Hey nobody needs to queue up to speak here - there are always multiple angles of looking at things!
Let’s start with Flame Torrent: it’s gotten a lot of attention lately, but IMO it’s massively overhyped and overvalued. Flame Torrent is great while leveling and for lower-DPS skeleton builds, but it significantly underperforms at endgame when you’re looking at high DPS pet builds (such as the 1M+ variety).
First, some guideposts. I had posted this earlier in another thread about benchmarking pet DPS:
(For consistency with the videos, we’ll use 1.0.2.1 builds and numbers.)
Let’s assume a consistent 4 concurrent procs of Flame Torrent for an average of 100k DPS. (Completely ballpark estimate, but this doesn’t matter very much for our purposes.) This means that, for the sake of argument, Skelemancer has 570k DPS with 100k additional coming from Flame Torrent, i.e. Flame Torrent represents an extra 17.5% DPS for that build. However, Flame Torrent would only add an additional 7.5% DPS to the Night King. That’s actually not very much: for example, the Night King’s DPS could be improved by 7.5 percent with just +16% more attack speed from items/devos.
Now look at what happens when you go in the opposite direction: for a build with only 200k DPS, another 100K DPS from Flame Torrent is going to represent a huge +50% increase to DPS, which means of course they’re going to sing mad praises about how amazing Flame Torrent is.
IMO Flame Torrent is a DPS crutch to help builds with weak DPS: how fervently someone advocates for Flame Torrent becomes a reflection of how badly their build relies on it for DPS. On the other hand, once you’ve reached 1M+ DPS by building up a serious reservoir of all the pet DPS components (flat damage, +% damage, RR, OA/crit, speed, and pet count), then a comparatively small percentage increase to any of those factors thereafter will further boost DPS by more than Flame Torrent ever could. It simply becomes more efficient to just pile on any of those 6 DPS factors even higher since they’re all multiplicative factors. Remember: multiplicative stacking is incredibly powerful, and pet builds excel at it better than any other build type.
Flame Torrent also suffers from some serious mechanical disadvantages:[ul]
[li]DPS loss. In order effectively make use of Flame Torrent, most players will walk all pets (including non-skeletons) directly next to their intended target so that skeletons alone can use their procs. However, this costs DPS due to (a) interruption of autoattack from all pets (at a cost of 1M+/s), and (b) exposure of all pets (including ranged pets) to additional damage and unnecessary casualties, especially from AOEs.[/ul][ul][/li]
[li]Limited coverage. Flame Torrent’s limited range means that outside of the “skeleton walking” tactic for burst damage, it actually affects relatively few targets compared to other devotions with better-coverage AOEs. In particular, Flame Torrent is largely wasted on ranged skeletons.[/ul][ul][/li]
[li]Limited RR. Most experienced players would consider about 60-100 RR to be a good benchmark for primary damage type(s) at endgame, but skeleton builds using Flame Torrent almost never achieve anywhere close to this much RR for fire and/or chaos. This means that while skeletal Flame Torrent looks good at early difficulties and against low-resistance targets like the practice dummy and Mad Queen, it significantly underperforms against many target types (such as Chthonians) during actual gameplay in Ultimate/Gladiator.[/ul][/li]None of this means that Flame Torrent is bad - it’s simply deeply flawed, and it’s better early on in a build’s life than for optimal max-DPS at endgame.
Moving on, let’s consider Rend as one of the alternatives to Flame Torrent. Rend has its own issues: as a debuff with an attached damage component, multiple Rend procs on the same target won’t stack damage, and it also won’t benefit from crit DOT mechanics for continuous crits. There will be no abusive single-target damage here. But even without multi-proc stacking or perma-crits, Rend provides some useful benefits:[ul]
[li]+5% pet OA (Huntress node). This increases Rend damage and also further multiplies single-target DPS.[/ul][ul][/li]
[li]-150 enemy OA. I discussed its value in “Debuff All the Things” in the Analysis section, and reapersgaze touched upon it earlier as well. In short,[/li] it reduces your resummoning of pets, and more importantly, it reduces the need to kite, which hugely improves DPS.[/ul][ul]
[li]-32% bleed RR. The built-in RR for its own damage type means that Rend’s damage is far more consistent against all enemies, regardless of situation.[/ul][/li]That’s all nice and good, but what about Rend’s actual damage output when used by the Night King, especially since pet bleed damage isn’t displayed on-screen? Here are the damage factors:[ul]
[li]285 base damage[/li]
[li]+2071% pet bleed damage, thanks to +% to all pet damage - remember that pets can easily inflict any damage type![/li]
[li]-131% bleed RR (CoF + Spellblade + Rend; reduced to -119% after 1.0.3.0 nerf)[/li]
[li]3.78x max pet crit (reduced to 3.62x max pet crit after 1.0.3.0 nerf) with over 4k effective pet OA[/ul][/li]Throw in Crucible bonuses, and this comes out to well over 50k/s max bleed per target per second. As for how many targets Rend can hit…
(Screencap from 3:30 of Gladiator run)
…approximately every enemy on-screen, permanently, as if the entire arena floor were made of lava for them. Considering there can often be 20+ enemies on-screen at once in Crucible, Rend is easily capable of 500k DPS or more, just by itself. Of course all that DPS is completely spread out as AOE, but when your build already has 1M+ single-target DPS, having a very large and steady supply of AOE damage is extremely useful for improving clear times and for protection from being swarmed.
In fact, Rend is so effective that sometimes the fastest and most effective way to deal with trash is to just cast a single CoF on them and let Rend finish them off. In the same Gladiator run, you can see this is exactly what I did with the NW and NE spawns in wave 141, and the SE spawns in wave 147. This allows pets to be more efficiently deployed against targets where their massive single-target DPS can be put to better use.
tl;dr: Many pet cabalists rely on Flame Torrent to supplement their lackluster skeleton DPS because their single-target DPS and boss kill times would be extremely lacking otherwise. However, once skeleton DPS is multiplied high enough to reach top tier levels, there are other, more interesting skeleton devotion options like Rend that open up.
Bysmiel’s Command provides better pure single-target DPS, while Huntress provides a smaller amount (via +% OA) but will also shred enemy OA and inflict AOE damage. IMO you receive better returns from providing badly-needed cover to one’s weakness than from fighting diminishing returns by trying to pile one’s strengths even higher.
If you’re playing HC, the OA shred from Rend becomes even more valuable.
As described in the guide, you’ll want to:[ul]
[li]Craft at Angrim for +% armor bonuses.[/li]
[li]The process is to repeatedly craft Master’s Spellblades until you have several with “Plaguebearer’s” prefixes. Other affixes will NOT generate as high of RR values.[/li][*]Wait until your intended character is at least level 65, then craft the items with a level 12 mule and give the best Spellblade to your intended level 65+ character to equip. The leveling guide explicitly lists level 65 for this task.[/ul]