AR Damage Over Time: Chip Edition

By BigStupidJellyfish | 2020-November-07 | Rewritten 2022-January-14

There are many self-buffing ARs in the game, with all sorts of combinations of stats, buffs, and cooldowns. The relatively new #2 Processor chip has thrown a wrench into the established AR 'meta,' benefitting slower ARs much more. A number of significant SPEQ adjustments were also released.

G41 is a classic doll to use as a baseline for comparison (taking the role of "a generic selfbuffing AR"). With her own ranking SPEQ seeing a major power spike, let's redo some comparisons and see how some ARs are affected.

Before we get into things, a disclaimer: I've only selected dolls/configurations that translate rather simply into theoretical DPS numbers. There are a lot of ARs whose skills rely on various external factors that can't be represented well here. To name a few: literally any grenadier (416, Sopmod, ADS,...), AK-15, AK-Alfa, ACR, 64 Shiki, AN-94, G11 mod, and EM-2. Even within the realm of dolls that mostly just shoot things, other skill details can complicate matters. Consider RFB's targeting, which may significantly reduce incoming damage from Jaeger-style backlines, or G11's triple hit, which improves consistency and reduces the effect of link protection. As always, for best results, test your teams in Target Saving.

Daytime#

Starting off with some various 5-star ARs and a fairy that doesn't change anything:

This is a chart of the total damage dealt by each AR at each time, relative to the total dealt by G41. By 4 seconds, for example, Type95 is expected to have dealt ~16% more total damage than G41. By 10 seconds, however, G41 will have more than caught up and will have dealt ~10% more damage compared to Type95. This is not a chart of the instantaneous damage inflicted by each.

While STAR can't use chips, she does get a decent boost out of an RMR if no crit rate tiles are involved. M4A1's profile may look abnormally flat - this assumes you're keeping the team around, and her skill1 has the exact same ICD/CD/duration/multiplier as G41's. G36 does quite well in this chart, though her position is rather unstable - ROF tiles, increasing crit damage, or kiting away one of the dolls assumed to be on her tiles will reduce/remove her advantage over G41.

Command is still a strong option for ARSMGs. Switching out some equipment as appropriate, we see a relative gain from dolls lucky enough to have a buffed SPEQ and from AR-15:

RO's neural upgrade is coming soon, so it's reasonable to expect some ROF tiles in your ARSMGs:

As expected, G36/AR-15/G11 drop off a bit due to their high base/in-skill ROF. This won't apply as much for G11 post-mod, as she gets a SPEQ that gives her extra FP and crit damage instead of ROF (reducing overcap when paired with ROF tiles). G36 goes completely flat in the right example as she just barely makes it to the ROF cap pre-skill ( \( \lfloor(76+25)*1.15\rfloor=116 \) ), so her skill2 makes no difference and her skill1 lines up perfectly with G41/M4's.

Another popular configuration is to pair UMP45 with M4A1. M4 can't exactly buff herself, so I'll put R5 in her place:

All this crit rate means STAR can go back to her mod SPEQ, and it looks like she and G41 like this configuration a lot.

Continuing to generate all these variant charts for other sets of ARs would take a while, so here are some more generic cases with new dolls:

And with Command:

  • For typical ARSMG battle durations (6-12s), AK-12 is now much worse than G41 (for more AK-12 equipment comparisons, see these charts).
  • MDR does quite nicely, especially with lower crit auras. Her partial ROF selfbuff means she'll do worse, relatively, as you add ROF tiles (similarly to G36 earlier).
  • TAR-21 doesn't lose ground when switching to Command, but she wasn't great to begin with.

A comparison focusing on AUG and Angelica:

Nighttime#

Chips are generally awesome at day, but they become even better at night - PEQs are almost mandatory, so crits are less valuable to start with. Let's see how some night ARs stack up.

Interestingly, ARs care about crits so little here that G41's SPEQ is suboptimal compared to a chip. It technically has a higher FP bonus, which helps against armor, but that's not a scenario you should be using G41 for anyways. With appropriate equipment G41 is solidly above AR-15, but below the night specialists as long as their skills get a chance to go off. If you weren't already convinced, TKB is very lackluster here for a 5-star night specialist. And if you weren't around for Deep Dive or are from the future with her chip-slot SPEQ, here are some more comparisons with 9A-91.

Crit damage auras only benefit AR-15 and AS Val here. AR-15 does gain enough to be preferable over G41 at any timeframe, assuming no armor (where she'll drop off much faster). TKB moves up by maybe a percentage point or two.

Note that I'm assuming 100% uptime on AS Val here. You may need to kite Python a little bit to delay some stacks.

Other Comments#

Most of the dolls here aren't all that meta - RFHG continues to dominate that landscape. Upcoming ranking events are unlikely to call for anything far beyond M4A1/G11/416. 416 and M4's cannon mode aren't covered here as there's no reasonable way to plot AOE on a chart like this. Similar deal with G11's mod and it's HP percentage-based damage.

Your dolls probably aren't in echelons all by themselves (unless it's some cheap farm route), so other skills and tilebuffs will change how things work out. Add a Fervor talent for even more fun. Some ARs will have better or worse tilebuffs themselves for your frontline. Enemy stats will also vary per encounter. These plots use the average total damage per time interval. Critical hits, different number of frames between attacks, accuracy rolls, and so on will make the percent difference vary per combat. The frame timings of each attack may also cause slight differences with simulators. Think of this as a more general/average relationship.

Old (pre-chip) version of this page is here.

Bonus representation - to see about where these plots converge to over very long periods, here are some example semilog plots going up to 10000 seconds:

In a sense, chips make ARSMG teambuilding super annoying. So many dolls have the potential to be strong, but are incredibly sensitive to a number of conditions (crit rate/crit damage/ROF buffs, mostly). Dolls will want to change equipment at different times, changing their relative performance.

... Does it count as a "bright side" that almost all remotely competitively viable ARs have chip-replacement SPEQs with no crit damage penalty, so you're free to just VFL+SPEQ?

More like this