I responded to a thread about Recoil/Spread, this is a great one to talk about movement accuracy:
Running Accuracy:
Our intention is that most weapons (shotguns excluded) are not reliable when shooting in a full movement state. If this becomes an effective or reliable way to shoot the weapons then we will definitely look to correct it.
Also, distinguishing between an opponent hitting a headshot on you while in a full run state, versus when they've come to a stop on their side but still look like they're moving for you. I've observed many cases where it's interpreted that the opponent shot you while moving, but they were already within an accurate standing state on their client. This is due to networking where you haven't quite got the information about the opponent coming to a stop, but they've already shot and killed you. We're going to explore opportunities to tighten this up and reduce the occurrence of this.
We'll continue to watch the movement accuracy tuning and feedback, please look out for scenarios where you think the opponent is just straight up running and hitting lucky shots, versus when they've likely to have had come to a stop on their screen and shot you.
Walking Accuracy:
Currently, you're able to shoot 'accurately' while walking (it's still more spread and recoil compared to standing still or crouching). Our thought here is that if you're taking the time to move methodically and pre-aim properly, you should be able to shoot a few accurate shots without needing to come to a full stop when walking.
Additionally, peeking an angle by walking should not be a high success chance strategy against an opponent holding the angle as they get a window of time where they'll see you first. Shooting an opponent who is walking will also tag them to a point where they're nearly rooted, and the movement in that case should be fairly irrelevant.
So in theory walk accuracy can be helpful for you when approach an angle and responding quickly if an opponent swings on you. It should not be very helpful to peek with or utilize in combat. This is our currently theory with walk shooting, and we'd like to observe over time how the behaviors pan out.
MB1 + Walk:
There has also been a theory floating around that if you bind mouse1 to walk you can immediately go into a walk accuracy state from a full run. Excal (Designer on the team) did some testing on this (All tests done with the Phantom):
Standard Counter-Strafing:
Run -> Walk (61%): 0.055s
Run -> DeadZone (25%): 0.104s
Run -> Stop (0%): 0.160s
Mouse1 Walk (No Counter-Strafe):
Run -> Walk (61%): 0.065s
Run -> DeadZone (25%): 0.110s
Run -> Stop (0%): 0.170s
Mouse1 Walk (Counter-Strafe):
Run -> Walk (61%): 0.060s
Run -> DeadZone (25%): 0.107s
Run -> Stop (0%): 0.166s
These results aren't going to be perfectly accurate due to inconsistent frame rates and this is all done with human input, but the gap is large enough that it's safe to say stopping by letting off of movement inputs or counter-strafing is faster than using walk to transition into an accurate state.
You will also fire your first bullet in the running state during the transition with this bind, so not only is it slower but you're losing your most accurate bullet when you begin firing as well.
As mentioned in the other thread, keep this context in mind when you encounter some of these scenarios. If they're not playing out as intended or you don't agree with the direction please let us know! This information helps inform how we tune the game over time.
Thanks for taking the time to read this wall of text!
devtrackers.gg