Forums » Suggestions
Capship turret accuracy calculation
TL:DR; make turret inaccuracy less precise/repeatable but make it so it can actually hit something from time to time instead of being a sure "miss".
Currently capship turret accuracy is somewhat binary in the sense that turrets seem to operate in two modes mode 1 consists of lazering the target with perfect accuracy that doesn't take into account acceleration or curved movement. (this is fine)
The other mode however is incredibly infuriating as it just offsets the projectile's direction by roughly 15~30 degrees off to the side (the real problem is that it's always the same degree of inaccuracy, essentially the inaccurate mode is way too repeatable for something that's not supposed to be accurate) making the turret in question miss absolutely everything when it enters this "mode" when you've been shooting at a connie or a HAC from 1000m and every single shot misses it gets... a little annoying.
The problem with this second mode is that the inaccuracy doesn't change it's as if my turret was shooting with a crooked barrel and the turrets AI's is too stupid to compensate on the following shot it gets annoying when can happen constantly when sitting still at closer range.
would be possible to make the inaccurate mode function with a point spread function the magnitude of the application of said said inaccuracy would be affected by two factors, the size of the targeted ship: small, medium, large, capital and the relative velocities between the ships. a relative velocity of zero should always result in the turret shooting with pinpoint accuracy (If a dent manages to strafe with a centurion the centurion's got other problems type of deal)
if implementing truly "random" inaccuracy is too much of a pita using a similar spread pattern to the queen gat would be the next best thing for the aformentioned "mode 2".
Currently capship turret accuracy is somewhat binary in the sense that turrets seem to operate in two modes mode 1 consists of lazering the target with perfect accuracy that doesn't take into account acceleration or curved movement. (this is fine)
The other mode however is incredibly infuriating as it just offsets the projectile's direction by roughly 15~30 degrees off to the side (the real problem is that it's always the same degree of inaccuracy, essentially the inaccurate mode is way too repeatable for something that's not supposed to be accurate) making the turret in question miss absolutely everything when it enters this "mode" when you've been shooting at a connie or a HAC from 1000m and every single shot misses it gets... a little annoying.
The problem with this second mode is that the inaccuracy doesn't change it's as if my turret was shooting with a crooked barrel and the turrets AI's is too stupid to compensate on the following shot it gets annoying when can happen constantly when sitting still at closer range.
would be possible to make the inaccurate mode function with a point spread function the magnitude of the application of said said inaccuracy would be affected by two factors, the size of the targeted ship: small, medium, large, capital and the relative velocities between the ships. a relative velocity of zero should always result in the turret shooting with pinpoint accuracy (If a dent manages to strafe with a centurion the centurion's got other problems type of deal)
if implementing truly "random" inaccuracy is too much of a pita using a similar spread pattern to the queen gat would be the next best thing for the aformentioned "mode 2".
I don't think there ARE different modes, its just some funkiness in how turret targetting works between two moving targets. Remind me tonight to make a visual demonstration and i'll see if I cant record what you're talking about - i've seen it.
It "feels" as if there's these two modes.
Really it's just that sometimes the turret's accurate and the rest of the time it's shooting off 15 degrees to the side constantly.
Really it's just that sometimes the turret's accurate and the rest of the time it's shooting off 15 degrees to the side constantly.
This is probably a bug. Probably some kind of degenerate math error, or a "bad" calculation of a relative velocity due to a jump in clock timing, or something along those lines. I highly doubt that there are two "modes", at least not in any way that was "intentionally implemented".
All of this basically means we would need a lot of detail about how to exactly reproduce the issue.
Like: specific ship types, angles, relative speeds, turrets in usage, /ping at the time, probability of the issue being repeatable at the time, etc. Some kind of clear-cut case where it's easy to demonstrate the issue.
Specific reproduction cases are super helpful. Videos are not very helpful. I already "believe" that there's an issue, but we need specific instructions to help see it in our own development clients.
All of this basically means we would need a lot of detail about how to exactly reproduce the issue.
Like: specific ship types, angles, relative speeds, turrets in usage, /ping at the time, probability of the issue being repeatable at the time, etc. Some kind of clear-cut case where it's easy to demonstrate the issue.
Specific reproduction cases are super helpful. Videos are not very helpful. I already "believe" that there's an issue, but we need specific instructions to help see it in our own development clients.
Argh i know there's no such thing as accuracy modes it's just it feels like the turrets work in two states.
I should be able to reproduce it in a sterile environment.
I should be able to reproduce it in a sterile environment.
After spending more time in Deneb on a slightly higher latency connection with more packetloss/out of order packets coming in/out I have a feeling this is potentially brought on by network issues, it seems to to mostly happen when relative speeds are high don't hold off on looking into it until I have reproduced it between player capships in "sterile" conditions with some footage to boot; with a bit of luck that'll make it easier to figure out.