此站点大量使用JavaScript。
请在您的浏览器中启用JavaScript。
正式服
PTR
10.2.7
PTR
10.2.6
Beta
Soloing Kil'jaeden Mythic - a discussion about theorycrafting and optimization
来自 Durendillehunt
[Last Updated]：
2019/06/14
变更日志
补丁：8.3.7
目录
评分：
Greetings,
My name is Durendil, hunter soloer, and recently I spent a few weeks on Kil'jaeden Mythic, in the Tomb of Sargeras. And he was hard - in order to kill him, I had to do a lot of testing and theorycrafting. This allowed me to collect and organize my thoughts about how I approach boss fights in WoW.
This isn't a Kil'jaeden soloing guide - although it can help - but a discussion on the interactions between gear, hps, dps, strategy, and execution in a difficult setting, with Kil'jaeden as the example.
I didn't have enough survivability, and I didn't have enough damage. I was so close to impossible I couldn't tell if I could make it or not.
But I had a major issue: with patch 8.2 arriving, I calculated that as soon as essences became available he would become far more doable - for example, you get
清醒梦境
immediately on your arrival in Nazjatar, and it heals for 8k at neck level 50 (which doesn't seem like a lot but would have made a massive difference).
So I tried Kil'jaeden Mythic again and again. And I am glad I did. Because I got him, which was awesome, but more importantly it was both extremely fun and extremely profitable. Fun is subjective, but I hadn't enjoyed a soloing like that since Helya (normal in Legion and Mythic in BFA) or more accurately Lei Shen (the Thunder King), which was four years ago. And as for profit, something incredible happened: I learned new stuff about mechanics (mostly healing interactions), and also practiced new techniques - which, after 8 years of soloing, doesn't happen often (but isn't as rare as you think, challenging content is great for improvement). Although on the monetary side I think I recovered less than 100 gold on the 150k of consumables invested, so I'll probably have to go farm a bit.
Video
I would suggest watching the video after the guide in order to understand better what's happening, until you want to watch it twice.
Introduction
There are two main aspects to every world of warcraft fights, to which the other aspects are subordinated: survival and damage. In its basic form, survival is linked to HPS and is about keeping your health above 0, while damage is about DPS and beating certain thresholds. Usually, you focus on one or the other - trading off survival for dps and vice versa. However for less straightforward fights dps and survival interact in a more complex manner.
I've divided my explanation into parts for an easier read, but during each section I will explain how it interacts with others. I will first give the requirements of the fight. The second part will be about gear, and how to balance it around healing and damage. The third part will be about survivability, and how to maximize your healing output. Then comes strategy, and how to minimize your damage input. I'll end with damage, and how to optimize it when having to deal with all the previous requirements. And during that time I'll take examples from Kil'jaeden.
Requirements
Kil'jaeden is a ridiculously high-damage fight. He will chain cast
末日决战
(circles on the ground), which deals between 590k and 620k damage over 30 seconds (depending on soaking). With
灵龟守护
having a 144 seconds cooldown (with talent), there are many moments where you will have to heal through 2 consecutive armageddons, which can still kill you easily on the slightest miscalculation, and sometimes 3 armageddons. The high point is 1M8 damage plus 60K from
爆裂恐惧烈焰
(circle around character) and
聚焦恐惧烈焰
(beam), spread over 1min40. In term of pure numbers, that's 18.6K hps needed over that time. Note that the month before, I considered pushing 10k hps to be extremely hard as a hunter for the solo of Mythic Desolate Host. And in that situation, the 60k may not seem like a lot, but it did all the difference. Icing on the cake, even when I turtled, It wasn't easy to be full health for the next armageddons.
Then there is the boss health issue. He berserks at 14min30, but you can survive about 30 seconds. And you can't damage him in transitions. But he had enough health that if I went absolutely full survival, I would die from lack of dps (or I thought I would, as explained in the conclusion).
Gear
So, in order to prepare I needed the BiS of BiS items. For that, I had to scour the world since the sources were all over the place.
Azerite
The "best" (as in highest output) trait for soloing hunters is, hands down,
冷漠面容
. However, ilvl 400 pieces with
古树壁垒
(+other ring 4 defensive) give more hps overall than a 415 piece with bulwark.
I still decided on full impassive visage. IV has as much burst healing than AB, the latter being more of sustained healing. 415 pieces have more stats, so more damage and a bigger health pool. And you lose a lot of damage by using a primary trait for healing and not dps. Finally, you cannot farm it since you can only get it through the warfront and overwhelming odds, so I could only try with the shoulders, and confirmed it was not worth it.
The three pieces below all have impassive visage and great dps traits (I'll talk about those in the dps section).
Helm: I was using
拉胡艾的电化王冠
at first, from Opulence. But I switched to
黑暗收割者兜帽
, bought at the vendor, in order to trade
鲜血虹吸
(1,10% leech) for
压倒能量
- I needed more damage, haste was my best stat even with the loss of stacks from dots, and leech was the least impactful defensive stat (for example, versatility was worth around twice as much as leech).
Shoulders: The BiS is hands down
聚合畸变护肩
, from the vendor. It has
污秽传输
, which for dps is not great but not terrible, and gives a good amount of extra healing (3,6%).
Chest: There is a single chests that can be farmed to 415 with IV,
惊恶角斗士的链甲上衣
. As the "gladiator" shows, it's a pvp chest. Which I started doing this expansion for the express purpose of getting soloing gear. The chest is extremely easy to get through conquest points, but getting it to 415 requires getting to 2100 rating (which wasn't easy as a pvp newbie, and I didn't want to play survival so I had to do it as beast mastery). Then you need to stay at that level for weeks until you loot a
督军的战利品
in the weekly chest (or the azerite piece directly) - took about 5 weeks for me. On the plus side, I went from bad at pvp to average, so yay improvement.
Other gear
法索乌的防洪裤
: 405 ilvl, Restless Cabal. A lot of healing (5,9)%. You lose dps, but at over 90% health the proc deals damage so it's perfect for soloing, giving heals when it needs to and damage when it doesn't. I only had the ilvl 405 version since it didn't proc and I can't get it in mythic - my guild has decided to take it easy and simply farm Jaina's mount in order to arrive rested on the next raid. I really enjoy the new direction Blizzard is going with items.
镶钻折射棱镜
: 415 ilvl, the BiS++ trinket from Opulence. It absorbs a lot, and the absorb stacks when you're not using it. And it has agility. It did 16,6% of my healing. Now, getting an extremely low drop rate trinket as a hunter is hard, but once you've mastered begging anyone who can switch to tank loot spec to use it on Opulence (which, to be fair, they already did for the actual tanks), then begging the guild leader to award you the trinket once the tanks, the offtanks, and their alts loot the trinket, you may get it in a few months. Also make sure you have Hati use his puppy dog eyes. A great thanks to Salucmwa form my guild for looting it and trading it.
虚空石
: 410 ilvl, from Uu'nat. It's tricky to use (see healing section), but it's massively powerful (7,9% healing) and on demand. Only drawback is it lacks agility, but even with the healing absorb RNG (ex: legs proc at the wrong time) it's better than anything else (but still half as powerful as the prism). And you get it the same way than Opulence trinket, but by begging the healers this time. Luckily it has a way better drop rate.
Stats
For this fight, the value of the stats were as follow:
haste: best dps stat (for me - stat weights are all over the place for BM hunters)
versatility: best survival stat
agility: more damage and more healing, but less worth than versatility
leech: not bad, but since BM hunters only benefit from about 40% of leech (insert joke about pets doing all the work), not worth it. You can increase its worth with mend pet, which is affected by leech, but if I use it on CD I lose about 1-2k dps due to GCDs.
In the end, I decided on 2 Haste gems, 1 versatility gem, 1 agility gem (mistake, should have gone for versatility), 2 versatility ring enchants, haste weapon enchant.
I also went for versatility food (
香薰鲷鱼
), which gave more survivability than the 100 agility from feast (and was way cheaper).
I finally chose a ring with 2% leech (2,1% of healing) over another one with 0,5% more simmed dps.
I finished with 10% haste (which is still low, but I had a lot of haste procs), and 9,54%/4,77% versatility, which equals to 400k damage avoided + 670k more healing, for a net gain of 1070k. All in all, with this and the azerite traits I sacrificed about 2% survival for about 2% dps.
I used the usual dps consumables - potions, scrolls, drums, rune. No Vantus though, since my guild leader still wants us to Vantus Jaina. I'll talk about healing consumables later.
Dwarf didn't give benefits for this fight. I used
天生狂野
to reduce turtle cooldown.
As you noticed, at some moments I chose dps over survival and others it's the opposite - it was about finding balance.
Healing
Abilities and items
I had to organize my healing very carefully in order to survive each armageddon wave (consecutive armageddons without immunity) and get back to 100% before the next.
Listing all the sources of healing would be nice, but useless. To optimize, I had to understand and use the interactions between the sources of healing.
For the unaware, on each cast armageddon creates 7 circles. Each unsoaked circle deals 16k damage and puts a "big dot' for 30 seconds. If you soak a circle, you suffer a "small dot", and you take less damage overall and it is delayed. But the small dot lasts for 1 full minute and stacks.
冷漠面容
,
镶钻折射棱镜
and the small dot: IV is a powerful heal that can crit, but it only procs when taking damage. Absorbs do not count as damage - so it won't proc when you have the shield from Opulence. But it will proc the moment it breaks. This is bad when the shield is active and the big dot is over, since one of your major sources of healing isn't working. If you take no stack of the small dot, the shield will recover between armageddons. If you take one stack, the shield will recover slower but you won't heal through IV. And if you take 2 stacks, you will heal and take nearly no damage - but the shield won't stack until it is over. So the question is: is it better to maximize IV heals by soaking two armageddons in a row or is it better to stack shields between waves ? Since not soaking = death, 1 stack was a necessity. And in that situation, 2 stacks was worth it - I had to recover fully between waves, even when I had "only" two in a row.
意气风发
and
灵魂治愈
: from an abstract point of view, since spirit mend has the lowest cooldown, you should use it first, then use the big cooldowns when necessary. But the issue with big cooldowns is that you need them to be up at set timers. In this fight, the cooldown of exhilaration wasn't an issue for a double armageddon since I could only use one. So I used spirit mend first and exhilaration when needed. But for triple armageddons, due to
自然治愈
, I could use two exhilarations. Yet if I delayed the first I would die before it came off cooldown. So I used exhilaration first, and delayed spirit mend pet. Which brings us to the problem with delaying. If I used exhilaration ASAP, I could lose healing by going over 100% (overhealing). If I used it at 70% and had an IV critical proc (or legs), I could also overheal. So I chose to wait until I had about 60%. Same with spirit mend - I had to use it as soon as possible, but it could crit - so it's better to wait until a crit wouldn't overheal. But if I delayed CDs too much I wouldn't have them available when needed, so I had to make calls.
There is also the interaction with
优胜劣汰
. It seems great, and its best value is when you aren't soaking (about 35k less damage). But if you use it when you already have to delay your heals, you will take less damage - and therefore delay more or overheal. And more importantly it has a long CD. Therefore, I had little choice if I wanted to use it as often as possible, except for the first triple armageddon, where I chose the second one since I didn't need it to get back to 100% before the second, but it wasn't enough to allow me to use a single spirit mend and have the cooldown ready for the second armageddon if I used the damage reduction on the first one.
灵魂治愈
and
狂野怒火
: this is a technique I used for the first time, using
狂野怒火
. Now, one of the best azerite traits hunters have is
怒火升腾
. It increases agility by a lot: therefore, it increases pet healing by a lot too since spirit mend is based on attack power (13%/stack, I had 2). Sadly, Bestial Wrath doesn't also increase pet healing by 25%, even if it does increase the healing from most of the hunter's base abilities (just not those that actually heal). It's worth mentioning because a similar ability -
野兽守护
- does increase spirit mend by 30%, so remember that sometimes spells have effects that are not listed in the tooltip. Anyways, it was important to delay healing for BW in order to heal more - unless it would delay too much. Finally,
野性守护
probably increases spirit mend crit chance by 10% (just the direct healing, not the HoT). But that's not a given - for example, since it increases crit chance of specific abilities, as well as "melee and ranged crit", I don't know if it increases the crit chance of, say, Impassive Visage. And other spells cannot crit, such as potions (not sure why, but I tested with a lot of them) or Exhilaration (% based healing).
However, there is another issue that will never come up in theorycrafting but does in the game, because it is so counterintuitive that it can only be discovered though testing. It's something I call "pet refresh rate". Now, a while ago I discovered that pet leech procced every 5 seconds instead of every second, which is bliss when you are tanking with a ferocity pet that relies on his massive leech to stay alive. Then he dies despite having a ton of leech healing that didn't heal him, and now you have half the island's map rushing you (island expeditions are actually quite fun as a BM hunter, you can take massive amount of mobs in normal with a ferocity pet). I'm mentioning it because it shows that blizzard coded pets to take less resources by checking their status less often - which is great for the servers when there are 30 people on a boss, but, unless I'm missing something (which, as an IT engineer, I know could be very possible), it serves no purpose in a small group. Now let's talk procs and pets. You may think that, if a proc increases your power, it also increases the pet's power? That if you get an agility buff it increases your pet's agility? Yes but no: it will, eventually. It's about a 1-2 seconds delay when it procs, and a 1-2 seconds delay when it falls off. Globally, you won't notice it. However, when I used BW and then immediately used my heal, I was getting less healing on the direct heal than expected. So, overall I discovered it by analyzing the replay, then tested it, and it would have been
rather
nice to know beforehand. But back to the spells.
虚空石
and pretty much everything: The void stone is a on-demand 145k absorb, which is absolutely monstrous. But it can absorb healing. So if I use it and then my legs proc and crit, I lose 20k healing and will take 20k more damage, which is more than a healing potion (36k). Luckily it doesn't proc IV. To keep it simple, I had 3 choices:
use it right before impact, having the shield break in 2 seconds
use it at impact, since it immediately breaks my other shield and procs impassive visage. During the shield, my other shield starts stacking and my cooldowns come back
wait an opportune moment, right after an IV proc
I found the second option to be the best. It's a bit riskier, since bad rng can kill you, but it doesn't happen too often and given how tight it was I had to take a few risks. I did have to do the third option if spirit mend and the trinket came off cooldown at the same time - the heal takes priority. And finally, due to the 2-minute cooldown I had to go with the first option once and precast it in order to have it back up before certain death.
It wasn't enough.
On-use consumables
Despite all my theorycrafting and tries, I couldn't survive the triple armageddon in phase 2 (the 18.6k hps one) or the last double armageddon (also about 18.6k hps, since they are very close together). And it's in this situation we use items that can only be used once per fight, at the moment they are needed.
微型等离子护盾
: Engineering shield. Affected by attack power (therefore agility) but I needed it at a specific moment so I didn't use the BW trick.
治疗石
: Healthstone! If you're not familiar with the technique, you can enter the proving grounds through a class trainer or specific npc, and grab a healthstone with three charges. So use the Garrison hearthstone, talk to the mission table npc, enter the proving grounds, use the Dalaran hearthstone, and get back in the raid any time you're out of charges. It's not very fun, and therefore, I never used it before in soloing (great for Mythic+ though), but KJ was my Godzilla threshold on the matter.
海滨治疗药水
: Your trusty healing potion. Does not share a cooldown with the healthstone. For some reason it never crits, and I'm pretty sure it could crit in Legion. Is that new, or is my information incorrect ?
海滨活力药水
: Does not share a cooldown with healing potions, but does share a cooldown with the dps potion. More on that below.
深海绷带
: I used it to recover if my health was below 100% for an armageddon, but only if I didn't have any dots. It's also a dps loss since it is channeled.
The entire survival strategy had one secondary objective: do not use the rejuvenation potion. I had to survive the triple armageddon with only the belt and healthstone (most powerful ones), and the last double with the healing potion.
For that, other than maximizing the use of my cooldowns, the important part was to plan the strategy.
Strategy
triple armageddon at start. The turtle on the first armageddon of the transition was non negotiable, and if I used turtle on the first armageddon, I had to do it on cast, losing 8 seconds of dps, and I had to delay the transition by about 10 seconds. With the triple armageddon, I could push the boss 15 seconds faster - and we'll see in the dps section that seconds is the greatest dps resource. Since the armageddons are far apart, survival was tight but doable with the right rotation.
I then used turtle immediately after the armageddon cast in transition. Since the next armageddons are far apart, it allowed me to only have a double armageddon afterwards. I also dpsed the boss a little, despite his immunity, to get my cooldowns and exhilaration back up.
the double armageddon wan't very hard to survive, since I had all cooldowns except the weakest (-20%, that I needed for later). However, during a certain period of time I couldn't dps the boss at the start, so no healing procs. This is also an example of me not being able to delay spirit mend - I needed it up for the next armageddons. And finally, due to the damage of his beams (50k) being full health for next armageddon required taking no chances - for example, I also turtled the dreadflame (circle), despite it probably not being worth losing 4 seconds of dps.
then came the difficult part, which is usually where I wiped: the triple armageddon. First, I made sure to go in at 100% health even if it took a bandage. Then delayed exhilaration and spirit mend - exhilaration so I had two, then spirit mend. I couldn't delay spirit mend more than about 20 seconds after the explosion, since if I did there would come a moment where I needed healing but it was still on cooldown. I used the shield on the explosion to make sure I didn't dip too low to have time to heal back. Then I had the issue of the heal and the trinket coming off cooldown at the same time. I chose to use the trinket first - timing it right after an IV proc - because I could delay the spirit mend until 15 seconds before the next impact (therefore having it available again in the middle of the dot) and use it during bestial wrath, maximizing my healing. Then came the usual wipe moment: I had all my healing on cooldown, and had to use the healthstone to hold on long enough. I was two ticks away from death - every crit, proc or even the +25% from the bestial wrath trait really counted, and so did every bit of healing I did during the first two armageddons - if I hadn't been at 88% health on impact (after healthstone), I would have died. After the heal got through, I had to focus on spending focus as fast as possible in hope to get exhilaration off cooldown in time.
after another turtle and getting back to full health while making sure I had spirit mend up for the next impact, I had a double armageddon. But the difficulty lied in the fact that I would get only 2 spirit mends and not 3 - which made it as hard as the triple armageddon, especially with only a healing potion left. I had to use exhilaration just to have enough health to survive the second impact, as well as the -20% defensive - I planned the timings of this defensive throughout the fight for the express purpose of having it here. I kept the trinket for after the heal had passed (+ IV proc). I could have exchanged the second heal with the trinket, but even after the big dot fades I had the little dots ticking for maybe 3 ticks - so having spirit mend up again was very important in case I ended up with 3% health (which would have been possible with one less proc/crit). The healing potion saved my life on the kill.
Now, even after choosing the armageddon rotation, and which CDs to use, I had to determine wether to soak or not. First, soaking is worth it: an armageddon deals 620k if not soaked and 592k if soaked. But in case of double soak, you can add 10-15k due to refreshing the stack. I have already explained why I went for two stacks, but there is also the fact that the little dot lasts for 1 minute, so it delays some of the damage - and once the big dot is gone, you are safe (but need to get to work on getting back to full). Finally, the last armageddon in a wave needs to be soaked - 7 big dots and 1 little dot is not survivable, and even if you refresh the little one 6 big dots and 2 little dots deals 4,1% less damage than 7 big ones (so 20k less damage during the big dot, which is huge when you are hanging in by a thread). So when there was a triple armageddon, I soaked the last two - use cooldowns on the first, use trinket + soak on second to delay damage have enough health to survive through the third until cooldowns are back up.
The last survival issue was to avoid damage from the
爆发的映像
s. They usually miss you with their cast, so I kept them alive as long as possible for dps purpose. But not too long, since if they do hit you it deals a lot of damage (it increases each cast, not each hit). Finally, their (red) orbs seem harmless but if timed with a bump from the black orbs it will probably send you in a red orb, probably starting a snowball effect where you take damage from the the explosions beneath your feet - and each hit is 16k damage. It also deals a little unavoidable damage, so no killing them during the P2 triple armageddon. But on the plus side, if killed right as all dots fade it can trigger and extra IV proc.
I won't talk about transition 2 and phase 3 in this section since survival is easy until berserk - or to be more precise, if you make a mistake you fall off the platform and get one shot or you lose time and get destroyed by the berserk, so healing isn't really going to help. I'll just say, for those who know the boss and are confused, that due to soloing him the rifts spawned through
撕裂裂隙
had no energy and disappeared as soon as they soaked a fire patch - which meant I kept stacking
千魂之暗
, which did massive damage after berserk.
Damage
Most people think dps is only about gear, rotation and cooldown management. To understand BM hunter dps, the one thing to know is
倒刺射击
- the more you keep the stacks, the more overall damage you deal. It has two charges, so if they drop you want to wait at least until you will be able to refresh the stack of the first shot with the next one. I had one trait of
吞食狂热
to make it easier.
However, raidbots sims and a real boss are very different situations. To understand the difference between theorical dps and practical dps I'll first start explaining how we lose dps.
Losing damage
By losing dps, I'm talking about mistakes or decisions that reduce the theorical maximum output your character could be doing - not unavoidable mechanics such as the boss having a shield, but being in a situation where you could be doing more damage, but aren't.
First, let's talk about deliberate dps loss.
As explained before, my soloing gear is not my raiding gear - I chose to sacrifice dps for healing, considering that I'd gain dps in the long run - either by not dying or, in the case of the first triple armageddon, pushing the boss faster by playing more aggressively. I'm also not going to use examples not present in this fight, such as delaying CDs for a burst phase.
You can lose dps by choice, deciding you'd rather use a GCD for a heal than for damage. For example, in phase 2 I was solely focused on healing. There came a time where I should have waited for my barbed shot, but using it immediately would have reset the cooldown of my BW, which would have increased the healing of my heal and allowed me to use it sooner. So I decided to go for the healing, and when the shot was off CD again my stack had dropped. This happens a lot during the fight, and since the shot can proc, the timings can change and therefore you have to make fast decisions. I cannot say if it was the right one.
You can also lose dps because of focus. You have your limits. If you are focusing on survival, you can the risk of making a mistake in the stacking, rotation or focus management. But if you focus entirely on dps, and make a mistake survival-wise, you can die. It's the reason why my raid leader has two calls he uses often: "Bourrez le boss" (smash the boss) for burst phases and "Lachez le boss" (forget the boss) when he needs us to stay alive - and is more than willing to lose dps if it means survival.
Then there is non-deliberate dps loss, or errors. Did you mismanage focus ? Mess up the rotation ? Lose a GCD ? You lost dps. However, it also means that if you fix these errors your dps will increase.
For example, the phase 3 of the kill I was extremely focused on the strat since it was my third phase 3. I knew I had just had a perfect transition, so I was trying too hard not to mess it up. Hence the many small mistakes, such as:
there is a moment where I used cobra shot instead of barbed shot, lost my stacks, and then to make things worse I used barbed shot, also losing that stack since the next charge was too far away
I lost line of sight or range of the boss at least 3 times, costing me about 0.5 seconds each
on berserk, I should have feigned death sooner after the turtle.
The boss did die, and it's what matters. But what if he didn't ? How do you dps more ?
All things considered, I made a lot of small mistakes. I could have done more damage - not by getting more gear or other, but by playing better. And that's the most important part, which is something I feel a lot of players forget (or at least the vocal ones). When it comes to gameplay we talk a lot about rotations and generalities, and so do guides because they need to be broad. But you aren't trying to down a target dummy - you are facing a very specific boss. And there is a difference between "doing the strategy right" and that plus "while doing as much damage as possible".
Increasing damage
Before continuing my previous thought, we need a frame of reference. When I sacrificed some survivability for damage, and switched a few azerite traits, I gained 2% theorical dps.
Then there is clicking. Now I've had a few good discussions on keybinding vs clicking - meaning ones with actual arguments that apply to my case (as opposed to generic comments that aren't helpful). Although I do use macros and keybind off-gcd spells (traps, interrupt, panic buttons), I don't do it for spells I nearly never use. Examples: when using the healing potion, I lost about 0,3 seconds on a GCD. And if I have to disengage with my back to the boss (which happened once in the fight), I also lose about 0,5 seconds of line of sight - I usually plan ahead, but it still did happen a couple of times. Now, switching to keybinding would have a refractory period, but these discussions have convinced me that I could gain a small amount of dps by going full keybinds, under the condition that I never make a mistake or lose time hitting the keys. Let's be broad and say 2% in an extreme case, since it's an estimation.
So we are at 2% thanks to gear + 2% possible by overhauling my gameplay (which I quite enjoy). That is a lot, and can make all the difference. But now I'm going to use math to show why execution is far more impactful.
Execution
Frame of reference
Let's take a common situation: a guild kills Jaina Mythic. They succeeded because they were 20 alive at the start of phase 3, with only one dps dead 2/3rds in (sacrificed for siegebreaker), and because the rest survives until the last seconds. Then they nearly all die, but they barely managed to do enough dps for victory to be claimed. On their fourth week, they have more or less the same gear - it's been months of farming. But one dps dies in phase 2, and there is no rez. Yet the boss dies with 19 people.
So I looked at the logs of phase 3. I only considered boss damage, and this is the result.
First kill
Fourth kill
Difference
Raid dps
325 551
351 303
7,91%
Average per dps
22 207
25 093
13%
To be more precise, if I only look at the 8 dps that were present on both pulls and didn't die on either, the dps increase is 6,7% average - but among those we have two people at -13% and three others at +13%.
Now, if you believe dps to be a fixed thing - "oh that guy does 20k dps", then the only explanation is that on the first kill many players had no idea how their character should be played and they suddenly learned in the following weeks (except the two who lost dps, they must have forgotten about some of their buttons).
Or you consider that the "players' dps" is an abstract concept, and what matters is the players' dps on Jaina Mythic in phase 3. That on the first kill, people were nervous, and, due to this being the most clean P3 start we had yet, didn't have the automatic reflexes you develop after 100+ pulls in a phase - so they focused on the strategy and survival far more than on their dps. And that's how you do 13% less damage than you could be doing if you knew the fight like the back of your hand (and down the boss, which is why strategy is way more important than output). It's all about the execution.
So, coming back to Kil'jaeden: could I have dpsed more with better execution ? How much dps would I have gained simply by playing better ?
Theorycrafting
Not counting pulls where I got to the second transition and died to the armageddon dots, I had three pulls where I got to phase 3. Let's check the timers. Note that on the first one I was using Heroism in phase 2 since I was doing 2 triple armageddons in this phase.
Also, I did all the maths using the blizzard UI from my recordings, not recount. For example, in the kill video Kil'jaeden takes (18966-15138) = 3828k damage, plus 249k for the reflection, yet I have 3690k damage on my recount. I don't know why, and if someone has an idea don't hesitate to comment, but in the meantime I trusted in the in-game health bars.
However I did take the same boss health for the three lines by choosing the average boss damage taken between transitions - 3810k, 7650k and 7500k. It makes hypotheticals possible and the variance was minimal.
Finally, the dps numbers can seem pretty high - a few weeks ago I did a pure dps race on M.O.T.H.E.R, doing the equivalent of 23,3k dps in full damage gear (not counting increase from room 3). However I usually have significantly more dps in legion raids, which could be due to the armor of Bosses not scaling (unconfirmed).
P1end
P1 dps
P2 end
P2 dps
T2 end
T2 duration
P3 end
P3 dps
19% try
02:23
26,64k
9:10
24,60k
10:53
01:43
13:08
27,70k
6% try
02:17
27,81k
09:16
23,68k
11:00
01:44
15:01
26,36
kill
02:10
29,31k
09:13
23,39k
10:25
01:12
15:03
26,98
So, let's see if I could have done better phase by phase:
Phase 1: I had pretty much mastered phase 1, and nearly always got the boss between 2:08 and 2:12. Managed a 2:06 once but it was probably RNG. So I'm going to keep the 29,31k dps as my realistic (non-rng) maximum.
Phase 2: I was focused entirely on survival here, in order to see phase 3 again, so even if I could dps more I'm not going to pretend that I would have managed it. Therefore, I'm going to take the best value between the tries with the final strat, 23,68k, since it's the value I proved I could do.
Transition 2: my transition was absolutely perfect. Ideally, you want to always be dpsing the shadows, with a 100% uptime of Illidan's buff. But you also need to manage the bumps - avoid falling off, but also have them bump you on Illidan to refresh the stack, or use
逃脱
to land on him. So I'm going to assume I couldn't have done better, despite it only being my third successful transition.
Phase 3: unlike the rest, I had a massive amount of improvement possible in phase 3. I was at 23% after two minutes on the kill, while I was at 20% during the first try at that time. Given that that first try ended with me being blown off the platform by a
恶魔方尖碑
, I triple checked every spawn - while I could have simply checked once. I also could have done a strategy using disengage by predicting the possible bump directions. So, to calculate what dps. Also, since my cooldown rotation is two minutes, then the 27,70k is an accurate representation of my dps over and entire cycle. But I did that without heroism or the potion. Therefore, I could definitively reach 28,70k, especially given what I showed in phase 1 with a full 2-min burst with hero/potion (and where I also lost a bit of boss dps due to the add). Add the fact that my main damage ability deals +50% more damage in phase 3, and we can confidently say that my maximum dps is 29k (especially since I have no need to focus on survival).
So what was my actual best possible performance? And how does it stack up to a flat +2% increase ?
P1 dps
Time
P2 dps
Time
T2 time
T2 end
P3 dps
P3 time
Seconds gained
Kill
29,31
130
23,39
552
72
624
26,98
902
0
Optimum
29,31
130
23,68
548
72
620
29
879
23
2% increase
29,90
127
23,86
543
72
615
27,52
888
14
Conclusion
One second gained is about the same as a 0,14% damage increase.
Healing-wise, I consider my execution to be near perfect since if it wasn't I would have died. Between bandages and pre-turtles, I lost about 7 seconds. My pet was still doing a little dps so let's say 4 or 5. So better healing would have been worth about 0,5%-0,7% dps.
From a practical point of view, was sacrificing survivability for a 2% increase worth it ? Without it, I wouldn't have gotten that third kill. But maybe I would have had less trouble in phase 2, increasing my damage a little, and seen phase 3 more often, increasing my damage by a lot.
And had I still lacked dps, I could have simply dpsed more. No need for better gear, massive overhauls or other, simply better execution.
But talking of execution, the deciding factor was the transition. If 10 seconds is worth 1,4% dps, then cutting the time by 30 seconds was what made it possible dps-wise.
In other words, going for more dps at the expense of survival may have been a mistake. I would have wiped on that third phase 3, but I would have seen far more phase 3s overall, leading to a better mastery of the fight, and therefore more damage. But I can't know, since I would also have had to reliably manage the second transition near-perfectly.
I didn't do everything right. But I did some things very right and that's what allowed me to down the boss. And it's nice to know I had some margin for improvement if I came up short. But as my raid leader said to the players that dealt 13% less damage than they could have during the burn phase: "GG" (plus a few expletives, but they were aimed at the boss). If a boss dies, it dies, and she wouldn't be more dead had she lasted 3 seconds less - but she would be alive if anyone had made one more mistake regarding the strategy, since we were out of rezzes due to previous mistakes.
So if you're having trouble with your damage on a fight, make sure you do the strategy right and survive, but then, if you want to improve your dps, play
the fight
better. Gear will improve your dps. Maybe changing your play style will increase your dps. But you can also improve your dps with better execution, or make it so you need less. Or improve your survival, and gain precious seconds. Or whatever else you need: there is not a single way you must pursue. Understand where you can improve, and choose some or all - but don't chase one thinking it will solve all your problems. Because whatever your objective, once it's accomplished, what matters is that you did it, and hopefully you had fun.
About the author
If you are a hunter, you can check out my soloing guide at
Hunter Extreme Soloing Guide - Battle for Azeroth
.
Durendil
has been playing his hunter since Vanilla. During WotLK, he decided to try soloing because the alternative was playing another class. He started with BC heroic dungeons, and followed with
熔火之心
and
卡拉赞
. It was then that he ran out of guides, so he started creating his own strategies for Zul'gurub (level 60 version), Zul'Aman (level 70 version) and
预言者斯克拉姆
, which required nature resistance gear. He rose as one of the top two hunter soloers along with
Caribald
, and made a name for himself in the hunter community thanks to his first hunter
solo of The Lich King
. He continued soloing expansion to expansion, getting the first hunter kills on
拉格纳罗斯
(10-25H),
雷神
(normal and heroic), and
海拉
. During that time Caribald retired from soloing and
Shoot
made his debut.
He plans on keeping on soloing and pushing his hunter to his limits.
标签：
PvE
高级
[Get Wowhead]
高级会员
[$2]
[A Month]
[Enjoy an ad-free experience, unlock premium features, & support the site!]
评论
评论来自
Vikinng
Awesome thank you so much
评论来自
struds
Congratulations, and thankyou for the thoughtful article :-)
评论来自
Dauthus
I must have missed it, what spec was he using?
评论来自
Cacarot
Any tips about the purple Orb pattern after KJ takes flight? I know there's 6 orbs. What is the pattern? Is it a set pattern? Does it change with each attempt? Is it set per lockout?
贡献
在发表评论前，请留心以下提示：
您的评论必须为简体中文，否则将会被删除。
不知道如何发评论？参考我们的
格式指南
！
发表前最好先自行校对一次。
有问题可以访问我们的
论坛
来寻求帮助。
发表评论
你没有登录。
请登录
或者
注册账号
来添加你的评论。
使用下面的表格浏览您的截屏。
[Screenshots containing UI elements are generally declined on sight, the same goes for screenshots from the modelviewer or character selection screen.]
质量越高越好！
[Please review our
Screenshot Guidelines
before submitting!]
您没有登录。请
登录
后提交截屏。
将视频URL输入下列表格即可。
URL：
支持：仅限 YouTube
说明：您的视频需通过审核才能在站点上显示。
我们用
Wowhead 客户端
保证数据库的及时更新，向您提供额外的有趣的功能！
两大目的：
它还维护WoW的一个插件
Wowhead Looter
, 在您游戏时采集数据！
它将
采集所得数据
上传至Wowhead，保证数据库时刻掌握最新信息！
您可以用它追踪完成的任务、配方、坐骑、伙伴宠物以及头衔！
您还在等什么？立即
下载客户端
整装待发吧。
我们用 Wowhead 客户端保证数据库的及时更新，向您提供额外的有趣的功能！
两大目的：
您可以用它追踪完成的任务、配方、坐骑、伙伴宠物以及头衔！
您还在等什么？立即 下载客户端 整装待发吧。