Strafe is an attack sequence which always consists of at least two missiles, whose length is the sum of the First and Last Frame sub-columns in the tables below; lengths added by additional missiles (due to additional targets or more minimum arrows) are in the Follow-up Frame sub-column, from left to right (the Last Frame sub-column is normally the combined length of the follow-up interval between the last two missiles and the end of the sequence animation).
For example, when two arrows are fired by a bow with 0 EIAS (6|3|10), sequence length is 16 (6+10) frames (the first arrow is fired in frame 6, the second 3 frames later and the remainder of the animation is 7 frames); when 10 arrows are fired, sequence length is 40 (6 + (3*8) + 10) frames. Average attack rate against dispersed targets is sequence length/missiles, so the two arrow sequence averages 8 (16/2) frames per arrow while the 10 arrow sequence averages 4 (40/10) frames per arrow.
Strafe has a 4 frame next delay, which prevents its missiles colliding with a single target more than once every 5 frames (this does not affect multiple dispersed targets since next delay will expire even when alternating between just two targets, although it may affect multiple targets when consecutive missiles have to pass through one target to reach another). This does not mean that once the follow-up interval between missiles is less than 5 frames, only alternating arrows will collide with the target: although arrows and bolts normally only make a single collision check, and this check cannot occur while next delay is active, it normally takes time for them to pass through a target and next delay can expire while this is happening.
The time it takes for an arrow or bolt (they are the same speed) to pass through a stationary target depends on its size: it always takes 3 frames to pass through a size 3 target and a single frame to pass through a size 1 target, while it takes 1-3 frames to pass through a size 2 target depending on its position.
Arrows or bolts take 3 frames to pass through a stationary size 2 target centered on dark green sub-tiles, 2 frames to pass through one centered on green sub-tiles and 1 frame to pass through one centered on empty sub-tiles. Effective attack rate against a single target is always maximised at 3 frames and minimised at 1 frame, so position your Amazon accordingly.
Effective attack rate is maximised by ensuring it takes as long as possible for each missile to pass through the target, allowing the maximum time possible for next delay activated by a previous missile to expire: attack sequences are a fixed length, so maximising collisions per sequence results in more collisions and thus more hits and damage over time. Since the vast majority of targets are size 2-3, what follows assumes it takes 3 frames for an arrow to pass through a single stationary target: black cells in the Missile row indicate frames in which no missile is passing through a target, black cells in the Collision row indicate frames in which next delay is active.
4 frame follow-up
Frame 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37
Missile 1 2 3 4 5 6 7 8 9 10
Collision 1 2 3 4 5 6 7 8
Thus when 10 missiles are fired at a single stationary target, only 8 (missiles 1, 2, 3, 5, 6, 7, 9 and 10) can actually collide with it.
3 frame follow-up
Frame 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
Missile 1 2 3 4 5 6 7 8 9 10
Collision 1 2 3 4 5 6
Thus when 10 missiles are fired at a single stationary target, only 6 (missiles 1, 2, 4, 6, 7 and 9) can actually collide with it. Whether or not attack rate improves with fewer collisions in less time varies depending on the EIAS breakpoint and number of missiles fired.
A further complication is introduced by a bug resulting in discrepancy in length between the sequence displayed by the client and that actually executed on the server. When the sequence is shorter on the client, when a single sequence ends the Amazon is unable to move or execute a different attack until the slower server sequence has ended. However, when consecutive sequences are fired, the server sequence is interrupted when the client sequence ends, and the next sequence begins on both the client and the server.
For example, a bow with 50 EIAS fires at 4|2|7 on the client and 4|3|7 on the server, resulting in a 10 arrow sequence lasting 27 (4 + (2*8) + 7) frames on the client and 35 (4 + (3*8) + 7) frames on the server. So after a single sequence, the Amazon will appear to have stopped firing after 27 frames but she will be unable to do anything else until the sequence actually finishes 8 frames later. Consecutive sequences will result in the server sequence being interrupted after 27 frames:
Arrow 1 2 3 4 5 6 7 8 9 10
Frame 4 7 10 13 16 19 22 25 28 31
The server sequence is interrupted before the last two arrows can be fired, resulting in an average of 3.375 (27/8) frames per arrow when fired at multiple dispersed targets; without the interruption, the average would be 3.5 (35/10). When firing at a single stationary target, only 5 of those 8 arrows (1, 2, 4, 6 and 7) can collide with it due to next delay, resulting in an average of 5.4 (27/5) frames per arrow; without the interruption, the average would be ~5.83 (35/6).
BOW
Frames/Attack EIAS IAS
First Follow-up Last 10 5 0 -10
7 4 4 4 4 4 4 4 4 12 -14
11 -8 3
6 3 3 3 3 3 3 3 3 10 0 11 6
9 11 26 19 13 2
5 21 42 34 26 13
8 23 46 37 29 15
7 38 80 68 56 37
4 2 2 2 2 2 2 2 2 50 120 102 86 60
6 58 157 133 113 80
50 and 58 EIAS breakpoints only apply to the displayed animation for follow-up attacks: they actually continue to be 3 frames apart, and this interval is never reduced to 2 frames. When Strafe is fired continuously, the next volley begins when the animation of the previous one finishes, which reduces the number of actual attacks per volley but may sometimes increase the number of actual attacks over time, resulting in a faster attack rate. [3]
Single target
The table below lists the average frames per arrow when firing continuously at a single target.
EIAS Frames/Attack Arrows
Server Client 2 3 4 5 6 7 8 9 10
-14 7|4|12 7|4|12 9.5 (19/2) 7.6 (23/3) 9 (27/3) 7.7 (31/4) 7 (35/5) 6.5 (39/6) 7.1 (43/6) 6.7 (47/7) 6.3 (51/8)
-8 7|4|11 7|4|11 9 (18/2) 7.3 (22/3) 8.6 (26/3) 7.5 (30/4) 6.8 (34/5) 6.3 (38/6) 7 (42/6) 6.5 (46/7) 6.2 (50/8)
0 6|3|10 6|3|10 8 (16/2) 9.5 (19/2) 7.3 (22/3) 8.3 (25/3) 7 (28/4) 6.2 (31/5) 6.8 (34/5) 6.1 (37/6) 6.6 (40/6)
11 6|3|9 6|3|9 7.5 (15/2) 9 (18/2) 7 (21/3) 8 (24/3) 6.7 (27/4) 6 (30/5) 6.6 (33/5) 6 (36/6) 6.5 (39/6)
21 5|3|9 5|3|9 7 (14/2) 8.5 (17/2) 6.6 (20/3) 7.6 (23/3) 6.5 (26/4) 5.8 (29/5) 6.4 (32/5) 5.8 (35/6) 6.3 (38/6)
23 5|3|8 5|3|8 6.5 (13/2) 8 (16/2) 6.3 (19/3) 7.3 (22/3) 6.2 (25/4) 5.6 (28/5) 6.2 (31/5) 5.6 (34/6) 6.1 (37/6)
38 5|3|7 5|3|7 6 (12/2) 7.5 (15/2) 6 (18/3) 7 (21/3) 6 (24/4) 5.4 (27/5) 6 (30/5) 5.5 (33/6) 6 (36/6)
50 4|3|7 4|2|7 5.5 (11/2) 6.5 (13/2) 5 (15/3) 5.6 (17/3) 5.0 (76/15) 5.2 (21/4) 5.1 (46/9) 5 (25/5) 5.4 (27/5)
58 4|3|6 4|2|6 5 (10/2) 6 (12/2) 5.2 (42/8) 5.3 (16/3) 6 (18/3) 5.2 (21/4) 5.0 (66/13) 5.3 (48/9) 5.2 (26/5)
[4]
Sequence length and thus average frames per arrow was found experimentally to be the same at the 50 and 58 EIAS breakpoint for 7 arrows (level 21-23) [5]
Multiple targets
The table below lists the average frames per arrow when firing continuously at multiple targets which are spread out so that no arrow has to pass through another target (thus even with two targets there's an interval during which the 4 frame next delay expires).
EIAS Frames/Attack Arrows
Server Client 2 3 4 5 6 7 8 9 10
-14 7|4|12 7|4|12 9.5 (19/2) 7.6 (23/3) 6.7 (27/4) 6.2 (31/5) 5.8 (35/6) 5.5 (39/7) 5.3 (43/8) 5.2 (47/9) 5.1 (51/10)
-8 7|4|11 7|4|11 9 (18/2) 7.3 (22/3) 6.5 (26/4) 6 (30/5) 5.6 (34/6) 5.4 (38/7) 5.2 (42/8) 5.1 (46/9) 5 (50/10)
0 6|3|10 6|3|10 8 (16/2) 6.3 (19/3) 5.5 (22/4) 5 (25/5) 4.6 (28/6) 4.4 (31/7) 4.2 (34/8) 4.1 (37/9) 4 (40/10)
11 6|3|9 6|3|9 7.5 (15/2) 6 (18/3) 5.2 (21/4) 4.8 (24/5) 4.5 (27/6) 4.2 (30/7) 4.1 (33/8) 4 (36/9) 3.9 (39/10)
21 5|3|9 5|3|9 7 (14/2) 5.6 (17/3) 5 (20/4) 4.6 (23/5) 4.3 (26/6) 4.1 (29/7) 4 (32/8) 3.8 (35/9) 3.8 (38/10)
23 5|3|8 5|3|8 6.5 (13/2) 5.3 (16/3) 4.7 (19/4) 4.4 (22/5) 4.1 (25/6) 4 (28/7) 3.8 (31/8) 3.7 (34/9) 3.7 (37/10)
38 5|3|7 5|3|7 6 (12/2) 5 (15/3) 4.5 (18/4) 4.2 (21/5) 4 (24/6) 3.8 (27/7) 3.7 (30/8) 3.6 (33/9) 3.6 (36/10)
50 4|3|7 4|2|7 5.5 (11/2) 4.3 (13/3) 3.7 (15/4) 3.4 (17/5) 3.1 (19/6) 3.5 (21/6) 3.2 (23/7) 3.1 (25/8) 3.3 (27/8)
58 4|3|6 4|2|6 5 (10/2) 4 (12/3) 3.5 (14/4) 3.2 (16/5) 3.6 (18/5) 3.5 (21/6) 3.1 (22/7) 3.4 (24/7) 3.2 (26/8)
[6]
Sequence length and thus average frames per arrow was found experimentally to be the same at the 50 and 58 EIAS breakpoint for 7 arrows (level 21-23) [7]
XBOW
Frames/Attack EIAS IAS
First Follow-up Last 10 0 -10 -40 -60
11 6 6 6 6 6 6 6 6 17 -10
10 -9 2
16 -5 6
9 5 5 5 5 5 5 5 5 15 0 11
14 8 22 9
4 5 5 5 5 5 5 5 12 27 14 3
8 5 5 5 5 5 5 5 5 13 29 15 4
13 15 32 18 6
5 4 5 4 5 4 5 4 21 42 26 13
12 24 48 30 16
4 4 4 4 4 4 4 4 11 25 50 32 18
7 12 29 58 39 23
11 34 70 48 30
3 4 4 4 4 4 4 4 43 95 68 46 4
10 46 105 75 52 7
6 4 3 4 3 4 3 4 3 50 120 86 60 11
9 61 174 125 89 26 2
3 3 3 3 3 3 3 3 8 67 215 152 109 35 8
>49 EIAS breakpoints only apply to the displayed animation for follow-up attacks: they actually continue to be 4 frames apart, and this interval is never reduced to 3 frames. When Strafe is fired continuously, the next volley begins when the animation of the previous one finishes, which reduces the number of actual attacks per volley but may sometimes increase the number of actual attacks over time, resulting in a faster attack rate. There are other breakpoints which only apply to the displayed animation rather than actual attacks as well. [8]
Single target
The table below lists the average frames per bolt when firing continuously at a single target.
EIAS Frames/Attack Arrows
Server Client 2 3 4 5 6 7 8 9 10
-10 11|6|6|17 11|6|6|17 14 (28/2) 11.3 (34/3) 10 (40/4) 9.2 (46/5) 8.6 (52/6) 8.2 (58/7) 8 (64/8) 7.7 (70/9) 5.6 (76/10)
-9 10|6|6|17 10|6|6|17 13.5 (27/2) 11 (33/3) 9.7 (39/4) 9 (45/5) 8.5 (51/6) 8.1 (57/7) 7.8 (63/8) 7.6 (69/9) 7.5 (75/10)
-5 10|6|6|16 10|6|6|16 13 (26/2) 10.6 (32/3) 9.5 (38/4) 8.8 (44/5) 8.3 (50/6) 8 (56/7) 7.7 (62/8) 7.5 (68/9) 7.4 (74/10)
0 9|5|5|15 9|5|5|15 12 (24/2) 9.6 (29/3) 8.5 (34/4) 7.8 (39/5) 7.3 (44/6) 7 (49/7) 6.7 (54/8) 6.5 (59/9) 6.4 (64/10)
8 9|5|5|14 9|5|5|14 11.5 (23/2) 9.3 (28/3) 8.2 (33/4) 7.6 (38/5) 7.1 (43/6) 6.8 (48/7) 6.6 (53/8) 6.4 (58/9) 6.3 (63/10)
12 9|4|5|14 9|4|5|14 11.5 (23/2) 9 (27/3) 8 (32/4) 7.4 (37/5) 7 (42/6) 6.7 (47/7) 6.5 (52/8) 6.3 (57/9) 6.2 (62/10)
13 8|5|5|14 8|5|5|14 11 (22/2) 9 (27/3) 8 (32/4) 7.4 (37/5) 7 (42/6) 6.7 (47/7) 6.5 (52/8) 6.3 (57/9) 6.2 (62/10)
15 8|5|5|13 8|5|5|13 10.5 (21/2) 8.6 (26/3) 7.7 (31/4) 7.2 (36/5) 6.8 (41/6) 6.5 (46/7) 6.3 (51/8) 6.2 (56/9) 6.1 (61/10)
21 8|5|5|13 8|5|4|13 10.5 (21/2) 8.6 (26/3) 7.5 (30/4) 7 (35/5) 6.5 (39/6) 6.2 (44/7) 6 (48/8) 5.8 (53/9) 5.7 (57/10)
24 8|5|5|12 8|5|4|12 10 (20/2) 8.3 (25/3) 7.2 (29/4) 6.8 (34/5) 6.3 (38/6) 6.1 (43/7) 5.8 (47/8) 5.7 (52/9) 5.6 (56/10)
25 8|4|4|11 8|4|4|11 9.5 (19/2) 7.6 (23/3) 9 (27/3) 7.7 (31/4) 7 (35/5) 6.5 (39/6) 7.1 (43/6) 6.7 (47/7) 6.3 (51/8)
29 7|4|4|12 7|4|4|12 9.5 (19/2) 7.6 (23/3) 9 (27/3) 7.7 (31/4) 7 (35/5) 6.5 (39/6) 7.1 (43/6) 6.7 (47/7) 6.3 (51/8)
34 7|4|4|11 7|4|4|11 9 (18/2) 7.3 (22/3) 8.6 (26/3) 7.5 (30/4) 6.8 (34/5) 6.3 (38/6) 7 (42/6) 6.5 (46/7) 6.2 (50/8)
43 7|3|4|11 7|3|4|11 9 (18/2) 7 (21/3) 8.3 (25/3) 7.2 (29/4) 6.6 (33/5) 6.1 (37/6) 6.8 (41/6) 6.4 (45/7) 6.1 (49/8)
46 7|3|4|10 7|3|4|10 8.5 (17/2) 6.6 (20/3) 8 (24/3) 7 (28/4) 6.4 (32/5) 6 (36/6) 6.6 (40/6) 6.2 (44/7) 6 (48/8)
50 6|4|4|10 6|4|3|10 8 (16/2) 6.6 (20/3) 7.6 (23/3) 6.7 (27/4) 6 (30/5) 5.6 (34/6) 6.1 (37/6) 5.8 (41/7) 5.5 (44/8)
61 6|4|4|9 6|4|3|9 7.5 (15/2) 6.3 (19/3) 7.3 (22/3) 6.5 (26/4) 5.8 (29/5) 5.5 (33/6) 6 (36/6) 5.7 (40/7) 5.3 (43/8)
67 6|4|4|8 6|3|3|8 7 (14/2) 5.6 (17/3) 6.6 (20/3) 5.7 (23/4) 5.2 (26/5) 5.8 (29/5) 5.3 (32/6) 5.8 (35/6) 5.4 (38/7)
[9]
Multiple targets
The table below lists the average frames per bolt when firing continuously at multiple targets which are spread out so that no bolt has to pass through another target (thus even with two targets there's an interval during which the 4 frame next delay expires).
EIAS Frames/Attack Arrows
Server Client 2 3 4 5 6 7 8 9 10
-10 11|6|6|17 11|6|6|17 14 (28/2) 11.3 (34/3) 10 (40/4) 9.2 (46/5) 8.6 (52/6) 8.2 (58/7) 8 (64/8) 7.7 (70/9) 7.6 (76/10)
-9 10|6|6|17 10|6|6|17 13.5 (27/2) 11 (33/3) 9.7 (39/4) 9 (45/5) 8.5 (51/6) 8.1 (57/7) 7.8 (63/8) 7.6 (69/9) 7.5 (75/10)
-5 10|6|6|16 10|6|6|16 13 (26/2) 10.6 (32/3) 9.5 (38/4) 8.8 (44/5) 8.3 (50/6) 8 (56/7) 7.7 (62/8) 7.5 (68/9) 7.4 (74/10)
0 9|5|5|15 9|5|5|15 12 (24/2) 9.6 (29/3) 8.5 (34/4) 7.8 (39/5) 7.3 (44/6) 7 (49/7) 6.7 (54/8) 6.5 (59/9) 6.4 (64/10)
8 9|5|5|14 9|5|5|14 11.5 (23/2) 9.3 (28/3) 8.2 (33/4) 7.6 (38/5) 7.1 (43/6) 6.8 (48/7) 6.6 (53/8) 6.4 (58/9) 6.3 (63/10)
12 9|4|5|14 9|4|5|14 11.5 (23/2) 9 (27/3) 8 (32/4) 7.4 (37/5) 7 (42/6) 6.7 (47/7) 6.5 (52/8) 6.3 (57/9) 6.2 (62/10)
13 8|5|5|14 8|5|5|14 11 (22/2) 9 (27/3) 8 (32/4) 7.4 (37/5) 7 (42/6) 6.7 (47/7) 6.5 (52/8) 6.3 (57/9) 6.2 (62/10)
15 8|5|5|13 8|5|5|13 10.5 (21/2) 8.6 (26/3) 7.7 (31/4) 7.2 (36/5) 6.8 (41/6) 6.5 (46/7) 6.3 (51/8) 6.2 (56/9) 6.1 (61/10)
21 8|5|5|13 8|5|4|13 10.5 (21/2) 8.6 (26/3) 7.5 (30/4) 7 (35/5) 6.5 (39/6) 6.2 (44/7) 6 (48/8) 5.8 (53/9) 5.7 (57/10)
24 8|5|5|12 8|5|4|12 10 (20/2) 8.3 (25/3) 7.2 (29/4) 6.8 (34/5) 6.3 (38/6) 6.1 (43/7) 5.8 (47/8) 5.7 (52/9) 5.6 (56/10)
25 8|4|4|11 8|4|4|11 9.5 (19/2) 7.6 (23/3) 6.7 (27/4) 6.2 (31/5) 5.8 (35/6) 5.5 (39/7) 5.3 (43/8) 5.2 (47/9) 5.1 (51/10)
29 7|4|4|12 7|4|4|12 9.5 (19/2) 7.6 (23/3) 6.7 (27/4) 6.2 (31/5) 5.8 (35/6) 5.5 (39/7) 5.3 (43/8) 5.2 (47/9) 5.1 (51/10)
34 7|4|4|11 7|4|4|11 9 (18/2) 7.3 (22/3) 6.5 (26/4) 6 (30/5) 5.6 (34/6) 5.4 (38/7) 5.2 (42/8) 5.1 (46/9) 5 (50/10)
43 7|3|4|11 7|3|4|11 9 (18/2) 7 (21/3) 6.2 (25/4) 5.8 (29/5) 5.5 (33/6) 5.2 (37/7) 5.1 (41/8) 5 (45/9) 4.9 (49/10)
46 7|3|4|10 7|3|4|10 8.5 (17/2) 6.6 (20/3) 6 (24/4) 5.6 (28/5) 5.3 (32/6) 5.1 (36/7) 5 (40/8) 4.8 (44/9) 4.8 (48/10)
50 6|4|4|10 6|4|3|10 8 (16/2) 6.6 (20/3) 5.7 (23/4) 5.4 (27/5) 5 (30/6) 4.8 (34/7) 4.6 (37/8) 4.5 (41/9) 4.4 (44/10)
61 6|4|4|9 6|4|3|9 7.5 (15/2) 6.3 (19/3) 5.5 (22/4) 5.2 (26/5) 4.8 (29/6) 4.7 (33/7) 4.5 (36/8) 4.4 (40/9) 4.3 (43/10)
67 6|4|4|8 6|3|3|8 7 (14/2) 5.6 (17/3) 5 (20/4) 4.6 (23/5) 4.3 (26/6) 4.8 (29/6) 4.5 (32/7) 4.3 (35/8) 4.2 (38/9)
D2:R