Arisaka Armor-Piercing Rounds

By BigStupidJellyfish | 2021-December-25

Dead Line Strike ICD 6s/CD 8s

Active: Increase rate of fire by 55% for 5 seconds.

Passive: Every third attack deals 1x damage with perfect accuracy and piercing effect. (If "Arisaka Rounds" are equipped, fire a piercing shot with every other attack while the active skill is in effect.)

After C14's release, the usual 3-star monthly login dolls have been replaced by new SPEQs, often including extra skills and special effects. The first of these are the Arisaka AP rounds, equippable by 4 Shiki. At maximum calibration/enhancement, they provide +180 AP and +10 FP - an upgrade over both the standard AP rounds and the M1022 Sniper Rounds. Additionally, with them equipped, 4 Shiki will fire a piercing shot every second attack during her active skill and every third otherwise, instead of every third regardless of her active skill's state. Also, note that these piercing shots cannot crit.

Today, let's look into how she compares with this against three main competitors:

  • Type88 mod
  • Normal selfbuffers
  • Herself, pre-SPEQ

Damage simulators are, as usual, useless here. This will need some testing.

Daytime#

Test set 1: no external crit damage, miscellaneous enemy groups.

Type88 has a clear performance advantage, even when not maxed out. Still, 4 Shiki's SPEQ is a clear improvement and she does about as well or better than SVD.

Test set 2: adding some crit damage with a Damage II Suee. It's a terrible fairy for combat, but I want the battles to last a bit longer so all dolls can get some value out of their skills.

4 Shiki still does better with her SPEQ than with normal AP ammo across the board, despite the crit damage aura. Type88, in either form, was still clearly the best between all options tested.

Some nightmares would get to munch on the pos 7 rifle, which Type88 avoided entirely with her melee mode.

Nighttime#

4 Shiki's piercing shots have perfect accuracy, which helps her more at night. On the other hand, Type88's skill multiplier goes up from 35% to 100%. Let's see how these factors balance out.

Against some relatively weak armor, Type88 takes the crown. Taking the extreme opposite case, 4 Shiki jumps to the front against some scouts. Either way, single-target rifles are not the way to go.

Next, some enemies with more moderate levels of evasion. Overall margins were smaller, with 4 Shiki looking best.

Final round:

There seemed to be very few differences in performance against the zombies. Against the Strelet/Gunners, 4 Shiki had a slight edge over Type88. Both were preferable when compared to Mk 12.

About Surehit#

To better understand the impact of 4 Shiki's surehit component, let's run some example numbers. Case 1: hit rates with P22/Calico/no accuracy aura, at night. The number in parentheses for 4 Shiki is the average DPS increase surehit gives her, not counting the loss of crits.

Doll Base ACC Hit rate vs 15 EVA vs 24 EVA vs 50 EVA
4 Shiki 78 51.6% (+94% DPS w/ surehit) 40.0% (+150%) 24.2% (+312%)
SVD 84 53.1% 41.5% 25.4%
Type88 73 50.0% 38.5% 23.1%

Crits are in the range of 75-150% more damage, so this matches up with my observations of 4 Shiki's strong night performance. Against decently dodgy enemies, her surehit was worth as much or more than the crits. Taking the same conditions to day is less exciting:

Doll Base ACC Hit rate vs 15 EVA vs 24 EVA vs 50 EVA
4 Shiki 78 91.6% (+9% DPS w/ surehit) 87.2% (+15%) 76.5% (+31%)
SVD 84 92.1% 88.0% 77.9%
Type88 73 91.1% 86.4% 75.4%

Note that surehit also has benefits in terms of consistency and link protection, which are hard to represent here.

... I should really re-investigate accuracy options for night RFHGs.

Conclusion#

Piercing is almost always going to be weird. Minor variations in hitbox spaghetti, kiting variations, enemy positions, and so on have a much larger impact on battle outcomes than what you get with standard single-target dolls. To be most correct, it's best to test out Type88, 4 Shiki, and whatever other alternatives you're considering if some specific fight needs major optimization.

For a general starting point, based on my tests, I'd categorize them something like this:

  • 4 Shiki's SPEQ is always preferable over standard AP ammo on her. It is technically bad for DPS against a single target (unless it has lots of evasion) as the shots can't crit, but in that case you shouldn't be using 4 Shiki at all. Use a standard selfbuffer or M200 there.
  • At night: 4 Shiki's pierce is strong against groups of enemies with substantial (>15) evasion. Otherwise, Type88 is likely to be better.
  • At day: Type88 appears to have a fairly universal advantage over 4 Shiki. She still does better than typical selfbuffers against decent groups (~30+ total links).
  • This seems to be commonly misunderstood, so to be clear: even if 4 Shiki can do better than Type88 against groups of evasive enemies, Type88 will still generally do much better against said enemies than standard selfbuffers. Hitting 2-5 enemies at once is more valuable than hitting 5-10% more of your shots.

While 4 Shiki's daytime performance was disappointing, I was pleasantly surprised by how much her surehit helped at night. M200 would be another interesting test subject for comparison, although she makes building a team fair to all options much harder.

Minor stuff

  • The enemies tested against are not representative of the total range of things you'd use RFHG echelons against. I focused on larger groups that can be handled without kiting, to get more consistent results. This hopefully shouldn't be too big of a limitation, as many of the harder/kiting-intensive enemies come in smaller numbers and thus see little benefit from piercing (Doppels, Uhlans, Hydras).
  • As with Type88, I would expect 4 Shiki to favor certain positions in different battles. To keep HG tiles constant, she was on position 7 in these tests.
  • Bosses and other lone targets were not covered here. Don't use piercing dolls for single-target DPS.
  • 4 Shiki has a sideways tile, which can give her an extra boost while on position 1.
  • Low-spec Type88 is lv110, SL10/1, not oathed. High-spec Type88 is lv116, SL10/10, oathed.
  • Type88 has an unusually high 6/8 uptime, while 4 Shiki has the typical 5/8 plus her passive.
  • I've seen this asked before, so one other detail on her pierce:

Her piercing shot does go beyond her target. Another, more extreme, example. No HOC support was used - 4 Shiki pierced and hit the far right cyclops while it was off the screen.

  • While technically released, I have yet to get any flare ammo. They will slightly lessen the relative value of 4 Shiki's surehit.

More like this