[2.3] Everything you need to know about Natalya Optimization
Aug 25, 2015 11:13:05 GMT 1
Post by wudijo on Aug 25, 2015 11:13:05 GMT 1
[Related Links]
Diablofans Link
Reddit Link
Video:
Greetings all,
you might remember my post about UE + M6 optimization in the beginning of S3, and I thought it’s about time that I do some stuff like that for Natalya as well. Now that the end of season 3 draws close, I guess many are trying their best to get on the leaderboards, and I decided to put some more time into it. N6 has been used basically the entire season for solo and speedruns, and later on also got included in group pushing as well, effectively competing with UE for the top DH set in a tight race. This post mainly focuses on greater rift pushing, which generally requires the most optimization (and low level content farming efficiency has more parameters than just pure dps). It contains points for the current situation (patch 2.2), but I’ve added a lot of things to look out for in the next season (patch 2.3) as well, and included a TL;DR at the end. I was working on this as a side-project for quite a while but never quite managed to finish it, so here it is.
The structure of the post is as follows:
Enjoy.
[Natalya For Dummies]
The Natalya set heavily relies on Rain of Vengeance (RoV) to deal damage, and generally you aim at reducing the cooldown as much as possible using the 2-piece bonus. To achieve this, heavy CDR stacking is required, and a combination of Strafe + generator (mostly Evasive Fire (EF) for increased hatred regen) has proven to be the best, because like that you can get more attacks per time unit than you normally would when only attack with a regular skill. The RoV, combined with the Crashing Rain (CR) belt (especially in solo), is your main source of damage. The generator casts and Strafe are basically irrelevant, and because of that you don’t require a weapon damage roll on your offhand weapon (RoV and CR scale only off your mainhand weapon damage). In your offhand, you generally take a Balefire Caster when you are the only dps (solo / 2man / maybe 3man) and a Calamity when you are not (maybe 3man / 4man, or whenever you are using a non-fire rune). The reason why a Balefire Caster is better is because elemental damage is a separate damage multiplier and you generally only have 20% from bracers (or 40% if you got lucky with your amulet), while you have lots and lots of additive damage (30% Strongarm, 55% RoV, 20% Steady Aim, 30% Wolf, in groups even more) which is the same category as Calamity.
Completing a combo of Strafe + EF + RoV is called a cycle, and there are mainly two variations: a 2:1 cycle and a 3:1 cycle. For a 2:1 cycle, that means you cast 2x EF and 1x RoV while Strafing permanently, then repeat (3x EF for 3:1). That playstyle is a little bit difficult to get into, but once you get into the rhythm things become a lot easier and you can actually focus on other things much more during the rift. A reset is what the N2 procs are called, and you generally try to chain them in a way that you start and end a cycle with Strafe.
[The Rain of Vengeance runes]
Before we start to look into the details, it’s a good idea to know how Rain of Vengeance actually works. Here is a table where I collected some data, and some of these findings are quite interesting. There are some more things to it as explained in the video, but I’ll stick to the most noteworthy ones here. So far, it seems like everything remains the same for 2.3:
Shade is pretty useless (mostly because it creates huge amounts of lags), but might be worth revisiting in 2.3, and Dark Cloud generally is only worth for speedfarm content (because it is very weak, but follows you around as you move). In patch 2.2, Blizzard has removed the “Dark Cloud counter” from the buff bar (but ironically added a "Sentry counter" for the thereafter never to be used M6), but my testing implies that you can still stack it up to 5 times (so in extreme scenarios e.g. with Channeling Pylon you will lose a little dps, but it doesn’t matter that much). In any case, you’ll need to use one of the “controlling” runes (Stampede / Flying Strike) for solo, and Anathema is strictly reserved for groups where you have someone grouping things up for you. Unlike the 5800% weapon damage specified in the tooltip, Anathema actually does 8700% (9570% with Grenadier), which can be seen in the video (Blizzard pls). My guess is that it was designed like that on purpose, and they just assumed that most targets will not get hit by all the grenades (20x290% = 5800%, which would mean a 66.67% hit chance out of the 30 grenades). However, as seen in the video, even small targets like skinny zombies are usually getting hit by more than 20/30 grenades, and on big targets it’s certainly possible to realize its full potential. It’s also worth noting that it has a crazy proc coefficient: 30x0.12=3.60, which is something that might be very interesting for cubed weapon effects in 2.3.
[Preferred stat rolls]
For the CDR rolls, many slots are basically a no-brainer: Both weapons (2x10%), shoulders (8%), diamond in helm (12.5%), gloves (8%). On your offhand, you want to have elite damage and (Calamity) / or (Balefire) area damage, or if that is too difficult to get, RCR. If you use a Calamity, having vitality is also going to help a little. Generally, elite damage / area damage are better than RCR even if you run out of hatred, just because we have very little of both and they're such a huge damage boost. Even trading an ancient offhand with RCR for a nonancient with elite damage is worth it (given equal CDR values). It’s generally advisable to have area damage or RCR on the shoulders, hatred regeneration on the chest helps as well. The rings are dealt with in more detail in a moment. Using Evasive Fire or Strafe damage bonuses on your items is not recommended, as their share of your total dps is negligible.
[Optimal CDR values]
This is where it gets tricky. There has been a lot of talk about which cycle is the best: clean 2:1, delayed 2:1, 3:1, or something else entirely? Well, I have generally been advising for a delayed 2:1 setup using 6 resets, and generally thought that a 2:1 setup is only worth using if you have a pretty GG Flavor of Time. I do have one with dexterity, double crit, CDR and socket, and whenever I tried it I was disappointed. This suggests that the 2:1 cycle might actually not be the best, and only be worth using if you Flavor of Time is far superior to your Hellfire Amulet. In case you don't have either of those, but do possess another amulet with elemental damage instead of mainstat, it's close to a dexterity Hellfire. To assess the proper CDR values, we'll have to take a little look at the numbers:
[The Math – 2:1 vs. 3:1]
So here is the math I've done on Natalya RoV. These results (at the end) speak mostly for solo, but can be applied to groups as well with little modification. The above link compares 4 different setups: Clean 2:1 (full CDR everywhere + Gogok, 5 N2 procs), Delayed 2:1 #1 (no CDR on jewelry + Gogok, 7 N2 procs), Delayed 2:1 #2 (CDR on one ring + Gogok, 6 N2 procs), Clean 3:1 (no CDR on jewelry + Taeguk, 7 N2 procs). Keep in mind that things like better crowd control and grouping from more RoV casts are not factored in. I have compared the CDR values ["CDR" part], then included the APS values with their respective strafe breakpoints ["IAS" part], estimated their cycle time by adding each individual step together, which gave me the RoV casts/Min number ["Cycle Time" part]. This number was further modified with the expected damage gains when moving from a full CDR setup to a lower CDR setup by comparing near best in slot values ["Dmg" part]. Taeguk was included as a flat 40% additive boost (40% is only rank 60, but you also need time for stacking and might lose it once etc.), Caltrops was calculated as a flat 11% multiplicative boost (even though the expected damage while the effect is active gain is approximately 14-18% depending on crit chance / Single Out etc.).
On the right I have gathered some numbers from empirical testing to verify if my cycle times are correct, and they seem to overlap by about 95% with my numbers. To get the data, I have reviewed several of my solo and other DHs' runs with the different setups and created some extra footage for the missing parts. The empirical data is not entirely accurate as I have not done a frame by frame analysis, but instead I was using strong slow motion and collecting time stamps when attacks started and ended. The numbers suggest that with 2.00 APS, you will need to stand still when casting Evasive Fire or RoV for 0,25 sec (50% of attacking time), while the remaining 0,25 sec can be used for movement. While Strafing, this share is reduced to what seems to be 33%, meaning that you can move and strafe again after only 0,167 sec (which is the actual reason why we use Strafe in RoV builds in the first place).
The damage numbers at the end were derived from my old M6/UE spreadsheet, which I modified a little to be used for N6 comparisons. The added benefit of dmg rings vs. CDR rings was calculated by adding perfect ancient average damage to 6 CC / 50 CHD rings, which nets a 7,3% increase in power for the first and another 6,8% increase for the second ring (discounting the inflated damage boost you get from a 672 dps offhand). I also found some more interesting gearing stats, for example nonancient avg dmg is about equal to ancient dex, both of which are about 2-3% worse than CHD / ancient avg dmg. If you have high crit chance from Single Out / Archery / Caltrops, CC obviously loses value but still remains strong, about equal to CHD / avg dmg. Generally it seems to be the best choice to balance all of the stats (excluding dex) on the rings, while putting an emphasis on CC if you don't use any of these skills / passives, and putting an emphasis on CHD / avg dmg otherwise. Iceblink is about 4-6% effective dps behind Taeguk, but adds the benefit that you have better slow uptime and can group things better as well, so the gap is smaller in real scenarios. The Bane of the Stricken is very difficult to theorize, but it's very likely going to be a much bigger damage boost than the Taeguk, even putting a further emphasis on replacing Gogok. I averaged the benefit of the Hellfire Amulet to 9%, which is about what we can expect to get from Ambush, Steady Aim or Single Out.
Comparing the actual resulting damage numbers (RoV casts per minute multiplied by the damage bonus factors of slower cycles) from the different setups (35,150 for Clean 2:1; 36,435 for Delayed 2:1 #1; 37,790 for Delayed 2:1 #2; 43,380 for Clean 3:1) suggests that the latter is in fact the strongest raw output wise, with a considerable margin of 23,4% over the Clean 2:1 and 14,8% over my old solo setup, Delayed 2:1 #2 (N2 6 procs). The 3:1 casts about 25% less RoV than the 2:1, however those are about 63% stronger on average (not necessarily the damage numbers, but also because you have higher crit chance). The damage increase from the additional stats on the rings and the Hellfire passive is quite considerable, but what really seals the deal are the Caltrops, which are used in place of Preparation - Punishment (that all of the other setups require). Obviously losing precious discipline to Caltrops will make survival a lot more difficult, but in good rifts we are often sitting at full discipline with no way of using it. Even when removing the skill entirely the 3:1 setup still remains the strongest, so the discipline can safely be spent for survival when necessary.
These numbers seem very weird at first but the 3:1 is in fact the best choice in the end, especially with the added benefit of Caltrops. Previously, most DHs have been using Preparation - Focused Mind instead of Punishment in its place when running this setup, and usually it felt about even with the 2:1 setups. Even if we can realize only half of the theoretical damage boost, we are still looking at one additional tier we can push higher. All in all, these results seem to confirm that the Flavor of Time setup is in fact the weakest, and only worth using if you have a good elemental damage roll (just like I've been suggesting all along), and that the Delayed 2:1 cycle I'm using is in fact the strongest before considering the new / old 3:1. The big advantage of FoT (with elemental damage) is merely that it is much more easily acquired than a Hellfire (yet still very difficult to get). What's important to note is that the 3:1 is the most unforgiving in CDR requirements. If you don't want to sacrifice a ring for it, you will need to have perfect rolls on every slot, and even then it's very close (2,20 sec RoV CD vs 2,263 cycle time, including 0,24sec of human error). This is something that I have noticed quite clearly when playing it in the past - when you manage to execute the cycle optimally, you will have a short delay with 46% CDR. Depending on your item rolls, it might be advisable to use one CDR ring and the best item rolls (in my case, I have 7% ancient Pauldrons of the Skeleton King and 6% ancient Natalya's Gloves). If you get 3-4% damage out of that it's definitely worth to go with one CDR ring, as it makes the cycle much smoother as well (and with enough practice you will regain most of the lost damage by cutting the cycle time closer to 2.05-2.15 sec), but I'd recommend to do that anyway. This takes away a little bit from the damage advantage of the 3:1 again, but eventually you will feel the difference when running only 46 CDR, and you get other perks like increased wolf uptime as well.
[Patch 2.2 solo: Stampede (Fire) vs. Flying Strike (Cold)]
There has also been a lot of debate about which element to use, and I’ll try to summarize the points here quickly. For 2.3, Flying Strike will be useless anyway because of the CC nerfs, Stampede still work almost as usual (sometimes for the better, sometime for the worse). For 2.2, there are two main differences:
1) Consistency: Flying Strike can deal with much more aggressive and dangerous rifts. You can permanently freeze any white monster and almost any blue elite, and can easily skip yellows.
2) Damage: Flying Strike naturally deals less damage because we can’t use Balefire Caster, and apparently it also has a hard time against Rift Guardians, which take significantly longer. However, its huge advantage is that once you have a good crowd in one spot, you can just kill them off very quickly without knocking everything all over the screen, splitting them up and losing all your AoE damage.
Personally I enjoy the cold playstyle a lot more. Even with the Caltrops setup, you can deal with difficult rifts quite well, while you basically get screwed with fire whenever you have to skip and can’t stabilize. However, the RG damage is a huge problem – basically you will lose your whole rift clearing speed advantage here, and the only solution is to fish for Stonesinger. You can see this quite clearly in the 65/66 runs I have done using cold, where I was able to get through the rift quite nicely, but took an eternity to kill Perendi, but still managed to kill Stonesinger very quickly with a Power Pylon. Basically, if you look at my successful tier 64 run, or the 66 run from before, those were dream rifts for cold (with the 66 being impossible to beat for fire, because you just knock the teddies away too much). Just in general, you will get more out of almost any rift with cold, especially those that are filled with a lot of small monsters, but will lose out on the RG. In order to maximize the cold build, it is advisable to use an Ess of Johan for group things up (the best rolls are 20% cold, 100 CHD, dexterity), because you really want to pull and keep everything in one spot.
[Groups]
For groups, the prime choice is Anathema + Grenadier, you don’t want to knockback anything and don’t want to use the weak cold / physical / lightning runes. Obviously for groups, the benefit of Taeguk and Caltrops will decrease a little, as you will have much more additive damage and crit chance from party buffs. Not only that, but you will generally run at the 2.5 breakpoint with 1-2 buffs, or in some cases even at the 3.0 breakpoint (with Gogok, Law, Voodoo and IAS Slow Time simultaneously). Judging from my experience, completing these cycles at these breakpoints without any double Strafe ticks at all is very difficult to do consistently. Since you risk dying in groups much more than in solo, and sometimes you have to stop dealing damage, Taeguk really falls behind and I suggest to use Bane of the Powerful or Bane of the Stricken instead. The elite damage really helps, as you generally don't have any other source of that (unless you got lucky with your offhand). Because all of this, running a 2:1 cycle with double ticks on Strafe is probably a good idea for those who can't master the 3:1 yet, which is still going to be stronger eventually. The thing is that you don't really want to use Preparation – Punishment in groups (since Marked for Death is a better choice), and you will run out of hatred because of that. If everything is permanently controlled or relatively safe, that's not a big deal – just fill up your hatred with Evasive Fire for a full cycle and continue like normal after that. To alleviate this problem, you can use Night Stalker when running at high breakpoints. With the 3:1 cycle, you won't have many hatred issues (yet still run low occasionally) even without the passive.
[Patch 2.3]
As for Season 4, we'll see what becomes of the Natalya set. With the cube, Bane of the Stricken and the removal of perma CC many things are going to change. Build-wise, we have a lot of different options to try out: You can go a heavy RCR setup with Yang's Recurve and include e.g. Fan of Knives, use the Demon's Demise to proc some more "passive" AoE damage, cube some cool melee weapons for insane procs, or run setups without Strafe at all by exploiting the benefits of Hunter's Wrath / Yang's Recurve / Helltrapper. We might even see some 1:1 setups, and maybe they turn out to be incredibly strong, however at this point I don’t think that they’re really going to be worth it. There are many different things to explore, and I expect to see some strong new builds emerge over the course of the coming weeks. I’ve tried to give some hints at what to it's going to look like most likely, and so far it seems that Natalya and Unhallowed Essence remain about equally strong for most activities, unless something completely unforeseen happens that will push one or the other set a lot higher (Natalya has a lot more options here, and benefits a lot more from the buffed Conduit Pylons as well). Obviously, for speedfarm content, Natalya is getting slightly more buffs (with In-Geom being much stronger for this set), while Unhallowed Essence is getting slightly more buffs for slower runs (with Convention of Elements being much stronger for long fights).
For Kanai’s Cube, we have a lot of options: Calamity, Helltrapper, Demon's Demise, In-Geom and some of the melee weapons are definitely worth looking at for the weapon slot, Hexing Pants are probably going to be the go-to choice for armor (and they might enable 2:1 setups to run without Preparation) or we will see some use of the Leoric's Crown, and we have Obsidian Ring / RoRG / Convention of Elements for the jewelry depending on your build.
We’ll see how things play out, currently I think that both sets are going to be used for both group and solo pushing, and will yield close to the same results depending on rift / RNG / party composition.
[TL;DR]
Anathema is bugged and does way more damage than it says, and also has a crazy proc coefficient. For 2.2, math and testing shows that a 3:1 setup surpasses any 2:1 (clean or delayed) in both solo and groups. You're aiming for 47-49% CDR, 46% is not enough because you will have some downtime. For 2.3, many things are out in the open; we might see some pretty cool builds with cubed melee weapons, the new Demon's Demise and potentially some that don't even include Strafe anymore (e.g. using Hunter’s Wrath, Yang’s Recurve and / or Fan of Knives). It's too early to tell exactly but we'll definitely see a lot of innovation with the Natalya set (most things were just impossible to test in full detail on the PTR because of the latency, missing items and lack of proper groups). All in all DH is in a good spot for speedfarm content, and in a very uncomfortable spot for high end content. DHs have a much harder time to survive and deal equal or less damage than other classes currently, so we’ll have to fight our way up there. In any case, I'm fairly confident that DH will still be an appreciated choice in groups, and that patch 2.3 will be quite enjoyable for us as well.
I hope you enjoyed the read and / or my video, and maybe got inspired to min/max the Natalya for yourself or try some of the new things in patch 2.3.
Cheers,
wudijo
[FAQ]
Q: What does the Natalya set do exactly?
A: (2) Set: Reduce the cooldown of Rain of Vengeance by 2 seconds when you hit with a Hatred-generating attack or Hatred-spending attack.
(4) Set: Rain of Vengeance deals 100% increased damage. [Multiplicative, counts for Crashing Rain as well]
(6) Set: After casting Rain of Vengeance, deal 400% increased damage and take 30% reduced damage for 5 seconds. [Multiplicative]
Q: Isn't it a good idea to try Ball Lightning with it, because it can hit multiple times and reduce the cooldown by a lot?
A: No, only the first hit of every attack counts. The exception are channeled abilities like Strafe, which repeat an animation cycle over and over.
Q: What does the 6-piece bonus boost exactly?
A: It boosts all of your damage, including gems, procs and the conduit pylon.
Q: What about items that change how skills spend resources (like Krider)?
A: Kridershot / Spines of Seething Hatred work as usual, Danetta's makes Vault work too.
Q: Why Strafe?
A: Strafe + any other spell makes for many more N2 procs, thus reducing the cooldown much quicker than it would if you just used a single ability.
Q: What are the important Strafe breakpoints?
A: 1.67 / 1.88 / 2.15 / 2.50 / 3.00 APS.
Q: If I have hatred issues, should I reduce attack speed in paragons?
A: No, that doesn't make much sense. With more attack speed, you can use your Evasive Fire more quickly, and regenerate more hatred that way. Obviously this will speed up the cycle a little bit, so the passive hatred regen won't have that much of an effect anymore, but you can also fill up quicker should you run out of hatred.
Q: Why should I roll off the damage range on my offhand?
A: RoV and CR take only your mainhand weapon damage into account. Since these two attacks make up for almost all of your damage (90-99% depending on number of targets), it's a much better idea to go for one more primary stat on your weapon.
Q: Which legendary gems should I use?
A: Bane of the Trapped + Zei's Stone of Vengeance + (in order of my priority) Bane of the Stricken / Taeguk / Iceblink / Gogok. For groups that heavily rely on no sources of crowd control at all (e.g. because you have a pulling support Barb), dropping the Zei's might be necessary.
Q: Which passives should I use?
A: Awareness, Cull the Weak, Ambush, and then one or two out of Steady Aim, Archery, Single Out, Night Stalker.
Q: How much RCR or hatred regen do I need?
A: This completely depends on your cycle and how well you can execute the playstyle. Generally, the quicker cycles (2:1) require you to have some sort of active hatred regen (e.g. Preparation – Punishment), and you can do the slower ones (3:1) without if you have a little bit of passive regen (e.g. Archery, RCR on shoulders, hatred on chest). Usually it's best to figure out what cooldown setup you want to run, and solve the resource issues after.
Q: Why shouldn't I choose the 20% buff from Calamity over a Balefire when running a fire rune?
A: Calamity stacks with most of your other damage amplifiers like Wolf, Steady Aim, Taeguk, Strongarms, etc. Balefire only stacks with your fire elemental damage. This greatly increases the relative impact of having more fire damage, making Balefire stronger.
Q: Do Calamity and Marked for Death stack?
A: Yes, you can have one of each effect up on a monster at the same time.
Q: I have a Calamity with elite damage and a Balefire without, what should I choose?
A: For solo, Balefire will prove to be better anyway, and maybe you get another decent stat from it (say RCR or AD). In 2.3 with Bane of the Stricken, the duration of elite fights will play a much smaller role anyway.
Q: Area damage seems really good, should I stack it?
A: Yes, now that the stat finally got fixed again, it's definitely worth going for it. The most obvious slot is on the shoulders, the next on the offhand (after CDR in both cases). If you have those, you are in a pretty good spot already, going for more is probably not worth it unless you trade some of the "bad" damage stats for it (e.g. nonancient dexterity on a ring).
Q: Is it worth using an amulet with CDR that is not a Flavor of Time?
A: Not really. As we have seen in the math section, the 2:1 cycles are behind the 3:1 anyway, and if you really need CDR on the jewelry, you'll want to take it on the rings. Amulets roll much higher values of dexterity, crit chance and crit damage, but exactly the same CDR.
Q: Why do you use [insert Pauldrons of the Skeleton King / Pauldrons of Zakara / Homing Pads / Corruption here]?
A: Generally, you should just use the shoulders with the best stats, their special effects are almost worthless, with the best one being the cheat death proc of the Pauldrons of the Skeleton King. The problem with these is that they always roll armor + vitality, so you can only have 8 CDR and 15 RoV damage and no RCR / AD in addition to that.
Q: Which follower should I use?
A: If you have hatred issues, Templar is the way to go. Other than that, Enchantress does a good job at proccing the slow with Thunderfury and generally attacks more often than the Templar when you need to outrange stuff.
Q: How do I proc the slow for Bane of the Trapped and Cull the Weak in solo?
A: You use your follower. On the PTR, many have been cubing the Thunderfury themselves as well, but I think that's definitely not the best option we have, and the Enchantress usually does pretty well in most rifts.
Diablofans Link
Reddit Link
Video:
Greetings all,
you might remember my post about UE + M6 optimization in the beginning of S3, and I thought it’s about time that I do some stuff like that for Natalya as well. Now that the end of season 3 draws close, I guess many are trying their best to get on the leaderboards, and I decided to put some more time into it. N6 has been used basically the entire season for solo and speedruns, and later on also got included in group pushing as well, effectively competing with UE for the top DH set in a tight race. This post mainly focuses on greater rift pushing, which generally requires the most optimization (and low level content farming efficiency has more parameters than just pure dps). It contains points for the current situation (patch 2.2), but I’ve added a lot of things to look out for in the next season (patch 2.3) as well, and included a TL;DR at the end. I was working on this as a side-project for quite a while but never quite managed to finish it, so here it is.
The structure of the post is as follows:
- Natalya For Dummies
- The Rain of Vengeance runes
- Preferred stat rolls
- Optimal CDR values
- The Math – 2:1 vs. 3:1
- Patch 2.2 solo: Stampede (fire) vs. Flying Strike (cold)
- Groups
- Patch 2.3
- TL;DR
- FAQ
Enjoy.
[Natalya For Dummies]
The Natalya set heavily relies on Rain of Vengeance (RoV) to deal damage, and generally you aim at reducing the cooldown as much as possible using the 2-piece bonus. To achieve this, heavy CDR stacking is required, and a combination of Strafe + generator (mostly Evasive Fire (EF) for increased hatred regen) has proven to be the best, because like that you can get more attacks per time unit than you normally would when only attack with a regular skill. The RoV, combined with the Crashing Rain (CR) belt (especially in solo), is your main source of damage. The generator casts and Strafe are basically irrelevant, and because of that you don’t require a weapon damage roll on your offhand weapon (RoV and CR scale only off your mainhand weapon damage). In your offhand, you generally take a Balefire Caster when you are the only dps (solo / 2man / maybe 3man) and a Calamity when you are not (maybe 3man / 4man, or whenever you are using a non-fire rune). The reason why a Balefire Caster is better is because elemental damage is a separate damage multiplier and you generally only have 20% from bracers (or 40% if you got lucky with your amulet), while you have lots and lots of additive damage (30% Strongarm, 55% RoV, 20% Steady Aim, 30% Wolf, in groups even more) which is the same category as Calamity.
Completing a combo of Strafe + EF + RoV is called a cycle, and there are mainly two variations: a 2:1 cycle and a 3:1 cycle. For a 2:1 cycle, that means you cast 2x EF and 1x RoV while Strafing permanently, then repeat (3x EF for 3:1). That playstyle is a little bit difficult to get into, but once you get into the rhythm things become a lot easier and you can actually focus on other things much more during the rift. A reset is what the N2 procs are called, and you generally try to chain them in a way that you start and end a cycle with Strafe.
[The Rain of Vengeance runes]
Before we start to look into the details, it’s a good idea to know how Rain of Vengeance actually works. Here is a table where I collected some data, and some of these findings are quite interesting. There are some more things to it as explained in the video, but I’ll stick to the most noteworthy ones here. So far, it seems like everything remains the same for 2.3:
Shade is pretty useless (mostly because it creates huge amounts of lags), but might be worth revisiting in 2.3, and Dark Cloud generally is only worth for speedfarm content (because it is very weak, but follows you around as you move). In patch 2.2, Blizzard has removed the “Dark Cloud counter” from the buff bar (but ironically added a "Sentry counter" for the thereafter never to be used M6), but my testing implies that you can still stack it up to 5 times (so in extreme scenarios e.g. with Channeling Pylon you will lose a little dps, but it doesn’t matter that much). In any case, you’ll need to use one of the “controlling” runes (Stampede / Flying Strike) for solo, and Anathema is strictly reserved for groups where you have someone grouping things up for you. Unlike the 5800% weapon damage specified in the tooltip, Anathema actually does 8700% (9570% with Grenadier), which can be seen in the video (Blizzard pls). My guess is that it was designed like that on purpose, and they just assumed that most targets will not get hit by all the grenades (20x290% = 5800%, which would mean a 66.67% hit chance out of the 30 grenades). However, as seen in the video, even small targets like skinny zombies are usually getting hit by more than 20/30 grenades, and on big targets it’s certainly possible to realize its full potential. It’s also worth noting that it has a crazy proc coefficient: 30x0.12=3.60, which is something that might be very interesting for cubed weapon effects in 2.3.
[Preferred stat rolls]
For the CDR rolls, many slots are basically a no-brainer: Both weapons (2x10%), shoulders (8%), diamond in helm (12.5%), gloves (8%). On your offhand, you want to have elite damage and (Calamity) / or (Balefire) area damage, or if that is too difficult to get, RCR. If you use a Calamity, having vitality is also going to help a little. Generally, elite damage / area damage are better than RCR even if you run out of hatred, just because we have very little of both and they're such a huge damage boost. Even trading an ancient offhand with RCR for a nonancient with elite damage is worth it (given equal CDR values). It’s generally advisable to have area damage or RCR on the shoulders, hatred regeneration on the chest helps as well. The rings are dealt with in more detail in a moment. Using Evasive Fire or Strafe damage bonuses on your items is not recommended, as their share of your total dps is negligible.
[Optimal CDR values]
This is where it gets tricky. There has been a lot of talk about which cycle is the best: clean 2:1, delayed 2:1, 3:1, or something else entirely? Well, I have generally been advising for a delayed 2:1 setup using 6 resets, and generally thought that a 2:1 setup is only worth using if you have a pretty GG Flavor of Time. I do have one with dexterity, double crit, CDR and socket, and whenever I tried it I was disappointed. This suggests that the 2:1 cycle might actually not be the best, and only be worth using if you Flavor of Time is far superior to your Hellfire Amulet. In case you don't have either of those, but do possess another amulet with elemental damage instead of mainstat, it's close to a dexterity Hellfire. To assess the proper CDR values, we'll have to take a little look at the numbers:
[The Math – 2:1 vs. 3:1]
So here is the math I've done on Natalya RoV. These results (at the end) speak mostly for solo, but can be applied to groups as well with little modification. The above link compares 4 different setups: Clean 2:1 (full CDR everywhere + Gogok, 5 N2 procs), Delayed 2:1 #1 (no CDR on jewelry + Gogok, 7 N2 procs), Delayed 2:1 #2 (CDR on one ring + Gogok, 6 N2 procs), Clean 3:1 (no CDR on jewelry + Taeguk, 7 N2 procs). Keep in mind that things like better crowd control and grouping from more RoV casts are not factored in. I have compared the CDR values ["CDR" part], then included the APS values with their respective strafe breakpoints ["IAS" part], estimated their cycle time by adding each individual step together, which gave me the RoV casts/Min number ["Cycle Time" part]. This number was further modified with the expected damage gains when moving from a full CDR setup to a lower CDR setup by comparing near best in slot values ["Dmg" part]. Taeguk was included as a flat 40% additive boost (40% is only rank 60, but you also need time for stacking and might lose it once etc.), Caltrops was calculated as a flat 11% multiplicative boost (even though the expected damage while the effect is active gain is approximately 14-18% depending on crit chance / Single Out etc.).
On the right I have gathered some numbers from empirical testing to verify if my cycle times are correct, and they seem to overlap by about 95% with my numbers. To get the data, I have reviewed several of my solo and other DHs' runs with the different setups and created some extra footage for the missing parts. The empirical data is not entirely accurate as I have not done a frame by frame analysis, but instead I was using strong slow motion and collecting time stamps when attacks started and ended. The numbers suggest that with 2.00 APS, you will need to stand still when casting Evasive Fire or RoV for 0,25 sec (50% of attacking time), while the remaining 0,25 sec can be used for movement. While Strafing, this share is reduced to what seems to be 33%, meaning that you can move and strafe again after only 0,167 sec (which is the actual reason why we use Strafe in RoV builds in the first place).
The damage numbers at the end were derived from my old M6/UE spreadsheet, which I modified a little to be used for N6 comparisons. The added benefit of dmg rings vs. CDR rings was calculated by adding perfect ancient average damage to 6 CC / 50 CHD rings, which nets a 7,3% increase in power for the first and another 6,8% increase for the second ring (discounting the inflated damage boost you get from a 672 dps offhand). I also found some more interesting gearing stats, for example nonancient avg dmg is about equal to ancient dex, both of which are about 2-3% worse than CHD / ancient avg dmg. If you have high crit chance from Single Out / Archery / Caltrops, CC obviously loses value but still remains strong, about equal to CHD / avg dmg. Generally it seems to be the best choice to balance all of the stats (excluding dex) on the rings, while putting an emphasis on CC if you don't use any of these skills / passives, and putting an emphasis on CHD / avg dmg otherwise. Iceblink is about 4-6% effective dps behind Taeguk, but adds the benefit that you have better slow uptime and can group things better as well, so the gap is smaller in real scenarios. The Bane of the Stricken is very difficult to theorize, but it's very likely going to be a much bigger damage boost than the Taeguk, even putting a further emphasis on replacing Gogok. I averaged the benefit of the Hellfire Amulet to 9%, which is about what we can expect to get from Ambush, Steady Aim or Single Out.
Comparing the actual resulting damage numbers (RoV casts per minute multiplied by the damage bonus factors of slower cycles) from the different setups (35,150 for Clean 2:1; 36,435 for Delayed 2:1 #1; 37,790 for Delayed 2:1 #2; 43,380 for Clean 3:1) suggests that the latter is in fact the strongest raw output wise, with a considerable margin of 23,4% over the Clean 2:1 and 14,8% over my old solo setup, Delayed 2:1 #2 (N2 6 procs). The 3:1 casts about 25% less RoV than the 2:1, however those are about 63% stronger on average (not necessarily the damage numbers, but also because you have higher crit chance). The damage increase from the additional stats on the rings and the Hellfire passive is quite considerable, but what really seals the deal are the Caltrops, which are used in place of Preparation - Punishment (that all of the other setups require). Obviously losing precious discipline to Caltrops will make survival a lot more difficult, but in good rifts we are often sitting at full discipline with no way of using it. Even when removing the skill entirely the 3:1 setup still remains the strongest, so the discipline can safely be spent for survival when necessary.
These numbers seem very weird at first but the 3:1 is in fact the best choice in the end, especially with the added benefit of Caltrops. Previously, most DHs have been using Preparation - Focused Mind instead of Punishment in its place when running this setup, and usually it felt about even with the 2:1 setups. Even if we can realize only half of the theoretical damage boost, we are still looking at one additional tier we can push higher. All in all, these results seem to confirm that the Flavor of Time setup is in fact the weakest, and only worth using if you have a good elemental damage roll (just like I've been suggesting all along), and that the Delayed 2:1 cycle I'm using is in fact the strongest before considering the new / old 3:1. The big advantage of FoT (with elemental damage) is merely that it is much more easily acquired than a Hellfire (yet still very difficult to get). What's important to note is that the 3:1 is the most unforgiving in CDR requirements. If you don't want to sacrifice a ring for it, you will need to have perfect rolls on every slot, and even then it's very close (2,20 sec RoV CD vs 2,263 cycle time, including 0,24sec of human error). This is something that I have noticed quite clearly when playing it in the past - when you manage to execute the cycle optimally, you will have a short delay with 46% CDR. Depending on your item rolls, it might be advisable to use one CDR ring and the best item rolls (in my case, I have 7% ancient Pauldrons of the Skeleton King and 6% ancient Natalya's Gloves). If you get 3-4% damage out of that it's definitely worth to go with one CDR ring, as it makes the cycle much smoother as well (and with enough practice you will regain most of the lost damage by cutting the cycle time closer to 2.05-2.15 sec), but I'd recommend to do that anyway. This takes away a little bit from the damage advantage of the 3:1 again, but eventually you will feel the difference when running only 46 CDR, and you get other perks like increased wolf uptime as well.
[Patch 2.2 solo: Stampede (Fire) vs. Flying Strike (Cold)]
There has also been a lot of debate about which element to use, and I’ll try to summarize the points here quickly. For 2.3, Flying Strike will be useless anyway because of the CC nerfs, Stampede still work almost as usual (sometimes for the better, sometime for the worse). For 2.2, there are two main differences:
1) Consistency: Flying Strike can deal with much more aggressive and dangerous rifts. You can permanently freeze any white monster and almost any blue elite, and can easily skip yellows.
2) Damage: Flying Strike naturally deals less damage because we can’t use Balefire Caster, and apparently it also has a hard time against Rift Guardians, which take significantly longer. However, its huge advantage is that once you have a good crowd in one spot, you can just kill them off very quickly without knocking everything all over the screen, splitting them up and losing all your AoE damage.
Personally I enjoy the cold playstyle a lot more. Even with the Caltrops setup, you can deal with difficult rifts quite well, while you basically get screwed with fire whenever you have to skip and can’t stabilize. However, the RG damage is a huge problem – basically you will lose your whole rift clearing speed advantage here, and the only solution is to fish for Stonesinger. You can see this quite clearly in the 65/66 runs I have done using cold, where I was able to get through the rift quite nicely, but took an eternity to kill Perendi, but still managed to kill Stonesinger very quickly with a Power Pylon. Basically, if you look at my successful tier 64 run, or the 66 run from before, those were dream rifts for cold (with the 66 being impossible to beat for fire, because you just knock the teddies away too much). Just in general, you will get more out of almost any rift with cold, especially those that are filled with a lot of small monsters, but will lose out on the RG. In order to maximize the cold build, it is advisable to use an Ess of Johan for group things up (the best rolls are 20% cold, 100 CHD, dexterity), because you really want to pull and keep everything in one spot.
[Groups]
For groups, the prime choice is Anathema + Grenadier, you don’t want to knockback anything and don’t want to use the weak cold / physical / lightning runes. Obviously for groups, the benefit of Taeguk and Caltrops will decrease a little, as you will have much more additive damage and crit chance from party buffs. Not only that, but you will generally run at the 2.5 breakpoint with 1-2 buffs, or in some cases even at the 3.0 breakpoint (with Gogok, Law, Voodoo and IAS Slow Time simultaneously). Judging from my experience, completing these cycles at these breakpoints without any double Strafe ticks at all is very difficult to do consistently. Since you risk dying in groups much more than in solo, and sometimes you have to stop dealing damage, Taeguk really falls behind and I suggest to use Bane of the Powerful or Bane of the Stricken instead. The elite damage really helps, as you generally don't have any other source of that (unless you got lucky with your offhand). Because all of this, running a 2:1 cycle with double ticks on Strafe is probably a good idea for those who can't master the 3:1 yet, which is still going to be stronger eventually. The thing is that you don't really want to use Preparation – Punishment in groups (since Marked for Death is a better choice), and you will run out of hatred because of that. If everything is permanently controlled or relatively safe, that's not a big deal – just fill up your hatred with Evasive Fire for a full cycle and continue like normal after that. To alleviate this problem, you can use Night Stalker when running at high breakpoints. With the 3:1 cycle, you won't have many hatred issues (yet still run low occasionally) even without the passive.
[Patch 2.3]
As for Season 4, we'll see what becomes of the Natalya set. With the cube, Bane of the Stricken and the removal of perma CC many things are going to change. Build-wise, we have a lot of different options to try out: You can go a heavy RCR setup with Yang's Recurve and include e.g. Fan of Knives, use the Demon's Demise to proc some more "passive" AoE damage, cube some cool melee weapons for insane procs, or run setups without Strafe at all by exploiting the benefits of Hunter's Wrath / Yang's Recurve / Helltrapper. We might even see some 1:1 setups, and maybe they turn out to be incredibly strong, however at this point I don’t think that they’re really going to be worth it. There are many different things to explore, and I expect to see some strong new builds emerge over the course of the coming weeks. I’ve tried to give some hints at what to it's going to look like most likely, and so far it seems that Natalya and Unhallowed Essence remain about equally strong for most activities, unless something completely unforeseen happens that will push one or the other set a lot higher (Natalya has a lot more options here, and benefits a lot more from the buffed Conduit Pylons as well). Obviously, for speedfarm content, Natalya is getting slightly more buffs (with In-Geom being much stronger for this set), while Unhallowed Essence is getting slightly more buffs for slower runs (with Convention of Elements being much stronger for long fights).
For Kanai’s Cube, we have a lot of options: Calamity, Helltrapper, Demon's Demise, In-Geom and some of the melee weapons are definitely worth looking at for the weapon slot, Hexing Pants are probably going to be the go-to choice for armor (and they might enable 2:1 setups to run without Preparation) or we will see some use of the Leoric's Crown, and we have Obsidian Ring / RoRG / Convention of Elements for the jewelry depending on your build.
We’ll see how things play out, currently I think that both sets are going to be used for both group and solo pushing, and will yield close to the same results depending on rift / RNG / party composition.
[TL;DR]
Anathema is bugged and does way more damage than it says, and also has a crazy proc coefficient. For 2.2, math and testing shows that a 3:1 setup surpasses any 2:1 (clean or delayed) in both solo and groups. You're aiming for 47-49% CDR, 46% is not enough because you will have some downtime. For 2.3, many things are out in the open; we might see some pretty cool builds with cubed melee weapons, the new Demon's Demise and potentially some that don't even include Strafe anymore (e.g. using Hunter’s Wrath, Yang’s Recurve and / or Fan of Knives). It's too early to tell exactly but we'll definitely see a lot of innovation with the Natalya set (most things were just impossible to test in full detail on the PTR because of the latency, missing items and lack of proper groups). All in all DH is in a good spot for speedfarm content, and in a very uncomfortable spot for high end content. DHs have a much harder time to survive and deal equal or less damage than other classes currently, so we’ll have to fight our way up there. In any case, I'm fairly confident that DH will still be an appreciated choice in groups, and that patch 2.3 will be quite enjoyable for us as well.
I hope you enjoyed the read and / or my video, and maybe got inspired to min/max the Natalya for yourself or try some of the new things in patch 2.3.
Cheers,
wudijo
[FAQ]
Q: What does the Natalya set do exactly?
A: (2) Set: Reduce the cooldown of Rain of Vengeance by 2 seconds when you hit with a Hatred-generating attack or Hatred-spending attack.
(4) Set: Rain of Vengeance deals 100% increased damage. [Multiplicative, counts for Crashing Rain as well]
(6) Set: After casting Rain of Vengeance, deal 400% increased damage and take 30% reduced damage for 5 seconds. [Multiplicative]
Q: Isn't it a good idea to try Ball Lightning with it, because it can hit multiple times and reduce the cooldown by a lot?
A: No, only the first hit of every attack counts. The exception are channeled abilities like Strafe, which repeat an animation cycle over and over.
Q: What does the 6-piece bonus boost exactly?
A: It boosts all of your damage, including gems, procs and the conduit pylon.
Q: What about items that change how skills spend resources (like Krider)?
A: Kridershot / Spines of Seething Hatred work as usual, Danetta's makes Vault work too.
Q: Why Strafe?
A: Strafe + any other spell makes for many more N2 procs, thus reducing the cooldown much quicker than it would if you just used a single ability.
Q: What are the important Strafe breakpoints?
A: 1.67 / 1.88 / 2.15 / 2.50 / 3.00 APS.
Q: If I have hatred issues, should I reduce attack speed in paragons?
A: No, that doesn't make much sense. With more attack speed, you can use your Evasive Fire more quickly, and regenerate more hatred that way. Obviously this will speed up the cycle a little bit, so the passive hatred regen won't have that much of an effect anymore, but you can also fill up quicker should you run out of hatred.
Q: Why should I roll off the damage range on my offhand?
A: RoV and CR take only your mainhand weapon damage into account. Since these two attacks make up for almost all of your damage (90-99% depending on number of targets), it's a much better idea to go for one more primary stat on your weapon.
Q: Which legendary gems should I use?
A: Bane of the Trapped + Zei's Stone of Vengeance + (in order of my priority) Bane of the Stricken / Taeguk / Iceblink / Gogok. For groups that heavily rely on no sources of crowd control at all (e.g. because you have a pulling support Barb), dropping the Zei's might be necessary.
Q: Which passives should I use?
A: Awareness, Cull the Weak, Ambush, and then one or two out of Steady Aim, Archery, Single Out, Night Stalker.
Q: How much RCR or hatred regen do I need?
A: This completely depends on your cycle and how well you can execute the playstyle. Generally, the quicker cycles (2:1) require you to have some sort of active hatred regen (e.g. Preparation – Punishment), and you can do the slower ones (3:1) without if you have a little bit of passive regen (e.g. Archery, RCR on shoulders, hatred on chest). Usually it's best to figure out what cooldown setup you want to run, and solve the resource issues after.
Q: Why shouldn't I choose the 20% buff from Calamity over a Balefire when running a fire rune?
A: Calamity stacks with most of your other damage amplifiers like Wolf, Steady Aim, Taeguk, Strongarms, etc. Balefire only stacks with your fire elemental damage. This greatly increases the relative impact of having more fire damage, making Balefire stronger.
Q: Do Calamity and Marked for Death stack?
A: Yes, you can have one of each effect up on a monster at the same time.
Q: I have a Calamity with elite damage and a Balefire without, what should I choose?
A: For solo, Balefire will prove to be better anyway, and maybe you get another decent stat from it (say RCR or AD). In 2.3 with Bane of the Stricken, the duration of elite fights will play a much smaller role anyway.
Q: Area damage seems really good, should I stack it?
A: Yes, now that the stat finally got fixed again, it's definitely worth going for it. The most obvious slot is on the shoulders, the next on the offhand (after CDR in both cases). If you have those, you are in a pretty good spot already, going for more is probably not worth it unless you trade some of the "bad" damage stats for it (e.g. nonancient dexterity on a ring).
Q: Is it worth using an amulet with CDR that is not a Flavor of Time?
A: Not really. As we have seen in the math section, the 2:1 cycles are behind the 3:1 anyway, and if you really need CDR on the jewelry, you'll want to take it on the rings. Amulets roll much higher values of dexterity, crit chance and crit damage, but exactly the same CDR.
Q: Why do you use [insert Pauldrons of the Skeleton King / Pauldrons of Zakara / Homing Pads / Corruption here]?
A: Generally, you should just use the shoulders with the best stats, their special effects are almost worthless, with the best one being the cheat death proc of the Pauldrons of the Skeleton King. The problem with these is that they always roll armor + vitality, so you can only have 8 CDR and 15 RoV damage and no RCR / AD in addition to that.
Q: Which follower should I use?
A: If you have hatred issues, Templar is the way to go. Other than that, Enchantress does a good job at proccing the slow with Thunderfury and generally attacks more often than the Templar when you need to outrange stuff.
Q: How do I proc the slow for Bane of the Trapped and Cull the Weak in solo?
A: You use your follower. On the PTR, many have been cubing the Thunderfury themselves as well, but I think that's definitely not the best option we have, and the Enchantress usually does pretty well in most rifts.