Conc changes megathread

hex

regarding the timer issue ill quote zike (https://discord.com/channels/235111289435717633/719632195982983308/860638757849268224) and repost his suggestion:

"in theory dropping the throw delay 0.1 would be the same as hhs and thrown concs having a 0.1 fuse difference"
"much cleaner solution than 2 different timers imo"

in total this means both timers are the same, probably 3.8, while hhs are still slightly faster (assuming you chainprime them which if you care about having them go off asap you will). thrown and hh timings are the same with the same beep alignment and hh's go off 0.1s sooner which helps / fixes the 1 conc -> 1 hh thing not giving as much height as is desired.

making conc 1:1 tfc is going to get close to impossible as we start finding tiny things like these that are just handled differently between the two engines. i would say that the mode seems to be for the most part close enough to tfc that we just need to be concerned about making it as enjoyable as possible for zoomers and boomers alike, not necessarily whatever tfc does.

wubs

hex With regards to the throw delay, Zike mentioned that potential solution to me lastnight and it sounds good honestly. It'd work fine for HHs in all the main scenarios - if you're doing the pesky 2 conc HH horizontally, it should (!) solve that issue with respect to boost, and you are still free to delay the HH conc slightly for a midair HH for maximum boost there etc.

I think the sentiment behind 1:1 has been what you've said for quite a while, from me certainly - I'd like as much of the game I know and love to be kept as far as is realistic to do, but certain things being changed either through having to find alternative solutions or even what we could consider improvements sound absolutely fine by me.

Ark_

hex making conc 1:1 tfc is going to get close to impossible as we start finding tiny things like these that are just handled differently between the two engines. i would say that the mode seems to be for the most part close enough to tfc that we just need to be concerned about making it as enjoyable as possible for zoomers and boomers alike, not necessarily whatever tfc does.

I don't care about being 1:1 TFC on paper, it's possible to do if you want to spend many hours on testing and changes, but what I am much more concerned with is having a reasonably similar experience. I'm generally a proponent of changes, like adding bhop and the currently proposed crouch sliding mechanic. In my view things we can do that don't change the experience in momentum that much but make playing older maps much more on par, such as the conc timing and knockback changes, are worth doing as a rule. Even though we could just make up some values that work and say "good enough" like FF did, the process we have gone through has been worth it.

Conc is otherwise generally very close to being finalized, god knows Goc is done with our constant wanting to test different variables. I don't see much else being very heavily discussed other than minor squabling over precise values that only make minor differences.

Kanin

Here's my input on the matter

We're not getting TFC movement due to how Source engine works. So because of that I am ok with the current AA and Decel rates. I am against raising the AA as it is already extremely powerful.

As for the timer issues. I don't have much of an issue with 1 or 2 timers (some people are strongly against it though). What I will say though is this: I have tested 3.8 and 3.9 on both the HH and Thrown timers.

3.8 HH feels great. I can't fault it.
3.9 HH EDIT: This is fine
3.8 Thrown feels a tiny bit inaccurate from TFC. For example, a lot of the time you hit a conc really low to the floor and it should be a fail, but you get more height than you should. I believe this is because the explosion isn't as synced with the timer as TFC (so it's a few miliseconds earlier than it should be. This is NOT a complaint by the way, just an observation) If anything it adds some leniency to the mode
3.9 Thrown feels "slightly" better (there is not a lot in it though). It feels more punishing than 3.8 and more TFC-like. If I was forced at gunpoint I would proba;ly prefer this, but due to how much better 3,8 HH feels that leads me to my final point on this matter

CONCLUSION: If we are using only 1 timer for both HH and Thrown (and I believe we are defintiely heading that way), then I vote for 3.8 (or 3.81 if you want to be sciency). 1 conc HH timing is more important for the game mode at lower tiers (which will have the more active player base) than a slighty more TFC-like high tier gameplay

Other people feel free to test these values yourself and confirm/deny my findings.

Mac

3.9 Thrown feels "slightly" better (there is not a lot in it though). It feels more punishing than 3.8 and more TFC-like.

I did some more testing with the thrown grenade timers, and basically the added difficulty only comes from the beep timings rather than the extra 0.1 seconds added onto the timer. I changed the length of the beeps to match 3.9 as they do with 3.8, where the explosion happens on the third fast beep, and I genuinely felt no difference between 3,8 and 3.9. If you want to try it yourself, setting pitch to roughly 97 in game_sounds_momentum_weapons.txt makes it line up almost identically.

If you really want to make a consistent timer between TFC and momentum, 3.9 default for both thrown and hh would technically be more accurate to TFC and hhs should feel absolutely fine with adjusted beep lengths, but I honestly can't tell the difference between 3.8 and 3.9 from my very brief testing.

As for the 0.7s throw time I'm all for it.

Kanin

Mac Which line do I change for the pitch? I found 2 lines that say "70,80" Do I change them both? and if so, do I change both the values or delete and just add "97"?

I'll test it more tomorrow

Mac

Under "ConcussionGrenade.Timer" set "pitch" to 97.

Just remember to change it back to "PITCH_NORM" when you're done.

Kanin

New day, more testing. Changing the pitch value of 3.9 does basically make it the same as 3.8, which seems kind of pointless for testing which value is better 😛

My 3.9 HH issues were user error, so I apologise for that. It's fine, All of my playtesting on 3.8 was screwing me over.

If anyone has experience with TFC and MM it would be nice if you could maybe do a runthrough of some maps on both. Don't forget to change both the values also. I still believe that 3.8 makes your concs detonate slightly too early and that 3.9 is more accurate.

Just be aware that youve had hours and hours of MM timings so far so it's going to feel slightly off at first.

Zike

Casting my Official conc_ninja World Record Holder Vote™ for the current defaults of 3.8 fuse time and 0.7 throw delay. As far as the throw delay goes, looks like it really only affects starts to certain prec jumps that relied on 0.8 second chain primes (ninja j2 and goc2 j1 for instance). These setups are kinda awkward especially compared to the chain prime. Having to delay a conc by 0.1 seconds is annoying, but these setups are still very much possible. They can also be pretty easily adjusted to work with a 0.7 throw delay to keep the setup as chain primes.

My argument for 3.8 fuse is so no one has to bother messing around with the beep sound to fit 3.9 fuse properly. This aside I also just like my easy game with 3.8 fuse. Realistically though definitely feels like 3.8/3.9 doesn't really matter.

Gocnak

So, let's try to aim for a selected timer length, whether mom_conc_use_tfc should be 1, and so on, by ₀.9.3. Can be done before then but I'm going to start hardcoding stuff again around then if it's not.

wubs

To save anyone reading the longer blogpost below: current default settings that I've been playing with are fine. Ship it.

The game is undeniably different in a few small ways which is translating to (juggles, at least) being a chunk easier in MM vs TFC, but ultimately, the question the TFC boomers have been discussing the most is simply: is it fun?

That's one thing everyone has agreed on, and feedback is resoundingly positive. Emp played again for a bit yesterday/lastnight and the verdict was basically "I had fun to the point where I want to play again, so I'm happy".

3.8 fuse is fine (and like Zike said above, involves a chunk less work, which is a plus at this point), mom_conc_use_tfc = 1 is what I've been using since it was implemented. I would say that conc behaviour specifically is in a very good spot, and I think everyone that generally talks about conc stuff would be more than happy with what's currently on the plate.

There's still some rumblings of things like brushing walls killing your speed and being very punishing as a result, but it's not something I've personally looked into yet so if Kanin or Gnu wanted to raise it here or in the channel that'd be good to get more eyes on it.

hex

wubs There's still some rumblings of things like brushing walls killing your speed and being very punishing as a result

i assume this is separate from texturebugs like here? if not we do plan on (trying) to fix texturebugs absolutely. if it is something else then idk what it is so someone would have to raise concerns to get it known about yeah

Brunch

My vote goes for lower timer (3.8) basically because I believe easier timer is more fun while being virtually the same as 3.9.

hex i assume this is separate from texturebugs like here? if not we do plan on (trying) to fix texturebugs absolutely. if it is something else then idk what it is so someone would have to raise concerns to get it known about yeah

Slamming against a wall diagonally absorbs your momentum parallel to the wall, which didn't happen in tfc (you'd slide fully keeping your parallel momentum). It doesn't really affect concing but it's mildly annoying to those not used to it.

Shegetz

I wouldn't have any issues at all if the mode is shipped as is besides issues like texturebugs which I know are broader in scope than just conc. The 3.8 timer feels good and I like having both thrown and held concs have the same timer length. Concerning every other cvar I like them as they're currently defaulted to, which includes the .07 throw delay. The only mechanic I personally want removed before the mode is complete is reduced AA on crouch. It's not a huge issue but its also not a very interesting or enjoyable mechanic to me personally.

Mac

Keep aa change on crouch, don't give in to the lesser skilled crouchlets.

wubs

because I know this is already going to get lost, Gnu has referenced the following:

"hitting a wall at a shallow angle sometimes triggers a landing sound (footsteps) and kills speed"

"hitting the wall and not getting this sound still slows you down more than normal, friction seems to be high"

I have not tested any of this, but feel it is probably best to document before 500 more messages drown it out

GNU WANTS TO MENTION

someone said sv_friction might be responsible for rampslide behaviour being different to TFC and this might be the same thing_

qr_iLoveYou

mom_conc_use_tfc = 1, ship it.
I'm impartial to timer length.
Leave reduced AA on crouch.

Zike

So somewhere in the sea of discord, tyrant mentioned he thought concs weren't falling at the right speed. This is kind of annoying to really investigate cause of concs in TFC just disappearing after a long enough fall. I came up with this side by side here:
https://streamable.com/1nhi2z

These 2 different comparisons probably aren't perfect, but it does look like the TFC grenade hits the ground ever so slightly sooner in both cases.

If you go frame by frame you see this, where the TFC nade is like ₆₄ units in the air while the MM nade is only just hitting the ground.

I have the test maps if anyone really wants them, and the starting height is 2880 above the ground.

If anything this would probably be bumping the gravity up from 0.81 to 0.82.

Gocnak

So I'm going to offer a compromise for now. I'll let the conc grenade settings be tweakable in -mapping mode:

mom_conc_setting_gravity
mom_conc_setting_friction
mom_conc_setting_elasticity
mom_conc_timer_prime_length
mom_conc_timer_hh_length
mom_conc_timer_thrown_length
mom_conc_bloat_bounds

In exchange for now just hardcoding the "confirmed good" settings:

mom_conc_use_tfc
mom_conc_use_consistent_hh_boost
and mom_conc_det_distance_check_type

By 0.9.3 we can get the conc grenade settings hardcoded again, fixup trimping and any other minor movement discrepancies, and call Conc more or less "ready" for 1.0.0, where dev effort can move over to get Tricksurf ready for 1.0.0 next.

wubs

Thought I'd replied to this, so apologies - hex just raised this in the discord channel as a prompt, and as has been said there, noone's really complaining about settings anymore.

The only potential conversation/thing to resolve in my mind (though I'm not sure it's that important to me tbqh) is the gravity thing above. From zike's test, 2900~ unit fall with a _64u difference is around 2% variance, which could easily be explained by a marginal difference in time when the conc(s) are released.

I suggested that any bigger effect might be tied to 'stretched textures' as it kinda resembles an existing problem in TFC, and tyrant initially spotted it on his own map which has huge wall textures.

It could be worth repeating for stretched wall textures (idk what tyrant used in xtyr but the old _x maps in TFC had massive stretching). xtyr definitely feels 'floatier' on j3/4 in particular, but if anything, this is more 1:1 to TFC than we'd bargained for in my mind.

I know B has played xtyr a fair bit so I'll try to catch up and ask for thoughts vs old TFC juggling etc etc.