Part 5: advanced topics

from samurai import *
from plotting import *

import matplotlib.pyplot as plt
import pandas as pd
import numpy as np
import seaborn as sb
from IPython.display import display
%matplotlib inline

pd.options.display.max_rows = None

Return home

Table of contents

• On buff priority
• On Meikyo Shisui
• On Kenki spending
• On Hissatsu: Gyoten / Yaten
• On Third Eye
• On Hagakure
• On Higanbana

On buff priority

SAM has two self buffs and one target debuff to apply. The slashing resistance down debuff may be applied by WAR and NIN within 2-3 GCDs, letting you defer Yukikaze. What order should you apply your buffs and debuffs from a fresh start? Let’s examine some rotations that differ only in the buff/debuff application order.

Shifu, Jinpu, Yukikaze

sam = Samurai(kenki_mastery=2)

actions = [('Hakaze'), ('Shifu'), ('Kasha'),
           ('Higanbana'),
           ('Hakaze'), ('Jinpu'), ('Gekko'),
           ('Hakaze'), ('Yukikaze'),
           ('Hakaze', 'Hissatsu: Shinten'), ('Shifu', 'Hissatsu: Shinten'), ('Kasha', 'Hagakure'),
           ('Hakaze', 'Hissatsu: Guren'), ('Yukikaze', 'Hissatsu: Shinten'),
           ('Hakaze', 'Hissatsu: Shinten'), ('Jinpu'), ('Gekko'),
           ('Hakaze'), ('Shifu'), ('Kasha', 'Hissatsu: Kaiten'),
           ('Midare Setsugekka')]

df_buffs_0, average_potency, pps = sam.parse_rotation(actions)
display(df_buffs_0)

average potency per GCD = 517.7471428571428
average potency per second = 238.436184211

	Time	Weaponskill	Abilities	Potency	Jinpu	Shifu	Yukikaze	Higanbana	Kenki	Total Potency
0	0.00	Hakaze		150.000	False	False	False	0	0	150.000
1	2.40	Shifu		280.000	False	False	False	0	5	430.000
2	4.56	Kasha		400.000	False	True	False	0	10	830.000
3	6.72	Higanbana		425.200	False	True	False	0	20	1255.200
4	8.88	Hakaze		175.200	False	True	False	1	20	1430.400
5	11.04	Jinpu		305.200	False	True	False	1	25	1735.600
6	13.20	Gekko		485.200	True	True	False	1	30	2220.800
7	15.36	Hakaze		197.700	True	True	False	1	40	2418.500
8	17.52	Yukikaze		416.200	True	True	False	1	45	2834.700
9	19.68	Hakaze	(Hissatsu: Shinten,)	599.625	True	True	True	1	55	3434.325
10	21.84	Shifu	(Hissatsu: Shinten,)	765.570	True	True	True	1	35	4199.895
11	24.00	Kasha	(Hagakure,)	535.800	True	True	True	1	15	4735.695
12	26.16	Hakaze	(Hissatsu: Guren,)	1237.875	True	True	True	1	85	5973.570
13	28.32	Yukikaze	(Hissatsu: Shinten,)	842.160	True	True	True	1	40	6815.730
14	30.48	Hakaze	(Hissatsu: Shinten,)	599.625	True	True	True	1	25	7415.355
15	32.64	Jinpu		382.620	True	True	True	1	5	7797.975
16	34.80	Gekko		535.800	True	True	True	1	10	8333.775
17	36.96	Hakaze		216.675	True	True	True	1	20	8550.450
18	39.12	Shifu		382.620	True	True	True	1	25	8933.070
19	41.28	Kasha	(Hissatsu: Kaiten,)	535.800	True	True	True	1	30	9468.870
20	43.44	Midare Setsugekka		1403.820	True	True	True	1	20	10872.690

Jinpu, Shifu, Yukikaze

sam = Samurai(kenki_mastery=2)

actions = [('Hakaze'), ('Jinpu'), ('Gekko'),
           ('Higanbana'),
           ('Hakaze'), ('Shifu'), ('Kasha'),           
           ('Hakaze'), ('Yukikaze'),
           ('Hakaze', 'Hissatsu: Shinten'), ('Jinpu', 'Hissatsu: Shinten'), ('Gekko', 'Hagakure'),
           ('Hakaze', 'Hissatsu: Guren'), ('Yukikaze', 'Hissatsu: Shinten'),
           ('Hakaze', 'Hissatsu: Shinten'), ('Shifu'), ('Kasha'),
           ('Hakaze'), ('Jinpu'), ('Gekko', 'Hissatsu: Kaiten'),
           ('Midare Setsugekka')]

df_buffs_1, average_potency, pps = sam.parse_rotation(actions)
display(df_buffs_1)

average potency per GCD = 530.2328571428571
average potency per second = 239.151417526

	Time	Weaponskill	Abilities	Potency	Jinpu	Shifu	Yukikaze	Higanbana	Kenki	Total Potency
0	0.00	Hakaze		150.000	False	False	False	0	0	150.000
1	2.40	Jinpu		280.000	False	False	False	0	5	430.000
2	4.80	Gekko		460.000	True	False	False	0	10	890.000
3	7.20	Higanbana		492.200	True	False	False	0	20	1382.200
4	9.60	Hakaze		204.700	True	False	False	1	20	1586.900
5	12.00	Shifu		354.200	True	False	False	1	25	1941.100
6	14.16	Kasha		488.980	True	True	False	1	30	2430.080
7	16.32	Hakaze		201.480	True	True	False	1	40	2631.560
8	18.48	Yukikaze		419.980	True	True	False	1	45	3051.540
9	20.64	Hakaze	(Hissatsu: Shinten,)	603.405	True	True	True	1	55	3654.945
10	22.80	Jinpu	(Hissatsu: Shinten,)	769.350	True	True	True	1	35	4424.295
11	24.96	Gekko	(Hagakure,)	539.580	True	True	True	1	15	4963.875
12	27.12	Hakaze	(Hissatsu: Guren,)	1241.655	True	True	True	1	85	6205.530
13	29.28	Yukikaze	(Hissatsu: Shinten,)	845.940	True	True	True	1	40	7051.470
14	31.44	Hakaze	(Hissatsu: Shinten,)	603.405	True	True	True	1	25	7654.875
15	33.60	Shifu		386.400	True	True	True	1	5	8041.275
16	35.76	Kasha		539.580	True	True	True	1	10	8580.855
17	37.92	Hakaze		220.455	True	True	True	1	20	8801.310
18	40.08	Jinpu		386.400	True	True	True	1	25	9187.710
19	42.24	Gekko	(Hissatsu: Kaiten,)	539.580	True	True	True	1	30	9727.290
20	44.40	Midare Setsugekka		1407.600	True	True	True	1	20	11134.890

Yukikaze, Jinpu, Shifu

sam = Samurai(kenki_mastery=2)

actions = [('Hakaze'), ('Yukikaze'),
           ('Higanbana'),
           ('Hakaze'), ('Jinpu'), ('Gekko'),
           ('Hakaze'), ('Shifu'), ('Kasha', 'Hissatsu: Shinten'),           
           ('Hakaze', 'Hissatsu: Shinten'), ('Yukikaze', 'Hagakure'),
           ('Hakaze', 'Hissatsu: Guren'), ('Jinpu', 'Hissatsu: Shinten'), ('Gekko', 'Hissatsu: Shinten'),
           ('Hakaze'), ('Yukikaze'),
           ('Hakaze'), ('Shifu'), ('Kasha', 'Hissatsu: Kaiten'),
           ('Midare Setsugekka'),
           ('Hakaze')]

df_buffs_2, average_potency, pps = sam.parse_rotation(actions)
display(df_buffs_2)

average potency per GCD = 517.4309523809524
average potency per second = 230.995960884

	Time	Weaponskill	Abilities	Potency	Jinpu	Shifu	Yukikaze	Higanbana	Kenki	Total Potency
0	0.00	Hakaze		150.000	False	False	False	0	0	150.000
1	2.40	Yukikaze		340.000	False	False	False	0	5	490.000
2	4.80	Higanbana		368.000	False	False	True	0	15	858.000
3	7.20	Hakaze		194.500	False	False	True	1	15	1052.500
4	9.60	Jinpu		338.800	False	False	True	1	20	1391.300
5	12.00	Gekko		538.600	True	False	True	1	25	1929.900
6	14.40	Hakaze		219.475	True	False	True	1	35	2149.375
7	16.80	Shifu		385.420	True	False	True	1	40	2534.795
8	18.96	Kasha	(Hissatsu: Shinten,)	918.750	True	True	True	1	45	3453.545
9	21.12	Hakaze	(Hissatsu: Shinten,)	599.625	True	True	True	1	30	4053.170
10	23.28	Yukikaze	(Hagakure,)	459.210	True	True	True	1	10	4512.380
11	25.44	Hakaze	(Hissatsu: Guren,)	1237.875	True	True	True	1	80	5750.255
12	27.60	Jinpu	(Hissatsu: Shinten,)	765.570	True	True	True	1	35	6515.825
13	29.76	Gekko	(Hissatsu: Shinten,)	918.750	True	True	True	1	15	7434.575
14	31.92	Hakaze		216.675	True	True	True	1	0	7651.250
15	34.08	Yukikaze		459.210	True	True	True	1	5	8110.460
16	36.24	Hakaze		216.675	True	True	True	1	15	8327.135
17	38.40	Shifu		382.620	True	True	True	1	20	8709.755
18	40.56	Kasha	(Hissatsu: Kaiten,)	535.800	True	True	True	1	25	9245.555
19	42.72	Midare Setsugekka		1403.820	True	True	True	1	15	10649.375
20	44.88	Hakaze		216.675	True	True	True	1	15	10866.050

Yukikaze, Shifu, Jinpu

sam = Samurai(kenki_mastery=2)

actions = [('Hakaze'), ('Yukikaze'),
           ('Higanbana'),
           ('Hakaze'), ('Shifu'), ('Kasha'),
           ('Hakaze'), ('Jinpu'), ('Gekko', 'Hissatsu: Shinten'),           
           ('Hakaze', 'Hissatsu: Shinten'), ('Yukikaze', 'Hagakure'),
           ('Hakaze', 'Hissatsu: Guren'), ('Jinpu', 'Hissatsu: Shinten'), ('Gekko', 'Hissatsu: Shinten'),
           ('Hakaze'), ('Yukikaze'),
           ('Hakaze'), ('Shifu'), ('Kasha', 'Hissatsu: Kaiten'),
           ('Midare Setsugekka'),
           ('Hakaze')]

df_buffs_3, average_potency, pps = sam.parse_rotation(actions)
display(df_buffs_3)

average potency per GCD = 510.450238095238
average potency per second = 231.421740069

	Time	Weaponskill	Abilities	Potency	Jinpu	Shifu	Yukikaze	Higanbana	Kenki	Total Potency
0	0.00	Hakaze		150.000	False	False	False	0	0	150.000
1	2.40	Yukikaze		340.000	False	False	False	0	5	490.000
2	4.80	Higanbana		368.000	False	False	True	0	15	858.000
3	7.20	Hakaze		194.500	False	False	True	1	15	1052.500
4	9.60	Shifu		338.800	False	False	True	1	20	1391.300
5	11.76	Kasha		469.200	False	True	True	1	25	1860.500
6	13.92	Hakaze		191.700	False	True	True	1	35	2052.200
7	16.08	Jinpu		336.000	False	True	True	1	40	2388.200
8	18.24	Gekko	(Hissatsu: Shinten,)	918.750	True	True	True	1	45	3306.950
9	20.40	Hakaze	(Hissatsu: Shinten,)	599.625	True	True	True	1	30	3906.575
10	22.56	Yukikaze	(Hagakure,)	459.210	True	True	True	1	10	4365.785
11	24.72	Hakaze	(Hissatsu: Guren,)	1237.875	True	True	True	1	80	5603.660
12	26.88	Jinpu	(Hissatsu: Shinten,)	765.570	True	True	True	1	35	6369.230
13	29.04	Gekko	(Hissatsu: Shinten,)	918.750	True	True	True	1	15	7287.980
14	31.20	Hakaze		216.675	True	True	True	1	0	7504.655
15	33.36	Yukikaze		459.210	True	True	True	1	5	7963.865
16	35.52	Hakaze		216.675	True	True	True	1	15	8180.540
17	37.68	Shifu		382.620	True	True	True	1	20	8563.160
18	39.84	Kasha	(Hissatsu: Kaiten,)	535.800	True	True	True	1	25	9098.960
19	42.00	Midare Setsugekka		1403.820	True	True	True	1	15	10502.780
20	44.16	Hakaze		216.675	True	True	True	1	15	10719.455

compare_n_potencies([df_buffs_0, df_buffs_1, df_buffs_2, df_buffs_3], ['Shifu, Jinpu, Yukikaze', 'Jinpu, Shifu, Yukikaze', 'Yukikaze, Jinpu, Shifu', 'Yukikaze, Shifu, Jinpu'])

The results are quite close but Jinpu and Shifu appear to be the highest priority. For total party DPS, contributions due to (~16s) earlier application of the slashing resistance down debuff should be accounted for, assuming no WAR or NIN in the party. How do priorities differ if someone else, e.g., WAR, can apply the debuff instead?

External force applying slashing resistance down early

sam = Samurai(kenki_mastery=2)
sam.applied_yukikaze = True

actions = [('Hakaze'), ('Shifu'), ('Kasha'),
           ('Higanbana'),
           ('Hakaze'), ('Jinpu'), ('Gekko'),
           ('Hakaze'), ('Yukikaze'),
           ('Hakaze', 'Hissatsu: Shinten'), ('Shifu', 'Hissatsu: Shinten'), ('Kasha', 'Hagakure'),
           ('Hakaze', 'Hissatsu: Guren'), ('Yukikaze', 'Hissatsu: Shinten'),
           ('Hakaze', 'Hissatsu: Shinten'), ('Jinpu'), ('Gekko'),
           ('Hakaze'), ('Shifu'), ('Kasha', 'Hissatsu: Kaiten'),
           ('Midare Setsugekka')]

df_buffs_0_sl, average_potency, pps = sam.parse_rotation(actions)

sam = Samurai(kenki_mastery=2)
sam.applied_yukikaze = True

actions = [('Hakaze'), ('Jinpu'), ('Gekko'),
           ('Higanbana'),
           ('Hakaze'), ('Shifu'), ('Kasha'),           
           ('Hakaze'), ('Yukikaze'),
           ('Hakaze', 'Hissatsu: Shinten'), ('Jinpu', 'Hissatsu: Shinten'), ('Gekko', 'Hagakure'),
           ('Hakaze', 'Hissatsu: Guren'), ('Yukikaze', 'Hissatsu: Shinten'),
           ('Hakaze', 'Hissatsu: Shinten'), ('Shifu'), ('Kasha'),
           ('Hakaze'), ('Jinpu'), ('Gekko', 'Hissatsu: Kaiten'),
           ('Midare Setsugekka')]

df_buffs_1_sl, average_potency, pps = sam.parse_rotation(actions)

sam = Samurai(kenki_mastery=2)
sam.applied_yukikaze = True

actions = [('Hakaze'), ('Yukikaze'),
           ('Higanbana'),
           ('Hakaze'), ('Jinpu'), ('Gekko'),
           ('Hakaze'), ('Shifu'), ('Kasha', 'Hissatsu: Shinten'),           
           ('Hakaze', 'Hissatsu: Shinten'), ('Yukikaze', 'Hagakure'),
           ('Hakaze', 'Hissatsu: Guren'), ('Jinpu', 'Hissatsu: Shinten'), ('Gekko', 'Hissatsu: Shinten'),
           ('Hakaze'), ('Yukikaze'),
           ('Hakaze'), ('Shifu'), ('Kasha', 'Hissatsu: Kaiten'),
           ('Midare Setsugekka'),
           ('Hakaze')]

df_buffs_2_sl, average_potency, pps = sam.parse_rotation(actions)

sam = Samurai(kenki_mastery=2)
sam.applied_yukikaze = True

actions = [('Hakaze'), ('Yukikaze'),
           ('Higanbana'),
           ('Hakaze'), ('Shifu'), ('Kasha'),
           ('Hakaze'), ('Jinpu'), ('Gekko', 'Hissatsu: Shinten'),           
           ('Hakaze', 'Hissatsu: Shinten'), ('Yukikaze', 'Hagakure'),
           ('Hakaze', 'Hissatsu: Guren'), ('Jinpu', 'Hissatsu: Shinten'), ('Gekko', 'Hissatsu: Shinten'),
           ('Hakaze'), ('Yukikaze'),
           ('Hakaze'), ('Shifu'), ('Kasha', 'Hissatsu: Kaiten'),
           ('Midare Setsugekka'),
           ('Hakaze')]

df_buffs_3_sl, average_potency, pps = sam.parse_rotation(actions)

average potency per GCD = 529.5511904761904
average potency per second = 243.872258772
average potency per GCD = 543.0033333333333
average potency per second = 244.911297251
average potency per GCD = 517.4309523809524
average potency per second = 230.995960884
average potency per GCD = 510.450238095238
average potency per second = 231.421740069

compare_n_potencies([df_buffs_0_sl, df_buffs_1_sl, df_buffs_2_sl, df_buffs_3_sl], ['Shifu, Jinpu, Yukikaze', 'Jinpu, Shifu, Yukikaze', 'Yukikaze, Jinpu, Shifu', 'Yukikaze, Shifu, Jinpu'])

Here I made the assumption that the slashing resistance down debuff was already applied at the beginning of each rotation. We see that the value of deferring Yukikaze increases if someone else can apply the debuff instead. Although the average potency per GCD is much higher when applying Jinpu before Shifu while the delta in pps is much smaller, the results are likely to change as your skill speed changes. Here are some points to consider:

• Jinpu increases the potency per weaponskill and ability because it is a damage dealt increased buff.
  • Increases the value of Kenki spenders while Shifu does nothing for them.
  • Increases the value of Higanbana and its DoT while Shifu directly does not.
• Shifu provides a haste-type buff. This increases auto attack and weaponskill frequency. This may allow you to fit more actions into a buff window.
  • This also increases the rate of Kenki and Sen accrual.
  • This reduced weaponskill recast time may exacerbate GCD clipping when double weaving.

Because the outcome from either priority system is so similar, external factors such as party buffs and encounter duration will be the primary driving force behind which policy you follow. The above points of consideration plus those external factors should play into how you design openers on a case-by-case basis.

back to the top…

On Meikyo Shisui

Some of the initial questions of Meikyo Shisui were if Iaijutsu would consume charges of it and if combo actions initiated during it would continue after the buff expired. Testing on the live server showed:

• Iaijutsu does not consume charges
• Combo actions do not continue after the buff expires

The first revelation increases the flexibility of Meikyo Shisui. With the buff’s 10s duration, it is possible to spend your three charges while using an Iaijutsu sometime in between. The implications of that are immense:

1) High, raw damage in 3-4 GCDs - e.g., with Setsu already open, do Kasha, Gekko+Hissatsu: Kaiten, Midare Setsugekka, Kasha. That’s 1200+1080=2280 potency in 4 GCDs plus 30 Kenki.

2) Up to 30 Kenki generated in 3-4 GCDs

3) Expedited Sen opening for rapid Iaijutsu and/or Kenki generation via Hagakure

The sheer potency and utility of Kasha, Gekko, and (to a lesser extent) Yukikaze exemplify how powerful Meikyo Shisui is. At 400 potency (340 for Yukikaze) and 10 Kenki generated for each, the opportunity to skip the lower potency and Kenki production of Hakaze, Shifu, and Jinpu is huge. From a damage and potential damage perspective, it is optimal to use Meikyo Shisui to do some combination of Kasha, Gekko, and Yukikaze with an Iaijutsu in between.

Example 1: Two of Kasha or Gekko, one of the other

sam = Samurai(kenki_mastery=2)

actions = [('Hakaze'), ('Jinpu'), ('Gekko'),
           ('Hakaze'), ('Shifu'), ('Kasha'),
           ('Hakaze'), ('Yukikaze', 'Hissatsu: Guren'),
           ('Hakaze', 'Hagakure'), ('Yukikaze', 'Meikyo Shisui'),
           ('Gekko', 'Hissatsu: Shinten'), ('Kasha', 'Hissatsu: Kaiten'),
           ('Midare Setsugekka'),
           ('Kasha', 'Hissatsu: Shinten'),
           ('Hakaze', 'Hissatsu: Shinten'), ('Jinpu'), ('Gekko'),
           ('Hakaze'), ('Yukikaze'),
           ('Hakaze'), ('Shifu', 'Hissatsu: Kaiten'),
           ('Midare Setsugekka'),
           ('Kasha', 'Hissatsu: Shinten')]

df_ms, average_potency, pps = sam.parse_rotation(actions)
display(df_ms)

average potency per GCD = 532.045652173913
average potency per second = 241.647906793

	Time	Weaponskill	Abilities	Potency	Jinpu	Shifu	Yukikaze	Higanbana	Kenki	Total Potency
0	0.00	Hakaze		150.000	False	False	False	0	0	150.000
1	2.40	Jinpu		280.000	False	False	False	0	5	430.000
2	4.80	Gekko		460.000	True	False	False	0	10	890.000
3	7.20	Hakaze		172.500	True	False	False	0	20	1062.500
4	9.60	Shifu		322.000	True	False	False	0	25	1384.500
5	11.76	Kasha		460.000	True	True	False	0	30	1844.500
6	13.92	Hakaze		172.500	True	True	False	0	40	2017.000
7	16.08	Yukikaze	(Hissatsu: Guren,)	1412.200	True	True	False	0	45	3429.200
8	18.24	Hakaze	(Hagakure,)	191.475	True	True	True	0	5	3620.675
9	20.40	Yukikaze	(Meikyo Shisui,)	434.010	True	True	True	0	70	4054.685
10	22.56	Gekko	(Hissatsu: Shinten,)	893.550	True	True	True	0	80	4948.235
11	24.72	Kasha	(Hissatsu: Kaiten,)	510.600	True	True	True	0	65	5458.835
12	26.88	Midare Setsugekka		1378.620	True	True	True	0	55	6837.455
13	29.04	Kasha	(Hissatsu: Shinten,)	510.600	True	True	True	0	55	7348.055
14	31.20	Hakaze	(Hissatsu: Shinten,)	574.425	True	True	True	0	30	7922.480
15	33.36	Jinpu		357.420	False	True	True	0	10	8279.900
16	35.52	Gekko		510.600	True	True	True	0	15	8790.500
17	37.68	Hakaze		191.475	True	True	True	0	25	8981.975
18	39.84	Yukikaze		434.010	True	False	True	0	30	9415.985
19	42.00	Hakaze		191.475	True	False	True	0	40	9607.460
20	44.16	Shifu	(Hissatsu: Kaiten,)	357.420	True	False	True	0	45	9964.880
21	46.32	Midare Setsugekka		1378.620	True	True	True	0	30	11343.500
22	48.48	Kasha	(Hissatsu: Shinten,)	893.550	True	True	True	0	30	12237.050

plot_potency(df_ms)

By using some combination of Kasha and Gekko only for our Meikyo Shisui charges, we end up with the situation where one of our Shifu or Jinpu buffs will fall off. This is due to us avoiding combos that terminate in opening a Sen that has already been opened; in this example, it was the Kasha combo leading to Shifu being dropped for about 4-5 seconds. In this instance, we moved the final Midare Setsugekka to after Hakaze -> Shifu to minimize Shifu downtime.

The penalty for dropping the Shifu or Jinpu buff is in the neighborhood of 9-15% potency per GCD. For example, dropping Jinpu for 4-5 seconds is about two GCDs plus auto attacks; Hakaze -> Jinpu is 150+280=430 potency plus maybe 2 auto attacks. This puts the raw potency lost due to a dropped buff in the order of 90-100 (auto attacks not accounted for in my modeling). This is larger than the difference between Yukikaze and Kasha or Gekko.

Example: One Yukikaze, one Kasha, one Gekko

sam = Samurai(kenki_mastery=2)

actions = [('Hakaze'), ('Jinpu'), ('Gekko'),
           ('Hakaze'), ('Shifu'), ('Kasha'),
           ('Hakaze'), ('Yukikaze', 'Hissatsu: Guren'),
           ('Hakaze', 'Hagakure'), ('Jinpu'), ('Gekko', 'Meikyo Shisui'),
           ('Yukikaze', 'Hissatsu: Shinten'), ('Kasha', 'Hissatsu: Kaiten'),
           ('Midare Setsugekka'),
           ('Gekko', 'Hissatsu: Shinten'),
           ('Hakaze', 'Hissatsu: Shinten'), ('Shifu'), ('Kasha'),
           ('Hakaze'), ('Yukikaze'),
           ('Hakaze', 'Hissatsu: Shinten'), ('Jinpu', 'Hissatsu: Kaiten'),
           ('Midare Setsugekka'),
           ('Gekko', 'Hissatsu: Shinten')]

df_ms2, average_potency, pps = sam.parse_rotation(actions)
display(df_ms2)

average potency per GCD = 556.6820833333333
average potency per second = 253.037310606

	Time	Weaponskill	Abilities	Potency	Jinpu	Shifu	Yukikaze	Higanbana	Kenki	Total Potency
0	0.00	Hakaze		150.000	False	False	False	0	0	150.000
1	2.40	Jinpu		280.000	False	False	False	0	5	430.000
2	4.80	Gekko		460.000	True	False	False	0	10	890.000
3	7.20	Hakaze		172.500	True	False	False	0	20	1062.500
4	9.60	Shifu		322.000	True	False	False	0	25	1384.500
5	11.76	Kasha		460.000	True	True	False	0	30	1844.500
6	13.92	Hakaze		172.500	True	True	False	0	40	2017.000
7	16.08	Yukikaze	(Hissatsu: Guren,)	1412.200	True	True	False	0	45	3429.200
8	18.24	Hakaze	(Hagakure,)	191.475	True	True	True	0	5	3620.675
9	20.40	Jinpu		357.420	True	True	True	0	70	3978.095
10	22.56	Gekko	(Meikyo Shisui,)	510.600	True	True	True	0	75	4488.695
11	24.72	Yukikaze	(Hissatsu: Shinten,)	816.960	True	True	True	0	85	5305.655
12	26.88	Kasha	(Hissatsu: Kaiten,)	510.600	True	True	True	0	70	5816.255
13	29.04	Midare Setsugekka		1378.620	True	True	True	0	60	7194.875
14	31.20	Gekko	(Hissatsu: Shinten,)	893.550	True	True	True	0	60	8088.425
15	33.36	Hakaze	(Hissatsu: Shinten,)	574.425	True	True	True	0	45	8662.850
16	35.52	Shifu		357.420	True	True	True	0	25	9020.270
17	37.68	Kasha		510.600	True	True	True	0	30	9530.870
18	39.84	Hakaze		191.475	True	True	True	0	40	9722.345
19	42.00	Yukikaze		434.010	True	True	True	0	45	10156.355
20	44.16	Hakaze	(Hissatsu: Shinten,)	574.425	True	True	True	0	55	10730.780
21	46.32	Jinpu	(Hissatsu: Kaiten,)	357.420	True	True	True	0	35	11088.200
22	48.48	Midare Setsugekka		1378.620	True	True	True	0	20	12466.820
23	50.64	Gekko	(Hissatsu: Shinten,)	893.550	True	True	True	0	20	13360.370

plot_potency(df_ms2)

compare_n_potencies([df_ms, df_ms2], ['Gekko/Kasha only', 'Yukikaze + Gekko/Kasha'])

The advantage of using Yukikaze in your Meikyo Shisui window ensures you don’t need to worry about dropping a buff. The final weaponskill you use on Meikyo Shisui should be the one of Kasha or Gekko corresponding to the buff with the longest remaining duration. For example, if Shifu has more duration than Jinpu, finish the window with Kasha then follow up with the Gekko combo. The auto-attack penalty from dropping a buff isn’t accounted for here but regardless, the second policy comes out on top. (Generating extra Kenki for Hissatsu: Shinten doesn’t hurt either.)

back to the top…

On Kenki spending

As I’ve iterated, Kenki grants access to just about all of your damage-dealing abilities which are, cumulatively, an enormous share of your potential DPS. Economical spending of Kenki is therefore essential to maximizing your DPS.

Throughout the leveling process, we gain a number of Kenki spenders. We examine their value by their potency per Kenki (ppK) in order to evaluate their usefulness.

For single target situations:

1) Hissatsu: Guren (16 ppK)

1) Although at a lower ppK than Hissatsu: Kaiten on Iaijutsu, its potency is from the ability and not a weaponskill. It does not directly compete with Iaijutsu; it only competes via Hagakure usage.

2) Hissatsu: Kaiten on Higanbana (full duration) (23.5 ppK)

2) Hissatsu: Kaiten on Midare Setsugekka (18 ppK)

3) Hissatsu: Seigan (~13 ppK)

4) Hissatsu: Shinten (12 ppK)

5) Hissatsu: Gyoten / Yaten (10 ppK)

5) Hissatsu: Kaiten on Gekko / Kasha (10 ppK)

For multiple target situations:

1) Hissatsu: Guren (>16 ppK)

2) Hissatsu: Kaiten on Tenka Goken (>9 ppK)

3) Hissatsu: Kyuten (6n ppK, where n is the number of targets)

1) We examined the tradeoff between 2) and 3) in Part 4. In short, always have Kenki to use on Hissatsu: Kaiten for Tenka Goken. Excess Kenki goes to Hissatsu: Kyuten. With the Fuga, Oka, Fuga, Mangetsu combo generating 30 Kenki total, this should always be possible.

Wasted Kenki

Kenki capping is a risk especially when using Hagakure to convert three Sen. The penalty for doing so may be determined via the ppK values above. Since Kenki is quantized in units of 5, the potency losses are not insignificant. These losses are, however, only realized if they would have been spendable in the first place. For example, if you overflowed Kenki by 5 but ended the encounter with 0 Kenki, then you haven’t lost anything because that 5 Kenki could not have been spent anyway. However, if you instead ended with 5 Kenki, it is a loss of 100 potency because you could have spent the 10 Kenki on Hissatsu: Gyoten / Yaten.

Along those lines, while we strive to convert all Kenki into Hissatsu: Seigan / Shinten (in single target situations), ending an encounter with Kenki left on the table is a definite loss. Although typically less than ideal, dump Kenki into Hissatsu: Gyoten if it means Kenki goes unused otherwise. Alternatively, if you have 20 Kenki to spend, use it on Hissatsu: Kaiten for Gekko or Kasha.

back to the top…

On Hissatsu: Gyoten / Yaten

The difference in ppK between Hissatsu: Shinten and Hissatsu: Gyoten is miniscule but there. The utility of a gap closer is increasing uptime via reduction of transit time. The increase in auto attacks from using Hissatsu: Gyoten should alone make it worth it; the increased weaponskill uptime for Sen and Kenki generation seals the deal. Use Hissatsu: Gyoten whenever you need the gap closer.

Hissatsu: Yaten is a little more subtle. The Enhanced Enpi proc granted increases the potency of the next Enpi you use to 300. Using this Enhanced Enpi gives you time to move back into range and refunds the Kenki cost of using Hissatsu: Yaten. Furthermore, the increased potency offsets the loss from not being able to do your weaponskill combos. (The Gekko/Kasha combos average ~276 potency per GCD plus the Kenki generated.) You may also argue that if you use Hissatsu: Yaten to stay active longer before disengaging to dodge adds benefit from increased uptime.

If our goal is to not take damage from point blank AoE attacks, then our options are:

1) Disengage early to dodge, then:

1) Run back into range

2) Hissatsu: Gyoten back into range

2) Use Hissatsu: Yaten to dodge, then:

1) Run back into range

2) Enpi, then run back into range

3) Enpi, then Hissatsu: Gyoten back into range

Disengaging early and running back into range saves Kenki at the cost of reduced uptime. If that saved Kenki is instead spent on Hissatsu: Shinten, that is a gain of 2 ppK. If we lost weaponskills due to lost uptime, e.g., we lost between 5 and 10 Kenki in addition to the weaponskill potency and auto attacks. It is difficult to imagine a scenario where we would want to take the loss in uptime.

Using Hissatsu: Yaten to dodge incurs the cost of the reduced ppK but not of reduced uptime or related weaponskill losses (since it is off the GCD). The next decision is if you use the Enhanced Enpi proc or not. As stated, the increased potency plus Kenki refund is more than sufficient to make it worth using, especially if the alternative is doing nothing. Doing nothing in lieu of Enpi may be attractive if you are mid-combo since Enpi breaks combo. Breaking combo delays Sen opening, Kenki generation, and devalues the combo actions already taken.

If you backstep mid-combo prior to Kasha or Gekko, you may preserve your combo by doing nothing and waiting to reengage. The outcome is 400 potency, an open Sen, and the 10 Kenki cost of Hissatsu: Yaten refunded. The opened Sen may be valued by its Kenki equivalent (via Hagakure) of 20 – a value of 240 potency if converted into Hissatsu: Shinten. Alternatively, it may be valued for its progress towards Midare Setsugekka – a value of 240 or 360 potency. Compare this to the 300 potency of Enhanced Enpi plus refunded Kenki cost with a broken combo. When faced with these two options, preserving your combo by doing nothing is the attractive option.

Breaking combo after only Hakaze appears to lean differently. It is a similar value proposition as the previous scenario if the intended combo path were Yukikaze. In this case, you should elect to preserve your combo for Yukikaze. However, if the intended combo paths end in Kasha or Gekko, breaking combo with Enhanced Enpi looks more attractive:

• The next combo step after Hakaze is 280 potency plus 5 Kenki and the opportunity to finish the combo.
• There is less of a sunk cost if the combo is broken (only Hakaze will be “wasted”). Essentially, you defer the combo by two GCDs and write off the Hakaze.

The Enhanced Enpi covers the immediate next GCD (Jinpu or Shifu) with its increased potency and higher Kenki generated. It also covers the cost of doing nothing instead. So while you ideally should not use Hissatsu: Yaten to disengage while in the middle of a combo, you should not use Enhanced Enpi unless you’re only on the Hakaze step of the combo and intend to finish the combo with Kasha or Gekko. In that case, you should break combo with Enhanced Enpi.

The final decision is if you use Hissatsu: Gyoten to get back into range. Once again, the tradeoff is reduced Kenki efficiency versus increased uptime. If the dodge maneuver is executed efficienctly, then the GCD after using Enpi will most likely be enough to return to the target without clipping the following GCD. Therefore, the tension is between auto attack uptime and Kenki savings. If auto attacks are approximately 100 potency, then using Hissatsu: Gyoten to close the gap is worth it if at least one auto attack is gained as a result. However, if the Kenki spent on Hissatsu: Gyoten prevents you from using a Hissatsu: Shinten later on, then running back into range would have been the better option. The payoff in using the gap closer is minimal compared to the potential penalty for inefficient Kenki use (~100 potency for a gained auto attack versus 200 potency loss for insufficient Kenki for Hissatsu: Shinten). I would imagine that the expected payoff is optimal if you elect to run back into range instead of using Hissatsu : Gyoten.

back to the top…

On Third Eye

This defensive ability with a short cooldown offers a modest one time damage taken reduction. Additionally, it grants the Open Eyes status, enabling the use of either Merciful Eyes or Hissatsu: Seigan.

Merciful Eyes’s healing potency was reduced since the media tour tool tip release by a factor of 3. What was once a frequently available Second Wind alternative has been reduced to a pitiful 200 potency heal. In my experience, this is enough to offset an auto attack while, for example, face-tanking a FATE boss. As such, use Open Eyes for Merciful Eyes if you are desparate.

On the other hand, the ability formerly known as Starry Eyes, Hissatsu: Seigan, which was useless due to its high Kenki cost according to the media tour tool tips, has since found its calling along with its renaming to be inline with the other Kenki spenders. At a potency of 200 with a Kenki cost of 15, its potency per Kenki ratio of ~13 ranks it third among Kenki spenders, beating out the gold standard of Hissatsu: Shinten (at a potency per Kenki ratio of 12).

This grants an avenue of optimization in which you try to make sure you activate Third Eye before as many (unavoidable) damage situations as you can. Kenki that would have gone to a Hissatsu: Shinten dump may then be used on Hissatsu: Seigan instead for a theoretical DPS gain. The modest duration of the Open Eyes status (15s) enables you to spend the Open Eyes at your convenience (e.g., you currently only have enough Kenki for an upcoming Hissatsu: Kaiten for Midare Setsugekka, which takes priority).

Of course, this theoretical DPS gain means nothing if the Kenki difference is not made up by the end of the encounter; if you end the encounter with 10 spare Kenki, then a Hissatsu: Shinten should’ve been used instead.

Try taking advantage of Third Eye on fights like Susano EX! There’s raidwide damage with cast bars (Unkehi), targeted damage (stack after knockback, targeted lightning), raidwide damage without cast bars (during the parry active time maneuver).

back to the top…

On Hagakure

As discussed in part 4, Hagakure is a subtly vital cooldown. By converting open Sen into Kenki, it essentially is shifting damage potential from being on the GCD (via Iaijutsu weaponskills) to being off the GCD (via Kenki spenders). This effectively makes the damage potential “free” because it’s no longer constrained to the GCD. This pushes your rotation up a GCD because you no longer need to cast the Iaijutsu to realize that potential.

Besides Higanbana, Midare Setsugekka is our most powerful Iaijutsu at 720 potency for three Sen. If Hagakure is used to convert this three Sen into Kenki, 60 Kenki is generated. If this Kenki is used for Hissatsu: Shinten at 12 ppK, the equivalent 720 potency is there. Thus, with our “gold standard” of Kenki spenders, all the potency is accounted for. In fact, one does not need to exclusively convert three Sen using Hagakure because it would still be efficient to convert one or two Sen due to the exchange rate. If the generated Kenki is used on Hissatsu: Seigan or other higher efficiency Kenki spenders, the value of Hagakure increases further.

When might it be useful to use Hagakure on one or two Sen?

• Getting Hagakure on cooldown sooner

• Adjusting your Midare Setsugekka timing to align with encounter mechanics and/or party buffs

• Fast loans of Kenki for on-demand burst potential (e.g., add phase)

The value of Hagakure depends on being able to convert the Kenki back into potency. You may look at the ppK list and say, “Hissatsu: Kaiten plus Midare Setsugekka is the highest efficiency use of Kenki. Why should we ever convert three Sen into Kenki?” While it’s true that Hissatsu: Kaiten + Midare Setsugekka is the best bang for your buck in terms of ppK, the ability to move the damage from Midare Setsugekka off the GCD provides the actual gains in pps, which is what actually matters.

Example

sam = Samurai(kenki_mastery=2)

actions = [('Hakaze'), ('Jinpu'), ('Gekko'),
           ('Hakaze'), ('Shifu'), ('Kasha'),
           ('Hakaze'), ('Yukikaze', 'Hissatsu: Guren'),
           ('Hakaze', 'Hagakure'), ('Jinpu', 'Hissatsu: Shinten'), ('Gekko', 'Meikyo Shisui'),
           ('Yukikaze', 'Hissatsu: Shinten'), ('Kasha', 'Hissatsu: Kaiten'),
           ('Midare Setsugekka'),
           ('Hakaze', 'Hissatsu: Shinten'), ('Yukikaze', 'Hissatsu: Shinten')]

df_hgk1, average_potency, pps = sam.parse_rotation(actions)
display(df_hgk1)

average potency per GCD = 560.5756250000001
average potency per second = 252.511542793

	Time	Weaponskill	Abilities	Potency	Jinpu	Shifu	Yukikaze	Higanbana	Kenki	Total Potency
0	0.00	Hakaze		150.000	False	False	False	0	0	150.000
1	2.40	Jinpu		280.000	False	False	False	0	5	430.000
2	4.80	Gekko		460.000	True	False	False	0	10	890.000
3	7.20	Hakaze		172.500	True	False	False	0	20	1062.500
4	9.60	Shifu		322.000	True	False	False	0	25	1384.500
5	11.76	Kasha		460.000	True	True	False	0	30	1844.500
6	13.92	Hakaze		172.500	True	True	False	0	40	2017.000
7	16.08	Yukikaze	(Hissatsu: Guren,)	1412.200	True	True	False	0	45	3429.200
8	18.24	Hakaze	(Hagakure,)	191.475	True	True	True	0	5	3620.675
9	20.40	Jinpu	(Hissatsu: Shinten,)	740.370	True	True	True	0	70	4361.045
10	22.56	Gekko	(Meikyo Shisui,)	510.600	True	True	True	0	50	4871.645
11	24.72	Yukikaze	(Hissatsu: Shinten,)	816.960	True	True	True	0	60	5688.605
12	26.88	Kasha	(Hissatsu: Kaiten,)	510.600	True	True	True	0	45	6199.205
13	29.04	Midare Setsugekka		1378.620	True	True	True	0	35	7577.825
14	31.20	Hakaze	(Hissatsu: Shinten,)	574.425	True	True	True	0	35	8152.250
15	33.36	Yukikaze	(Hissatsu: Shinten,)	816.960	True	True	True	0	15	8969.210

sam = Samurai(kenki_mastery=2)

actions = [('Hakaze'), ('Jinpu'), ('Gekko'),
           ('Hakaze'), ('Shifu'), ('Kasha'),
           ('Hakaze'), ('Yukikaze', 'Hissatsu: Kaiten'),
           ('Midare Setsugekka'),
           ('Hakaze'), ('Jinpu'), ('Gekko', 'Meikyo Shisui'),
           ('Yukikaze', 'Hissatsu: Guren'), ('Kasha', 'Hissatsu: Kaiten'),
           ('Midare Setsugekka'), ('Gekko')]

df_hgk2, average_potency, pps = sam.parse_rotation(actions)
display(df_hgk2)

average potency per GCD = 543.8215625
average potency per second = 244.964667793

	Time	Weaponskill	Abilities	Potency	Jinpu	Shifu	Yukikaze	Higanbana	Kenki	Total Potency
0	0.00	Hakaze		150.000	False	False	False	0	0	150.000
1	2.40	Jinpu		280.000	False	False	False	0	5	430.000
2	4.80	Gekko		460.000	True	False	False	0	10	890.000
3	7.20	Hakaze		172.500	True	False	False	0	20	1062.500
4	9.60	Shifu		322.000	True	False	False	0	25	1384.500
5	11.76	Kasha		460.000	True	True	False	0	30	1844.500
6	13.92	Hakaze		172.500	True	True	False	0	40	2017.000
7	16.08	Yukikaze	(Hissatsu: Kaiten,)	391.000	True	True	False	0	45	2408.000
8	18.24	Midare Setsugekka		1378.620	True	True	True	0	35	3786.620
9	20.40	Hakaze		191.475	True	True	True	0	35	3978.095
10	22.56	Jinpu		357.420	True	True	True	0	40	4335.515
11	24.72	Gekko	(Meikyo Shisui,)	510.600	True	True	True	0	45	4846.115
12	26.88	Yukikaze	(Hissatsu: Guren,)	1455.210	True	True	True	0	55	6301.325
13	29.04	Kasha	(Hissatsu: Kaiten,)	510.600	True	True	True	0	15	6811.925
14	31.20	Midare Setsugekka		1378.620	True	True	True	0	5	8190.545
15	33.36	Gekko		510.600	True	True	True	0	5	8701.145

compare_n_potencies([df_hgk1, df_hgk2], ['Using Hagakure', 'No Hagakure'])

The two policies diverge in a number of spots:

• Deferring the first Midare Setsugekka lets us spend Kenki on Hissatsu: Guren earlier. This basically already covers the difference of using Hagakure instead of Midare Setsugekka.

• Using Hagakure gave us four Hissatsu: Shinten compared to not using it at the expense of one Midare Setsugekka. This is essentially where the cumulative potency was derived from. In addition to the Kenki converted using Hagakure, we were able to generate more Kenki from weaponskills alone.

• At the end of these simulations, the one with the policy of using Hagakure ends with more Kenki (=> potential damage) and marginally higher pps.

To sum up:

• Hagakure converts damage potential from being on the GCD with Iaijutsu to being off the GCD with Kenki spenders.
  • More specifically, it essentially turns Midare Setsugekka into Kenki.

• The value in the conversion is only realized if the Kenki is actually spent. While the conversion gives you flexibility in when/how you spend it (i.e., you can bank the Kenki at your convenience), the payoff period takes a little longer. For example, you would need to use Hissatsu: Shinten 3-4 times over 3-4 GCDs. This has the potential to lead to wasted Kenki if you’re not careful.

• Hagakure does not need to always be spent on three Sen for benefit.

• You can use Hagakure to adjust your rotation timings by clearing open Sen.

Using Hagakure at fewer than 3 Sen

This is a mildly controversial topic. Here are the facts:

• Hagakure has a short 40s recast time.

• Therefore, waiting to open 3 Sen to use Hagakure while it is available may lead to fewer uses of Hagakure during the encounter.

  • Fewer uses => less potential damage.
• Hagakure is always a DPS gain if the Kenki generated with it is consumed.
• Hagakure has a linear relationship between utility and the number of Sen converted due to the 20:1 Kenki:Sen conversion rate.
• Because Hagakure has a recast time and has a finite number of uses, we need to consider that we cannot always convert Sen on demand with it.

• There are other subtle, less tangible aspects affected by these decisions.

  • Higanbana DoT clipping vs. dropping due to Sen alignment
  • Encounter specifics such as forced downtime and encounter duration

Therefore, the driving factor is not the number of uses of Hagakure in an encounter, but how many Sen are converted with it. This is unlike more straightforward damage cooldowns like Internal Release where you can pretty much use them on cooldown without stringent regard for timing.

For example, if we squeeze out an extra Hagakure use because we used Hagakure at one Sen twice compared to holding it to convert three Sen, we actually converted one fewer Sen even though we used Hagakure one more time. (i.e., if we held Hagakure to convert at three Sen instead, we would have converted 3+3=6 Sen total. On the other hand, we converted 1+1+3=5 Sen total, even with the extra Hagakure usage.)

It takes 2-3 GCDs, or roughly 4-7 seconds to open a Sen. Worst case scenario: Hagakure comes off cooldown when you already have three Sen open and just complete your next combo. This means you need 8 GCDs to open all three Sen if you intend on holding Hagakure to convert them; this is about 17-18 seconds. If this happens twice in an encounter, that holding time is about equal to an extra Hagakure.

With a perfect mapping of cooldowns throughout an encounter, you can probably decide the best ways to use Hagakure. Without prior knowledge of how an encounter will play out, you can safely follow some rules of thumb:

• Use Hagakure at one or two Sen if you would not otherwise be able to reach three Sen and use Midare Setsugekka.
  • e.g., the end of an encounter. Having three Sen open only pays off if you use Midare Setsugekka or if you convert it with Hagakure and spend all the resulting Kenki. This lets you get a return on your investment rather than forfeiting it, so to speak.

• Use Hagakure at one or two Sen before forced downtime, especially if the Getsu (from Gekko) and/or Ka (from Kasha) Sen are open.

  • If you will need to reapply Shifu and/or Jinpu when the encounter resumes, you will waste the Sen anyway.
  • If the downtime is one where you can use Meditate for a while, be wary of capping Kenki.
  • As an example, if you have Sen going into the sword phase of Susano EX, it’s a good idea to convert them during the transition. Since this transition is also a good time to get a full Meditate, be sure to dump Kenki beforehand.

• For near 100% uptime encounters, it’s a safe bet to hold Hagakure for three Sen each time unless you really need the Kenki on demand. One may argue that you may miss out on Hagakure usages if you hold it off cooldown for too much time, the potential payoff does not match the risk in my mind.

  • Keeping Hagakure aggressively on cooldown is banking on extra Hagakure uses leading to more total Sen converted. If you keep using Hagakure at one Sen while following this policy, your chances of breaking even in Sen conversion becomes more and more unlikely. Two “premature” conversions at one Sen will require at least two extra rewarded Hagakure uses to make up the difference.

• For the “worst case scenario” where Hagakure comes off cooldown immediately after you spent three Sen, using Hagakure at two Sen may be the best compromise.

  • You shave three GCDs off of holding time compared to waiting for three Sen.
  • You only need to gain a Hagakure use and convert at one Sen to break even.

back to the top…

On Higanbana

At over effectively 1410 potency in a single GCD, Higanbana’s efficiency is unmatched. However, since its payoff occurs over a 60s DoT, if most of the DoT does not tick, then we would have been better off not applying Higanbana. How much is enough?

Applying Higanbana

We’ll initialize with 20 Kenki in the gauge and Jinpu and Shifu already active but no Sen currently open. We’ll assume, without loss of generality, that Hagakure, Meikyo Shisui, and Hissatsu: Guren are available.

sam = Samurai(base_gcd=2.40, kenki_mastery=2, kenki_gauge=20)

sam.has_jinpu = True
sam.has_shifu = True

actions = [('Hakaze'), ('Yukikaze', 'Hissatsu: Kaiten'), ('Higanbana'), 
           ('Hakaze'), ('Jinpu'), ('Gekko'),
           ('Hakaze'), ('Shifu'), ('Kasha', 'Hissatsu: Guren'),
           ('Hakaze'), ('Yukikaze', 'Hagakure'),
           ('Hakaze', 'Hissatsu: Shinten'), ('Jinpu', 'Hissatsu: Shinten'), ('Gekko', 'Meikyo Shisui'),
           ('Yukikaze', 'Hissatsu: Shinten'), ('Kasha', 'Hissatsu: Kaiten'), ('Midare Setsugekka'), ('Gekko', 'Hissatsu: Shinten'),
           ('Hakaze'), ('Yukikaze'),
           ('Hakaze'), ('Shifu'), ('Kasha', 'Hissatsu: Shinten'),
           ('Hakaze'), ('Jinpu'), ('Gekko', 'Hissatsu: Shinten'),
           ('Hakaze'), ('Shifu', 'Hagakure'), ('Kasha', 'Hissatsu: Kaiten'), ('Higanbana')]

df_higanbana_true, _, _ = sam.parse_rotation(actions)
display(df_higanbana_true)

average potency per GCD = 536.2691666666665
average potency per second = 248.272762346

	Time	Weaponskill	Abilities	Potency	Jinpu	Shifu	Yukikaze	Higanbana	Kenki	Total Potency
0	0.00	Hakaze		150.000	True	True	False	0	20	150.000
1	2.16	Yukikaze	(Hissatsu: Kaiten,)	340.000	False	False	False	0	25	490.000
2	4.32	Higanbana		377.800	False	False	True	0	15	867.800
3	6.48	Hakaze		204.300	False	False	True	1	15	1072.100
4	8.64	Jinpu		348.600	False	False	True	1	20	1420.700
5	10.80	Gekko		548.400	True	False	True	1	25	1969.100
6	12.96	Hakaze		229.275	True	False	True	1	35	2198.375
7	15.12	Shifu		395.220	True	False	True	1	40	2593.595
8	17.28	Kasha	(Hissatsu: Guren,)	1569.600	True	True	True	1	45	4163.195
9	19.44	Hakaze		229.275	True	True	True	1	5	4392.470
10	21.60	Yukikaze	(Hagakure,)	471.810	True	True	True	1	10	4864.280
11	23.76	Hakaze	(Hissatsu: Shinten,)	612.225	True	True	True	1	80	5476.505
12	25.92	Jinpu	(Hissatsu: Shinten,)	778.170	True	True	True	1	60	6254.675
13	28.08	Gekko	(Meikyo Shisui,)	548.400	True	True	True	1	40	6803.075
14	30.24	Yukikaze	(Hissatsu: Shinten,)	854.760	True	True	True	1	50	7657.835
15	32.40	Kasha	(Hissatsu: Kaiten,)	548.400	True	True	True	1	35	8206.235
16	34.56	Midare Setsugekka		1416.420	True	True	True	1	25	9622.655
17	36.72	Gekko	(Hissatsu: Shinten,)	931.350	True	True	True	1	25	10554.005
18	38.88	Hakaze		229.275	True	True	True	1	10	10783.280
19	41.04	Yukikaze		471.810	True	True	True	1	15	11255.090
20	43.20	Hakaze		229.275	True	True	True	1	25	11484.365
21	45.36	Shifu		395.220	True	False	True	1	30	11879.585
22	47.52	Kasha	(Hissatsu: Shinten,)	931.350	True	True	True	1	35	12810.935
23	49.68	Hakaze		229.275	True	True	True	1	20	13040.210
24	51.84	Jinpu		395.220	True	True	True	1	25	13435.430
25	54.00	Gekko	(Hissatsu: Shinten,)	931.350	True	True	True	1	30	14366.780
26	56.16	Hakaze		229.275	True	True	True	1	15	14596.055
27	58.32	Shifu	(Hagakure,)	395.220	True	True	True	1	20	14991.275
28	60.48	Kasha	(Hissatsu: Kaiten,)	548.400	True	True	True	1	85	15539.675
29	62.64	Higanbana		548.400	True	True	True	1	75	16088.075

Ignoring Higanbana

We’ll start with the same initial conditions but this time decide against applying Higanbana.

sam = Samurai(base_gcd=2.40, kenki_mastery=2, kenki_gauge=20)

sam.has_jinpu = True
sam.has_shifu = True

actions = [('Hakaze'), ('Yukikaze', 'Hissatsu: Shinten'),
           ('Hakaze'), ('Jinpu'), ('Gekko', 'Hissatsu: Shinten'),
           ('Hakaze'), ('Shifu'), ('Kasha', 'Hagakure'),
           ('Hakaze', 'Hissatsu: Guren'), ('Jinpu', 'Hissatsu: Shinten'), ('Gekko', 'Meikyo Shisui'),
           ('Yukikaze', 'Hissatsu: Shinten'), ('Kasha', 'Hissatsu: Kaiten'), ('Midare Setsugekka'), ('Gekko'),
           ('Hakaze'), ('Shifu'), ('Kasha', 'Hissatsu: Shinten'),
           ('Hakaze'), ('Yukikaze'),
           ('Hakaze'), ('Jinpu'), ('Gekko', 'Hissatsu: Kaiten'), ('Midare Setsugekka'),
           ('Hakaze'), ('Shifu'), ('Kasha', 'Hissatsu: Shinten'),
           ('Hakaze'), ('Yukikaze', 'Hissatsu: Shinten'),
           ('Hakaze')]

df_higanbana_false, _, _ = sam.parse_rotation(actions)
display(df_higanbana_false)

average potency per GCD = 536.0723333333334
average potency per second = 248.181635802

	Time	Weaponskill	Abilities	Potency	Jinpu	Shifu	Yukikaze	Higanbana	Kenki	Total Potency
0	0.00	Hakaze		150.000	True	True	False	0	20	150.000
1	2.16	Yukikaze	(Hissatsu: Shinten,)	673.000	False	False	False	0	25	823.000
2	4.32	Hakaze		166.500	False	False	True	0	10	989.500
3	6.48	Jinpu		310.800	False	False	True	0	15	1300.300
4	8.64	Gekko	(Hissatsu: Shinten,)	893.550	True	False	True	0	20	2193.850
5	10.80	Hakaze		191.475	True	False	True	0	5	2385.325
6	12.96	Shifu		357.420	True	False	True	0	10	2742.745
7	15.12	Kasha	(Hagakure,)	510.600	True	True	True	0	15	3253.345
8	17.28	Hakaze	(Hissatsu: Guren,)	1212.675	True	True	True	0	85	4466.020
9	19.44	Jinpu	(Hissatsu: Shinten,)	740.370	True	True	True	0	40	5206.390
10	21.60	Gekko	(Meikyo Shisui,)	510.600	True	True	True	0	20	5716.990
11	23.76	Yukikaze	(Hissatsu: Shinten,)	816.960	True	True	True	0	30	6533.950
12	25.92	Kasha	(Hissatsu: Kaiten,)	510.600	True	True	True	0	15	7044.550
13	28.08	Midare Setsugekka		1378.620	True	True	True	0	5	8423.170
14	30.24	Gekko		510.600	True	True	True	0	5	8933.770
15	32.40	Hakaze		191.475	True	True	True	0	15	9125.245
16	34.56	Shifu		357.420	True	True	True	0	20	9482.665
17	36.72	Kasha	(Hissatsu: Shinten,)	893.550	True	True	True	0	25	10376.215
18	38.88	Hakaze		191.475	True	True	True	0	10	10567.690
19	41.04	Yukikaze		434.010	True	True	True	0	15	11001.700
20	43.20	Hakaze		191.475	True	True	True	0	25	11193.175
21	45.36	Jinpu		357.420	True	True	True	0	30	11550.595
22	47.52	Gekko	(Hissatsu: Kaiten,)	510.600	True	True	True	0	35	12061.195
23	49.68	Midare Setsugekka		1378.620	True	True	True	0	25	13439.815
24	51.84	Hakaze		191.475	True	True	True	0	25	13631.290
25	54.00	Shifu		357.420	True	True	True	0	30	13988.710
26	56.16	Kasha	(Hissatsu: Shinten,)	893.550	True	True	True	0	35	14882.260
27	58.32	Hakaze		191.475	True	True	True	0	20	15073.735
28	60.48	Yukikaze	(Hissatsu: Shinten,)	816.960	True	True	True	0	25	15890.695
29	62.64	Hakaze		191.475	True	True	True	0	10	16082.170

compare_potencies([df_higanbana_true, df_higanbana_false], ['Higanbana', 'No Higanbana'])

I applied the same policy to each scenario and limited the horizon to the duration of one Higanbana. I did not make a conscious effort to expend all resources (i.e., Sen and Kenki) at the end of this horizon so this is assuming that the encounter continues on. One small adjustment that could be made in each case is using Meikyo Shisui earlier, or even double weaving it with Hagakure if you are comfortable.

Here we see that the cumulative potency and resulting pps metrics end up being more or less equal. However, the final conditions differ in each case:

Applying Higanbana:

• 75 Kenki
• 0 Sen open
• Higanbana freshly applied

Ignoring Higanbana:

• 15 Kenki
• Setsu, Ka open

We see the immediate gains of ignoring Higanbana as that simulation takes an early lead for the first portion of the simulated duration. However, we observe the track with applied Higanbana catch up at around 34 seconds, or 30 seconds after the first Higanbana was applied. This is the point where we apply our first Midare Setsugekka to match the other policy.

Furthermore, the ending conditions of the policy with Higanbana put it favorably ahead in terms of potential damage. It ends with effectively 115 Kenki (if we did not do the second application of Higanbana) compared to the effective 55 Kenki of the other policy. (Converting Sen to Kenki here.)

Conclusions

I will go with the general recommendation of applying Higanbana if at least half of its DoT’s duration will tick and if it does not preclude you from getting a Midare Setsugekka off in that timespan. The scale tilts in favor of applying Higanbana if there is any forced downtime (but the target is still targetable) due to mechanics; the DoT pays off even if you’re not hitting a target.

One interesting observation is that even at the full duration of Higanbana, both policies are essentially equivalent in their cumulative potency outcome. If we know that we have about about a minute left in the encounter, we’ll have to adjust our decisions accordingly to avoid leaving resources on the table. For example, instead of using Hagakure and reapplying Higanbana, we probably should’ve just used Midare Setsugekka. Here’s how that would like:

sam = Samurai(base_gcd=2.40, kenki_mastery=2, kenki_gauge=20)

sam.has_jinpu = True
sam.has_shifu = True

actions = [('Hakaze'), ('Yukikaze', 'Hissatsu: Kaiten'), ('Higanbana'), 
           ('Hakaze'), ('Jinpu'), ('Gekko'),
           ('Hakaze'), ('Shifu'), ('Kasha', 'Hissatsu: Guren'),
           ('Hakaze'), ('Yukikaze', 'Hagakure'),
           ('Hakaze', 'Hissatsu: Shinten'), ('Jinpu', 'Hissatsu: Shinten'), ('Gekko', 'Meikyo Shisui'),
           ('Yukikaze', 'Hissatsu: Shinten'), ('Kasha', 'Hissatsu: Kaiten'), ('Midare Setsugekka'), ('Gekko', 'Hissatsu: Shinten'),
           ('Hakaze'), ('Yukikaze'),
           ('Hakaze'), ('Shifu'), ('Kasha', 'Hissatsu: Shinten'),
           ('Hakaze'), ('Jinpu'), ('Gekko', 'Hissatsu: Shinten'),
           ('Hakaze'), ('Shifu', 'Hissatsu: Kaiten'), ('Midare Setsugekka'), ('Kasha', 'Hissatsu: Gyoten')]

df_higanbana_true2, _, _ = sam.parse_rotation(actions)
#display(df_higanbana_true2)

average potency per GCD = 565.2031666666666
average potency per second = 261.668132716

	Time	Weaponskill	Abilities	Potency	Jinpu	Shifu	Yukikaze	Higanbana	Kenki	Total Potency
0	0.00	Hakaze		150.000	True	True	False	0	20	150.000
1	2.16	Yukikaze	(Hissatsu: Kaiten,)	340.000	False	False	False	0	25	490.000
2	4.32	Higanbana		377.800	False	False	True	0	15	867.800
3	6.48	Hakaze		204.300	False	False	True	1	15	1072.100
4	8.64	Jinpu		348.600	False	False	True	1	20	1420.700
5	10.80	Gekko		548.400	True	False	True	1	25	1969.100
6	12.96	Hakaze		229.275	True	False	True	1	35	2198.375
7	15.12	Shifu		395.220	True	False	True	1	40	2593.595
8	17.28	Kasha	(Hissatsu: Guren,)	1569.600	True	True	True	1	45	4163.195
9	19.44	Hakaze		229.275	True	True	True	1	5	4392.470
10	21.60	Yukikaze	(Hagakure,)	471.810	True	True	True	1	10	4864.280
11	23.76	Hakaze	(Hissatsu: Shinten,)	612.225	True	True	True	1	80	5476.505
12	25.92	Jinpu	(Hissatsu: Shinten,)	778.170	True	True	True	1	60	6254.675
13	28.08	Gekko	(Meikyo Shisui,)	548.400	True	True	True	1	40	6803.075
14	30.24	Yukikaze	(Hissatsu: Shinten,)	854.760	True	True	True	1	50	7657.835
15	32.40	Kasha	(Hissatsu: Kaiten,)	548.400	True	True	True	1	35	8206.235
16	34.56	Midare Setsugekka		1416.420	True	True	True	1	25	9622.655
17	36.72	Gekko	(Hissatsu: Shinten,)	931.350	True	True	True	1	25	10554.005
18	38.88	Hakaze		229.275	True	True	True	1	10	10783.280
19	41.04	Yukikaze		471.810	True	True	True	1	15	11255.090
20	43.20	Hakaze		229.275	True	True	True	1	25	11484.365
21	45.36	Shifu		395.220	True	False	True	1	30	11879.585
22	47.52	Kasha	(Hissatsu: Shinten,)	931.350	True	True	True	1	35	12810.935
23	49.68	Hakaze		229.275	True	True	True	1	20	13040.210
24	51.84	Jinpu		395.220	True	True	True	1	25	13435.430
25	54.00	Gekko	(Hissatsu: Shinten,)	931.350	True	True	True	1	30	14366.780
26	56.16	Hakaze		229.275	True	True	True	1	15	14596.055
27	58.32	Shifu	(Hissatsu: Kaiten,)	395.220	True	True	True	1	20	14991.275
28	60.48	Midare Setsugekka		1416.420	True	True	True	1	5	16407.695
29	62.64	Kasha		548.400	True	True	True	1	5	16956.095

compare_n_potencies([df_higanbana_true, df_higanbana_false, df_higanbana_true2], ['2 Higanbana', 'No Higanbana', '1 Higanbana'])

That’s a nontrivial increase in damage. This is just one of many examples of being cognizant of encounter duration when planning your resource usage.

back to the top…