It's the unlit world vpk that would be causing that. It works fine on a custom server because you're running sv_pure 0, but a dedicated server would be running sv_pure 2.

Surfaceproperties works perfectly fine in either sv_pure setting.

Correct me if I'm wrong, but doesn't it just load a background map so that preloading's streamlined and automatic?

The issue here is whether or not the pcf files can be preloaded at all, or if special workarounds need to be made, so a preload map would only really come in handy if everything else is solved.

Sorry for any mistakes with details here, it's been a long time since I worked with pcf's and I can't do much experimenting myself because I had to move to linux.

Removing the renderer/children/initializers/emitters headers will work on pcf's that are loaded and remove it in sv_pure. I never looked into which one will cause the particle to stop existing, but just clearing all of them works. Iirc, item_fx.pcf doesn't even need preloading (and maybe some of the dedicated unusual effect pcf's, not sure), at least for some effects. It's the exception however, and from my basic testing I didn't find any other files like that.

iirc particle_manifest doesn't get preloaded. That'd be your main hope, since the particles folder is very much in the sv_pure blacklist, so it shouldn't work but might.

I never tried creating new particles by the name of what I want to remove into a pcf that works (for example, make a new particle in item_fx.pcf called "rockettrail" to override the actual rockettrail in another pcf, and make sure its headers are clear so it won't render. Assuming that rockettrail isn't already in item_fx.pcf, I forget). And again, being on linux doesn't really give me much of a chance to test that.

http://i.imgur.com/TDWRydT.png

https://osu.ppy.sh/u/3064974
On the grind to get out of 50k

Tablets are a lot more intuitive for some people, and some people improve faster / get more consistent with one. But if you go through the rankings and see what top players use, a lot of them do use mice (including Angelsim currently at #2), so it doesn't really mean much and is entirely individual.

If you're used to drawing (and are the type of person who might want to get a drawing tablet anyway) it can be worthwhile to try it out. If you're not, and you're used to using a mouse in games (anyone on tftv should be), you'd be less likely to get anything from it. And the main limiting skills are reading, stream speed, and timing consistency, so your choice in input method won't make you instantly better or worse or anything like that.

Coolermaster's good but I can't recommend Corsair. Corsair's got awful keycaps + incompatible-with-basically-every-replacement-set keycaps because of a non-standard bottom row + painfully bulky cable + excess key wobble + extra pingy sounds on key release from the switches not being seated in the case well + bad software imo. They still make decent enough keyboards, but you can get much better for much lower prices.

A magicforce68 with Gaterons is the go-to budget board (gaterons are equal to cherry in every way except the switches are smoother and the browns' tactile bump is bigger). You'll have money left over but it's approximately the same as Coolermaster's quality, and a fair bit better than Corsair. The keycaps are its only real downfall in terms of quality, but they're not absolutely terrible and until you get into the $150 / £120 and up price range no keycaps are going to be particularly good.

The Ducky one as Koncept mentioned is a very good board also, though slightly out of your price range. Noppoo's are decent enough for budget boards, haven't heard any major complaints, but their switches are produced by Gaote (with older stuff by Kailh), which won't be as high quality as Cherry or Gateron (Gaote blues are decent but everything else isn't as good).

I know that this is a bit old but

Clean TF2's method of particle removal doesn't fix fps issues. It just tells the materials to not draw, when most of the fps drops are from the particle system itself and not the materials.
Clean TF2 helps fps through the unlit options (which is incredibly broken on sv_pure) and the model removals (which actually still works 100% fine in sv_pure with preloading), but not through particles.

You can edit the pcf's in a way that both works in sv_pure with preloading and does actually disable the particle system itself, but it doesn't work for most pcf files, which is kinda dumb.

No problem dude

http://www.koalastothemax.com/?aHR0cDovL2kuaW1ndXIuY29tL1hkTk9PZ2ouanBn
i improved it

They don't correlate, because the lag from each one won't align exactly with eachother and both are very inconsistent (even when fps is capped). The simple explanation (as long as your fps and hz are not the same, and both aren't extremely consistent, which fortunately isn't the case) is you add delays together to find the max delay (which is a worst case with the frame and poll being at just the wrong time), so doubling one and halving the other won't give you the same result.

So, using (1/fps + 1/polling)*1000, 120 fps + 1000hz would have a maximum delay/input lag of 9.333 ms, while 240 fps + 500hz would have a maximum delay of 6.167 ms.

Of course, if your mouse works flawlessy at 1000hz, there's no reason not to have it at 1000hz, which at 240 fps would bring the maximum delay down to 5.167 ms. And the minimum would always be whichever's slower of the two, so your minimum would be 8.333 at 120 fps and 4.167 at 240 fps as long as your polling rate's higher than those. 1000hz also has the added benefit of the delay being half as inconsistent (+/- 1ms instead of 2 for 500hz), so you really should go with that.

While they're not natural keyboard layouts exactly, ergo/split keyboards are all the rage in the mech scene right now (and imo better than the natural keyboard layout anyway, if you're going to angle your arms inwards you may as well go all the way be able to move it and angle it in the best position for you). If you're feeling some strain in your typing or would like to have your mouse more centered, an Ergodox (among other, usually more odd layouts like a split Planck) might be a good investment.

A lot of people change cl_yaw to 0.0165 to fix exactly that. And many don't because their aim isn't hindered and that's how they prefer it. It's not really a downside or an objective problem, it's personal.

That's your opinion, and while I also prefer to have 16:9 native, that's all it is and the objective advantages/disadvantages are much more minor than preference.

The extra fov from 16:9 will rarely mean the difference of seeing someone or not if they're close enough to be a pressing concern, and making sure to look around and knowing maps' hiding spots will mean that you basically never run into a problem from it.

I don't see what's so hard about this, some people prefer some things and some people prefer some other things and that's far more important than some occasional minor technical advantage.

Aspect ratio doesn't change the value, it's based entirely on the fov_desired value (which is 90 for almost everyone).

zoom_sensitivity_ratio 0.793471413188092380619906740908785437547348780566443970924

e: ninja'd

zoom_sensitivity_ratio 1.0 is accurate for fast flicks to the very edge of your screen on 4:3 (or the imaginary borders of a 4:3 screen for higher aspect ratios). 0.79347... is accurate for movements that approach no distance. Because you'll probably be doing more of that and adjusting your movements rather than closing your eyes and doing giant flicks, 0.79347... is usually better to have if you're not already used to something else as sniper.

It might just be regional but idk 75% of people I know say "zee" too.

You would put the .vpk there directly into your custom folder, no other steps needed.

I haven't maintained it at all for over a year now though so some things in it would be hit-or-miss.