Talk:Blink Arrow

Projectile Speed and its effect on travel speed

Back then I did tests, concluding that proj speed does nothing, then someone else corrected it to say proj speed does something.

To settle this, I did some more tests, and I'm documenting the results. This time I did a video recording and counted frames, which is a bit more accurate than the stopwatch approach I used in my previous tests - especially since I stopped based on the arrow landing, which turns out to be a meaningless animation effect. I counted the time from the first frame of the skill button greying out (aka start of the cooldown) to the first frame where the character (and the camera) arrives at the destination. Inaccuracies happen because I'm recording a 30 fps game on a 100 Hz monitor at 60 fps.

Picked a random character with no proj or attack speed, equipped a random bow from the vendor, used level 1 gems. Repeated each test 3 times.

• No link: 53, 53, 53 frames
• Level 1 Faster Projectile Support (50% inc proj speed): 46, 45, 46 frames -> 7 frames faster
• Level 1 Slower Projectile Support (20% less proj speed): 57, 57, 57 frames -> 4 frames slower

tl;dr: only a part of the delay is affected by projectile speed modifiers.

Repeated the same with a bunch of attack speed bonuses:

• No link: 37, 38, 37 frames
• Level 1 Faster Projectile Support (50% inc proj speed): 34, 34, 33 frames -> 3 frames faster
• Level 1 Slower Projectile Support (20% less proj speed): 40, 39, 39 frames -> 2 frames slower

tl;dr: attack speed modifiers affect the whole delay. Let's get a bit more precise and do two tests with the same attack speed, but different weapon base speeds:

• Bone Bow (1.2 aps), 35% global IAS for 1.62 aps: 42, 42, 41 frames
• Short Bow (1.5 aps) with 8% local IAS for 1.62 aps: 52, 52, 52 frames

tl;dr: a part of the delay is affected by global IAS, but not by attack speed (i.e. not the weapon base speed). Let's see if weapon base speed matters at all:

• Same Short Bow (1.62 aps), 35% global IAS for 2.187 aps: 37, 38, 38 frames

tl;dr: a part of the delay is affected by weapon base speeds

This is all consistent with the idea of separate stages: a regular attack animation, modified by weapon base speed and global IAS as usual, followed by a delay, which is affected by both proj speed and global IAS, but not by weapon base speeds.

Unfortunately, I'm unable to pin down the exact duration of the second stage.

• If +50% proj speed saves 7 frames, then the base should be 21 frames
• If -20% proj speed costs 4 frames, then the base should be 16 frames
• If we have 1.62 aps (37 frames for the attack) and need a total of 52 frames, then the base should be 15 frames

I'll adjust my previous ~200ms estimate, but it must remain an estimate for now.

WobblyOkapi (talk) 11:14, 8 August 2023 (UTC)