warriors and paladins have shields. Druids and death knights do NOT have shields. This means that druids and death knights are not capable of blocking, which in turn means that other stats must be modified so that we take the proper level of damage. While druids are not capable of parrying, that's a much weaker distinction, as parry is just another form of avoidance. In fact, parry is much weaker survivability-wise than dodge as it takes significantly more parry rating for 1% parry than dodge rating for 1% dodge and the DR cap on parry is roughly half that of dodge.
So, we have three manners in which we can reduce total incoming damage. The first is avoidance. Avoidance is dodge + parry + (opposed) miss. An avoided attack does not deal any damage. While avoidance is typically thought of as a straight percentage decrease to total physical damage, its power is inflated against enemies that hit for a large portion of your HP as having multiple strikes land in a row can be deadly, especially given situations where you cannot increase your HP to absorb an additional strike. Avoidance is based entirely on the RNG, and thus is extremely unpredictable and can be volatile at standard levels.
The second is percentage-based reduction. This works on any attack that is not avoided. The most common percentage is from armor. Increasing armor is a constant decrease, percentage-wise, on damage done. If I have 50% DR from armor, and move to 60% DR from armor, all physical damage coming in will be decreased by an additional 20%. Other forms of Percentage are many cooldowns. Shield Wall, for example, is a solid percentage-based DR, as is Icebound Fortitude, etc.
The last form of damage reduction is block. After an attack has landed, and has been decreased by all percentage modifiers, block is applied. Not all attacks will be blocked of course, but on a blocked attack, damage is reduced by a set amount; Namely your block value, which is [str/2 + shield block value (on gear) + the block number on your shield]. Of course, there are modifiers to block from talents etc etc etc. The point is, block is a static damage reduction.
So, this is why block is an issue. Death knights and druids only have avoidance and percentage DR, both of which apply as percentages to total incoming DPS. If we said that a boss does 50,000 DPS before armor and avoidance, and we know that a DK has 60% avoidance and 60% damage reduction from armor, we can prove that the total incoming damage is 50,000 * .4 * .4 = 8,000 DPS. Similarly, if a bear druid has 52% avoidance and 67% DR from armor, his total incoming DPS is 50,000 * .48 * .33 = 8,000 DPS. I'm excluding talents and cooldowns and things to illustrate a point; Including them is beyond the scope of this article. Anyway, by varying armor and avoidance but keeping the final result intact, we can balance classes to be equally powerful but with a unique feel.
Or can we? In the same situation, that boss is attacking a warrior. What's his final DPS? Well, we'll say he has 50% DR from armor and 50% avoidance. That's 12,500 DPS, but we haven't included block. But... we can't illustrate the effect of block without knowing the boss' attack speed! Let's say our trusty warrior has 2,000 block value. If the boss is attacking once a second, then we're removing 2000 every 1 second, thus we're down to 10,500 DPS. If the boss is attacking every 4 seconds, then our poor warrior isn't feeling that block much at all, he's still suffering 12,000 DPS. And if the boss is a ninja that attacks every 1/4 of a second, his block is inflated absurdly, decreasing our total to 4,500 DPS. Since Block is a static reduction rather than a percentage reduction, it is worth significantly more in situations with lower damage per strike, given equal damage per second.
I, personally, feel that the Avoid/Decrease%/DecreaseStatic set of damage reduction abilities is EXTREMELY solid. It provides us with three distinct and unique ways of mitigating damage, and allows us to reallocate based on the fight. Avoidance is much easier to stack than armor, and a percent of avoidance decreases damage just as much as a percent of armor, but it's spikier, causing healer issues, making the difference justified. Static damage reduction fluctuates based on what you're tanking, and you can have a high block set for trash or Heroics, and a high avoidance/armor set for tanking bosses.
What This Means
Everything up until now should have been review. But, it tells us some dangerous things. First off, DK and druid tanks will almost always have significantly higher DR from avoidance and armor than warriors and paladins, since they need to have enough DR to equal the effect of block. Since block is a static reduction instead of a percentage reduction, this means that bosses that hit ridiculously friggin' hard will almost always strike DKs and Druids for much less than they'll strike Warriors and Paladins. In the absurd other end of the spectrum, opponents whose strikes are very weak will almost always be tanked easier by Warriors or Paladins. This is part of the problem we're seeing with tanks right now. The other portion is the fact that magical damage doesn't affect all classes the same way. I'll get to that another day.
Anyway, regarding heavy hitting bosses, we can fix this by looking at warriors' block value. For example, in the explanation I gave before, a boss is dealing 50k DPS. A DK with 60% avoidance and 60% armor decreases that to 8k. A druid with 52% avoidance and 66% armor also decreases that to 8k. A warrior with 50% Avoidance and 50% armor decreases that to 12.5k. Given that his block value is 2,000, this means that for balance to be achieved, we can modify the opponents attack speed until damage is equalized. In this case, we need to remove 4500 damage, and we're blocking for 2k, which means that we need to block 2.25 attacks per second, or the boss needs to strike ever .44~ seconds. This is an extremely easy way to change the damage warriors and paladins take without affecting DKs and druids, and I believe it has already been stated that we'll be seeing this effect in Ulduar.
The Effect On Low-End Threat
But, one interesting note that doesn't get picked up a lot is that over time, block will become so powerful that low end content is decimated by it. Furthermore, since block is directly related to strength, warriors and paladins can increase their survivability in low-end, trivialized fights while increasing their threat. Forceful Deflection was intended to be a similar mechanic for Death Knights; Ditch some survivability for threat via strength and you'll get at least a little bit back. Unfortunately, Forceful Deflection works off a percentage decrease instead of a static decrease, thus its effect on survivability on low-end content is nowhere near as powerful as block. I don't know if druids have a mechanic that is at all similar to this. I'm not nearly as well versed with bear tanks as I ought to be.
I predict that you'll still see druids and DKs slightly overpowered tank-wise in Ulduar, but that bringing either one to tank older, farm-content is going to get more and more questionable. When we see druids and DKs tanking new content and warriors and paladins tanking old content, a small change will be made to normalize threat levels across varying content and increase warrior and paladin survivability. I'd really like to see the addition of static reductions for every class in unique, class-fitting manners, but I also feel like that won't be implemented as it's too much homogenization.
Saturday, March 7, 2009
Blood Tanking: God Mode
Today's foray into the world of balance will focus on Blood Death Knights. This is ripped directly from a post I made on the Elitist Jerks Forums.
--
For a long time, primary tanking theory has revolved around EHP. The concept is simple; Stack Armor and HP so that you have a large buffer zone. That way, if your avoidance fails and you get hit 4, 5, 6 times in a row, you won't get destroyed because healers have enough time to react to the drop in HP. I have always been a huge fan of avoidance tanking. In BC, the major issue with avoidance tanking was rage starvation; As a warrior, if you don't get hit, you don't get rage, and thus you don't get threat. EHP guaranteed that you would be hit often enough to retain solid rage levels and thus solid threat. The second issue was spikiness. Even with high levels of avoidance, it's still possible to be hit repeatedly. With EHP, you have a solid buffer for those situations, whereas avoidance tanking requires healers to watch for spikes and get quick heals off.
There's no question whatsoever that you take less damage when avoidance tanking than when EHP tanking. However, healers have no issue right now maintaining high mana levels while pumping huge heals, and thus people consider the HP buffer to be more important than the decrease in damage brought on by avoidance tanking. The proposed changes to healer mana regeneration may be enough to push more people toward avoidance.
It's relatively well known that avoidance scales exponentially with respect to itself. For example, going from 0 avoidance to 10% avoidance is a mere 10% damage reduction. However, going from 70% avoidance to 80% avoidance is a 33% damage reduction. Because of this fact, as your avoidance increases, avoidance becomes a better stat to stack. And because avoidance works as a multiplier on HP, as your HP goes up, avoidance becomes a more powerful stat to stack. Which means in the long run, everyone will be stacking avoidance, eventually. To combat this, and in an attempt to prevent people from reaching the mythical and oft-videod 100% avoidance, Blizzard implemented diminishing returns on avoidance. As your avoidance gets higher, you need more and more rating to add an additional percent. However, there's a serious latent issue with the DR calculation. Diminishing returns on avoidance excludes anything that gives a % value. For example, a talent that adds 5% to dodge does not affect diminishing returns.
The previous three tanking classes were all equal in terms of avoidance from talents. Warriors and Paladins receive 5% each to Dodge and Parry, while Druids receive 10% to dodge (since they are incapable of parrying). However, Death Knights can receive a whopping 22% avoidance that is unaffected by Diminishing Returns; 5% to Dodge from Anticipation, 10% to Parry from Blade Barrier, 3% to boss-miss from Frigid Dreadplate, and an additional 4% Parry from Runeforging.
At low gear levels, say 20% avoidance from gear, this isn't a HUGE deal. If a warrior has 30% avoidance and a DK has 42% avoidance, The DK is taking roughly 17% less damage. This is a number that could easily be made up for via things like the ability to block, the innate damage reduction on Defensive stance, or a number of other factors. However, when we reach the extreme end of avoidance, say 55% from gear, we're in dangerous territory. A warrior at this point would have 65% avoidance, while a DK would have 77% avoidance. The DK would be taking 35% less damage than the warrior. As avoidance available on gear increases, the gap widens even further.
The issue of spikiness can be mitigated two ways. First up is the talent Vampiric Blood, which I still hold is wildly overpowered. The general rule is that an avoidance tank will have significantly lower HP than an EHP tank, and thus has less of a buffer zone. Vampiric Blood allows you to have 15% more HP for 30 seconds out of every minute. My standard high-avoidance set has 27k HP, while my high-HP set has 32k HP (both unbuffed). Vampiric Blood raises the avoidance set to 31k, drastically reducing the difference in buffers.
The second manner in which spikiness is reduced is the talent Will of the Necropolis. WotN can actually be considered to be additional HP, based on your current HP. For example, from 100%, your HP is increased by roughly 17%, as an attack dealing 117% of your HP will be reduced to (117 * .85) = 99.45% of your HP. The vast majority of the time, the talent will act as between half and a third of that value (5-8% effective increase in HP).
So, anyway, the whole point is that DKs can have way more avoidance than other classes due to talents that give us undiminished avoidance. Also, the spikiness issue is reduced by some exceptionally powerful talents. And lastly, DKs suffer no penalty from not being hit, and thus the only meaningful reason NOT to stack avoidance is threat concerns which are, honestly, laughable right now.
To test my theory that avoidance-based DKs are exceptionally powerful, I gathered a few friends to do some testing on Patchwerk. Not a kill, just 6 minutes of 1-tank, 1-healer PW.
Video: YouTube - Patchwerk 10 (High)
Healer was in about half blues ( Armory Link )
Final screenshot of recount showing avoidance levels: http://www.saichotictech.net/WoW/PatchwerkAvoidance.jpg
70% avoidance from regular melee, 73% from Hateful Strikes. As time goes on, it'll only get worse. Either Talents must be added to diminishing returns, or DKs need to have their avoidance talents nerfed significantly. Otherwise, we'll be pushing 100% avoidance with almost no concerns whatsoever come Ulduar.
On a side note, if any boss is ever given the DW attack penalty in WotLK, avoidance tanking will be even more hilariously overpowered.
--
As an update from the time I posted this, MMO-Champion is showing the latest news regarding Blood Tanking, and DK Tanking in general: Will of the Necropolis has a 15 second cooldown, significantly decreasing its power. Vampiric Blood has been nerfed from 30s/1m (50% uptime) to 30s/2m (25% uptime). This changes the worth of high avoidance stacking, since damage will be much spikier now. Furthermore, Blade Barrier has been dropped from 10% parry to 5% overall damage mitigation, leaving DKs at 9% standard avoidance and 12% max avoidance from talents, bringing us much closer in line with other tanking classes.
As a tank, I have a lot more to say about how stats and talents interact with each other and why I feel the warrior is the perfect balance of multiple stats, but that's for another day.
--
For a long time, primary tanking theory has revolved around EHP. The concept is simple; Stack Armor and HP so that you have a large buffer zone. That way, if your avoidance fails and you get hit 4, 5, 6 times in a row, you won't get destroyed because healers have enough time to react to the drop in HP. I have always been a huge fan of avoidance tanking. In BC, the major issue with avoidance tanking was rage starvation; As a warrior, if you don't get hit, you don't get rage, and thus you don't get threat. EHP guaranteed that you would be hit often enough to retain solid rage levels and thus solid threat. The second issue was spikiness. Even with high levels of avoidance, it's still possible to be hit repeatedly. With EHP, you have a solid buffer for those situations, whereas avoidance tanking requires healers to watch for spikes and get quick heals off.
There's no question whatsoever that you take less damage when avoidance tanking than when EHP tanking. However, healers have no issue right now maintaining high mana levels while pumping huge heals, and thus people consider the HP buffer to be more important than the decrease in damage brought on by avoidance tanking. The proposed changes to healer mana regeneration may be enough to push more people toward avoidance.
It's relatively well known that avoidance scales exponentially with respect to itself. For example, going from 0 avoidance to 10% avoidance is a mere 10% damage reduction. However, going from 70% avoidance to 80% avoidance is a 33% damage reduction. Because of this fact, as your avoidance increases, avoidance becomes a better stat to stack. And because avoidance works as a multiplier on HP, as your HP goes up, avoidance becomes a more powerful stat to stack. Which means in the long run, everyone will be stacking avoidance, eventually. To combat this, and in an attempt to prevent people from reaching the mythical and oft-videod 100% avoidance, Blizzard implemented diminishing returns on avoidance. As your avoidance gets higher, you need more and more rating to add an additional percent. However, there's a serious latent issue with the DR calculation. Diminishing returns on avoidance excludes anything that gives a % value. For example, a talent that adds 5% to dodge does not affect diminishing returns.
The previous three tanking classes were all equal in terms of avoidance from talents. Warriors and Paladins receive 5% each to Dodge and Parry, while Druids receive 10% to dodge (since they are incapable of parrying). However, Death Knights can receive a whopping 22% avoidance that is unaffected by Diminishing Returns; 5% to Dodge from Anticipation, 10% to Parry from Blade Barrier, 3% to boss-miss from Frigid Dreadplate, and an additional 4% Parry from Runeforging.
At low gear levels, say 20% avoidance from gear, this isn't a HUGE deal. If a warrior has 30% avoidance and a DK has 42% avoidance, The DK is taking roughly 17% less damage. This is a number that could easily be made up for via things like the ability to block, the innate damage reduction on Defensive stance, or a number of other factors. However, when we reach the extreme end of avoidance, say 55% from gear, we're in dangerous territory. A warrior at this point would have 65% avoidance, while a DK would have 77% avoidance. The DK would be taking 35% less damage than the warrior. As avoidance available on gear increases, the gap widens even further.
The issue of spikiness can be mitigated two ways. First up is the talent Vampiric Blood, which I still hold is wildly overpowered. The general rule is that an avoidance tank will have significantly lower HP than an EHP tank, and thus has less of a buffer zone. Vampiric Blood allows you to have 15% more HP for 30 seconds out of every minute. My standard high-avoidance set has 27k HP, while my high-HP set has 32k HP (both unbuffed). Vampiric Blood raises the avoidance set to 31k, drastically reducing the difference in buffers.
The second manner in which spikiness is reduced is the talent Will of the Necropolis. WotN can actually be considered to be additional HP, based on your current HP. For example, from 100%, your HP is increased by roughly 17%, as an attack dealing 117% of your HP will be reduced to (117 * .85) = 99.45% of your HP. The vast majority of the time, the talent will act as between half and a third of that value (5-8% effective increase in HP).
So, anyway, the whole point is that DKs can have way more avoidance than other classes due to talents that give us undiminished avoidance. Also, the spikiness issue is reduced by some exceptionally powerful talents. And lastly, DKs suffer no penalty from not being hit, and thus the only meaningful reason NOT to stack avoidance is threat concerns which are, honestly, laughable right now.
To test my theory that avoidance-based DKs are exceptionally powerful, I gathered a few friends to do some testing on Patchwerk. Not a kill, just 6 minutes of 1-tank, 1-healer PW.
Video: YouTube - Patchwerk 10 (High)
Healer was in about half blues ( Armory Link )
Final screenshot of recount showing avoidance levels: http://www.saichotictech.net/WoW/PatchwerkAvoidance.jpg
70% avoidance from regular melee, 73% from Hateful Strikes. As time goes on, it'll only get worse. Either Talents must be added to diminishing returns, or DKs need to have their avoidance talents nerfed significantly. Otherwise, we'll be pushing 100% avoidance with almost no concerns whatsoever come Ulduar.
On a side note, if any boss is ever given the DW attack penalty in WotLK, avoidance tanking will be even more hilariously overpowered.
--
As an update from the time I posted this, MMO-Champion is showing the latest news regarding Blood Tanking, and DK Tanking in general: Will of the Necropolis has a 15 second cooldown, significantly decreasing its power. Vampiric Blood has been nerfed from 30s/1m (50% uptime) to 30s/2m (25% uptime). This changes the worth of high avoidance stacking, since damage will be much spikier now. Furthermore, Blade Barrier has been dropped from 10% parry to 5% overall damage mitigation, leaving DKs at 9% standard avoidance and 12% max avoidance from talents, bringing us much closer in line with other tanking classes.
As a tank, I have a lot more to say about how stats and talents interact with each other and why I feel the warrior is the perfect balance of multiple stats, but that's for another day.
Introduction
In general, I will be illuminating issues with specific classes, where they're either too weak or too powerful, and comparing one class' potential against the potential in other classes. Then, I'll propose solutions, and see how close I get to Blizzard's final changes.
Issues will not always be balance. Sometimes, I'll highlight the effect of complexity, where one class spams one button while other classes have to perform difficult rotations or use abilities based on a priority system. Other times, I'll discuss areas where one class has a default state that feels incorrect, and explain what I feel the intended goal is for a class and how it's not measuring up.
I'm doing this for a variety of reasons. I often have an intriguing idea, work out the math for it, and have a minor revelation on the proper working of a class, but don't share it. What good is unshared knowledge? Other times, I'll discuss an idea with a few friends, and we'll come to a conclusion, and then only our guild benefits. Hopefully by making all this public, we'll get more input on our mad ideas, and we'll be able to help more people understand how best to play their class and how best to interact with other classes. Furthermore, I desperately want to work for Blizzard. I'm a software developer by trade, and working on WoW, or their upcoming MMO, would basically be a dream job. To that end, many of the things I discuss will delve into the terrifying worlds of Math and Programming.
To start off, I will be posting at least one entry a day for the next thirty days, without fail. Stick around, read up, offer suggestions, lets see what happens.
Issues will not always be balance. Sometimes, I'll highlight the effect of complexity, where one class spams one button while other classes have to perform difficult rotations or use abilities based on a priority system. Other times, I'll discuss areas where one class has a default state that feels incorrect, and explain what I feel the intended goal is for a class and how it's not measuring up.
I'm doing this for a variety of reasons. I often have an intriguing idea, work out the math for it, and have a minor revelation on the proper working of a class, but don't share it. What good is unshared knowledge? Other times, I'll discuss an idea with a few friends, and we'll come to a conclusion, and then only our guild benefits. Hopefully by making all this public, we'll get more input on our mad ideas, and we'll be able to help more people understand how best to play their class and how best to interact with other classes. Furthermore, I desperately want to work for Blizzard. I'm a software developer by trade, and working on WoW, or their upcoming MMO, would basically be a dream job. To that end, many of the things I discuss will delve into the terrifying worlds of Math and Programming.
To start off, I will be posting at least one entry a day for the next thirty days, without fail. Stick around, read up, offer suggestions, lets see what happens.
Subscribe to:
Posts (Atom)
