https://cdn.discordapp.com/attachments/703607334886309979/855704499871809546/n.png

I mean if he does it even though I plan to quit tf2 after trolling main hl next season I'd probably toss a hundred bucks in the prize pool. I don't have the organizational skills that would help, as I just do tech support, but what I do have is more disposable income than sense. I have thousands of dollars of peripherals in my room, wouldn't be too bad to throw some money at something which can do some real good.

duck

good to know, but even without an anticheat esea was overall better at dealing with cheaters. Maybe their human level anticheat team was more worthwhile.

edit: me brain damage have

wasn't there originally some promise that RGL would give a fuck about competitive integrity and handle sandbagging better than esea or ugc and then literally nothing happened with that and we have cheaters taking even longer to get banned than in esea because there is no anticheat and RGL is unwilling to force server providers to install even a reporting-only anticheat feeding into a discord channel as a suspicious tick farm so that evidence can be gathered efficiently and actionably and silently even without reports to be "rgl-approved" for matches?
Even if RGL isn't willing to make good decisions, listen to feedback, have reasonable rules, not fuck over the best casters in the scene like they did to Bowl for HL and Aim and probably others we didn't hear so much about, and aren't willing to be consistent on handling sandbaggers couldn't they at least try not to be a complete anticheat downgrade too?

Admins obviously can't follow their own rules with shit like Exa having been using metal footsteps and the whole situation where Dimento got clowned on in regards to River so that the ban was so delayed that they fucked over not just one playoffs, but two!

The only value proposition RGL really offers isn't even streams because the tftv guys are good too(still respect the dolphin though), but that a prize pool exists, and that doesn't matter to most people. Honestly, why even participate in RGL? If IM and Main and adv players quit en masse the league simply dies, and maybe another league will realize a prize pool is all it takes and crowdfund one.

To some extent I am tired out of my mind, my eyes are strained so I'm seeing words vertically double like shadows, and I am pretty sure I am 1am schizoposting, but chances are I said at least something worthwhile by accident here.

I'd rather not pay anything to rgl unless they plan to run the league sanely and we know that's not going to happen
probably just going back to pubbing after this season, but unironically if UGC was going to open up voluntary crowd funding of prize pools I'd probably be up to pay into UGC just to have something that isn't RGL.

You'll need to get MDL working on Linux to do that. The bundled with TF2 studiomdl.exe is why Yttrium's viewmodel hider wasn't expected to play nice on Linux either.

I can say, however, that this works via WINE although I'm not sure what specific requirements there are because I used the accursed hellscape of a wineprefix I have for running literally anything that will run that doesn't require hyperspecific requirements. I couldn't get the yttrium viewmodel hider to work, so that does make this interestingly different. It's probably because yttrium has a program separate from just using a simple batch file.

https://cdn.discordapp.com/attachments/192083556300750848/853655254927540234/2021-06-13-101834.png

Edit: forgot to mention that to get MDL I just force proton on tf2 and then switch back to no compatibility layer after I am done with MDL.

ok I just got off work and based on the syntax in https://github.com/n0la/rcon it could literally be as simple as this - https://pastebin.com/utVXZBRy
As you can probably see, it's literally just some number shifting so it's in a range of 1 to 4 sensitivity incrementing .5 at a time, for example. You can change that to some other more sane range if you'd like, and I haven't tested it to be sure rcon actually plays nice with this, but as I noted, just number shuffling and switching xinput out for rcon is all you'll need.

Also, the sleep changes how many seconds, so right now it's at 3-5 seconds per increment.

-usercon +rcon_password "whatever" +ip 127.0.0.1 +net_start

I forgot that was an option, but the big thing is I wanted it systemwide so I could use it in other games, kovaaks, etc. as well. In regards to doing it via rcon, note specifically that all of this goes to a singular input, "accel speed", so if you had a tolerable sensitivity range you could modify the sensitivity function to point your commands at your local rcon and simply adjust the number scales accordingly.

Essentially, 90% of the work is already done as most of what really needs changing is number scaling and the command/target input location.

the muscle memory meme is just that, a meme. Aim is almost entirely a factor of motor control, comfort, and being warm on the current sensitivity. A randomizer sacrifices warmth and some comfort to focus solely on improving the underlying motor control. In particular, the reason for the frequent but generally small sens bumps as currently designed is to keep the fine motor control as a focus while not changing from 6 inches to 18 inches in 1 second and simply being unreasonable.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4747782/ has some info, but iirc there were other studies done which showed greater overall task accuracy improvement overtime with frequent small adjustments as well.

I didn't entirely believe it at first either until I tried very rough prototype versions of what became this sens randomizer and simply left it on for multiple days straight. At some point, the adjustment becomes almost seamless. As I said, I've been pugging with this and I am not the kind of pugger who intentionally throws often. If it was really as big as a problem as most people think it would be I wouldn't pug soldier or scout on it.

I am also a proponent/sometimes user of mouse acceleration, although xorg pointeraccel is kind of dogshit and I strongly advise Rawaccel on Windows as the single best accel app available on any platform. I can't use it on linux, but I did use it to make my work computer's trackball more comfortable.

Not explicitly a TF2 customization, but I've been using it in pugs and finally got it about where I want it. There's some limitations because it entirely leverages the libinput flat acceleration profile which has bounds of 0 and 1 with a precision of 6 decimal places, but I think it's solidly good enough now.

https://pastebin.com/9mQ4GVvb is the link to the randomizer script.
This only works with raw input disabled ingame.

Main notes are:
1) I do not know the actual factor of the libinput sens multiplication on the flat accel profile. It wasn't on https://wayland.freedesktop.org/libinput/doc/latest/pointer-acceleration.html and I couldn't be assed to look further. There is more to it than (dx * factor, dy * factor) because 0 still works, my gut says something like the "accel speed" is actually a 1:1 additive increase of the vector based on the existing vector multiplied by factor added to the initial vector, but I feel like the sens variance might also be too high for that? I'm not a math brain and I didn't bother to find the relevant bits in the actual libinput source, so at the end of the day I simply don't know.
2) This does not reset your libinput accel speed at the end. You just ctrl+c out of the script and set it back to your preferred base setting using xinput.
3) Keep things integer. Bash no like decimals :(
4) The double brackets may not be portable to all systems, so the if statements on lines 13 and 24 may break. You just want to "and" that the "increment" variable is either true or false *and* that the bounding high/low variables are the relevant number, otherwise the incrementing and decrementing will break.

I'm sure someone will have some idea to improve it because I'm a very stupid dumbass. The first thing would be probably be bothering to do a portable anding of the if statement. It's 2AM though and that requires thought that I am currently not capable of.

Just one more note. The reason I didn't use the coordinate transformation matrix instead is that it breaks 3d games, otherwise I'd have decoupled my Y sens from my X sens globally long before I even considered sens randomizer.

PEBKAC

no, even if he's right there is an obligation we all have to downfrag AimIsADick regardless of what and where he posts.

Also, quantifying actual FPS impact would be more useful because generally TF2 isn't GPU-bound, so there is a serious question of whether the increased GPU workload has more than nominal practical FPS impact on anything above Intel integrated graphics.

it does what he says it does. The only command is typeperf which exists solely to deposit performance data and all he does is manipulate standard output with no output redirecting to a file, no piping to other commands, etc.

https://futurism.com/neoscope/neurotic-introverted-men-risk-gaming-addiction