im not, im making an updated version in another video. cs2 does still have a mouse input problem, and the wider community agrees on that. I just didn't measure it correctly in this video. When this video was created, nobody knew about muzzle flash desync, and I'll be crafting better tests to investigate further - but that will take some time. also the description is updated!
@@qkNorris Few questions. Are you using any "tweak" packs / input lag guides? Quite often they are placebo or can even make things worse. From what you have shown in the video, those changes are fine. I would recommend High Performance Power Plan in Windows. Make sure Reflex + Boost is enabled. We don't want any downclocking / power savings as those can create input lag. Try capping the FPS to 200, then try 300 and play the game. I know in a **controlled** environment the input lag is a little higher but game stability is the more important factor here. Give it a real good shot and see if it happens in the same spots. Also close any background software you have. RGB and anti-virus software has been notorious for creating issues. I had similar issues in another game and spent years going down rabbit holes , but unfortunately I don't have access to the CS2 beta to test.
You are the first one I've seen providing some detailed information for this problem. Good shit, I bet more ppl have been scratching their head about how shit cs2 feels to play atm.
I felt that in CS2, for whatever reason, my AK spray just felt all over the place. I would literally go into CSGO and Yprac Recoil train, making sure i was getting consistent 170-230 scores, come back to CS2, and the spray is just not the same. I also found that I was having trouble acquiring targets with precision, almost always overmoving the mouse past the target and then having to recalibrate, costing a LOT of gunfights considering my spray is starting off-target and the adjustments are almost a full bullet behind. Ive also noticed that you can almost Shoot and drag across the enemies hitbox and get hits. Ive gotten BLATANT missed shots registering as headshots on targets never near the original bullet.
Someone should record like 20 matches and download replays. Play em side to side to see if there is a difference. There is something horribly wrong sometimes server side because you can get 2 kills spraying and then have 0.5s head start on someone and lose it without registering a single hit when it looks like you blasted 10 shots through the enemy. Either antiaim is crazy good in cs2 or there is some sort of sync problem.
I saw some guy figured it out online, rhe first 9 ish bullets of the akrecoil in cs2 have far more vertical recoil than in csgo, test it out yourself, it feels off because you pull down harder and faster than you do in csgo, this easily accounts for so much of how I personally would miss in cs2 whereas in go I would be fine.
@@jepulis6674on god I feel the same way about my bullets not registering then I get random ass headshots aiming and shooting no where near a player. It’s probably the sub shit tick that everyone’s praising volvo for.
@@MontySlythonye i found out how its fully vertical with follow recoil, i would just blast bullets to the right of people based on muscle memory of csgo spray. Just adjusted my spray now its all good like csgo spray to me after using follow recoil to learn where ur actually missing
Don't know if the muzzle flash gives any accurate results on demonstrating input lag but i do know there is something wrong with the game. Any FPS enthusiast knows what i mean when i say comparing 300fps CS2 with 300fps CS:GO is nothing like. CS2 feels like the actions i take (shooting, peeking, turning around) are 60% me having 5K hours in the game and 40% someone grabbing and holding me back making everything feel floaty, unnecessarily light and smooth against my will. Like the game turning on mechanical motion blur and showing me a cinematic 65mm IMAX version of what i actually tried to achieve when peeking around the corner. Been enjoying the game so far, but i am faraway from grinding CS competitively again if this doesn't change.
300 fps in CS:GO vs 300 FPS in CS2 doesn't have the same input delay and never will. It's like landing on the moon and expecting gravity to be the same as earth. Stop converting your sensitivity to the decimal point and go off of feeling. Muscle memory also isn't a thing, your brain does the work. It's why you can't convert it to other games either and why it feels "off" in those games too.
the first problem of cs2 are the hackers, the other things are little things in a BETA, so it's obvious because the objective of the BETA is clean the wrongly behaviours of the game in a previous version.... tomorrow is a good day, will be release we are waiting the new vaclive!
Great video, and I appreciate the deep dive into the subject matter. However, there's a potential issue with your test being conducted on a 128-tick server for CSGO because the data for the two games were not collected under the same tick rate. Ideally, the tick rate should be the same for both games in order to make a fair comparison. Since CS2 was tested on a 64-tick server, then CSGO should also be tested on a 64-tick server. Nonetheless, you've taken a step that no one else has, kudos!
yeah but they played matchmaking and it felt different in both games , even if he tested on 128 accidentally it's still sooooo visible on 64 ticks. also cs2 uses like a sub tick thing idk about that one
Fair point I think. If the tickrate of the local CSGO server was 128 and the CS2 server 64 then the latency difference would be 7.8ms ((1000/64) - (1000/128)) - which is approximately what the difference in max latency would account for in your test settings.
I get your Point, but it shouldn´t make a difference, because the input lag and tickrate are 2 different things. Would be very weird if this changes anything, because the input lag is client-sided. I think the 128 tick debate is kinda exaggerated. It does not change everything. It will help to make a better game or atleast feeling, but it will not improve the client-sided issues. Appreciate your work too.
Also he needs to test without using the muzzle flash, there is a comment up above thats explains better. I don't have the tools to test it otherwise i would test it as well
Something which I noticed when messing with settings between CS2 and CSGO was that if you turn off Raw Input in CSGO it might feel a lot more like CS2, but nothing I did could make it feel like CSGO. Even if there is nothing changing to the input, having raw input off made the aim feel slightly less instant and tactile. Might be worth an investigation
@@antonioluiz3877 you are wrong, if raw input is turned off, then your mouse movement is tied to frame rate. So when u have lower fps u have higher input lag.
I knew it! Apart from the lower frame rate, it almost feels like there's built in mouse acceleration or something. Another thing to note is that "voice_loopback 1" so you can hear your microphone, there seems to be a BIG delay, as well as built in noise supression, which is set far to agressively. You have to yell into the mic for it to pick up half your sounds, no more 3am whispering. Cheers for the video, hope Valve addresses all the issues. CS2 is looking grim at the current state, much like CSGO was on beta. I do like how the M4A1-S feels more like an M4A1-S should be. The metal finishes on weapons look fantastic (like the Deagle)
Bro... use full screen windowed turn on vsynch. If you have vsynch off you get tearing with 300 fps. Problem solved, no need to call Microsoft tech support. Oh an turn off msaa.
Its not placebo, there is something off for sure. Going from a night of CS2 to CS:GO feels night and day and CS:GO INSTANTLY feels so much better and honestly feels so much easier.
You need to run another test that doesn't use muzzle flash. It could be possible that cs2's muzzle flashes are inconsistent and might not have anything to do with input delay. Ideally something like camera movement, its unlikely that the simple act of looking around has changed while the muzzle flash very well could have since most visuals in cs2 have. maybe use a turn bind on left click and make the tool read a high contrast area? If this input is still as inconsistent in both games then there might actually be a proble
Yes, there was a reddit post some time ago, that showed the AK Spray was exactly the same (also input) but the muzzle flash was delayed. I think this video's test is just flawed.
On MrMaxims video "Is this why shooting feels off in CS2?" He explained this for CSGO: If an enemy runs by your crosshair between two ticks (for simplicity, let’s say right in the middle of two ticks: @tick 1.5 ). And you fire at exactly that point (1.5tick). The shot however will not be updated before the next tick (2 tick). If the enemy has passed your crosshair at that time it will register as a miss, even though you shot at the correct time (1,5tick). In CS2 however the if you shoot at the head of an enemy at tick 1.5tick. The sub tick will register the shot fired, but not update the game before tick 2. Even though your enemy is out of your crosshair at tick 2, and it will still register the kill. The strange part is that the fire animation and the actual kill is not displayed before the next tick. 1s/64tick= 15,625ms. The time between you clicked the button on your mouse and the fire animation showing up on the screen in my example would be 7,81ms. Therefore, if you shot right after a tick you could get a delay approaching 15,625ms. This delay between the shot registration and the firing animation makes the shooting seem floaty. In CSGO you at least get the firing animation where your crosshair is at tick 2. My take away from his video: The main thing you are testing here is the actual delays caused by settings and hardware, even though majority of the delay comes from the "shot registration to firing animation"-delay caused by sub tick. Unless I'm wrong, and the delay is only present while hitting a enemy target. Of course, the sub tick makes it fair, the one who shot the other first wins the fight. Not like in CS: GO, the one who still has his crosshair on the enemy by the time the tick updates, wins the fight. If they could process shot registration and firing animation on a sub tick level would be insane. I say make the fire animation, hit reg and bullet holes/blood splatter client sided, update to server at next tick. At this point we are just waiting for the server so we can see our fire animation on the shots we knew we hit.
This happend when me and my friends came back to csgo after a full week of cs2 rank grind. While we goof a lot about how everything loks worse, we notice that every guns feels ver responsive compared to cs2, especially on deathmatch. Headshots feels easier, aiming feels smooth, whie pistols first shot accuracy wasn't as best as cs2, you notice the bullets actually hitting your enemy or not. Hope valve notice thsi since they will remove csgo completely
I tried playing csgo with a friend, it felt like the ground was made of butter, felt like a lamborghini flying around corners, got 16 kills in the end of that wingman. 😂
@@brucieplays3720 Currently still 4 wins to go and i don't play much fps games in a week. Was MG1 in csgo and some of my friend got 3k-6k. I'm assuming i'll get around that since we always queue together.
I unfortunatly don't have the tools to test it myself, but I can definitely agree i feel somethings off (Most likely input lag). Thank you for this test! I really admire your persistence to make this test so detailed and elaborate. Let's just hope Valve can manage to fix this for the full game release :D
Thank you very much for some quantitative testing. What I would find interesting is the distribution of the measured input lag (a histogramm of these measurements), since if there is a spike in the histogram, this could point to even more of a problem. Additionally, as some others have said, this could still be a bug that happens with all animations. As a next step, looking at when bullet holes are appearing might be interesting.
Just got on CS2 for the first time today and it was extremely noticeable. We can’t go back to CSGO now either so we just have to twiddle our thumbs until Valve put out a fix… thanks for putting together the data. Hopefully Valve takes notice.
This is just subtick working as intended. The gun animation arrives at the next tick not on the client side the moment you click. That's it. That's why there is variance. It varies based on where you clicked between ticks. If you click closer to the beginning of the next tick, the animation appears sooner.
Yup, but doesn't change the fact that while your action is based on subtick system (lame way of saying) while animation is based on an actual tick rate system (Again, lame way of saying) they do not align, and if there is a variance as you put it, that means there is a thing that needs to be fixed. And regardless how close you do it next to the tick (Assuming you're capable of tracking and calculating the ticks in cs, and if you can, jesus christ, apply for Nasa already, we need to find some aliens.) it will never be aligned with your actions; which in higher end players who didn't have to put up with something to this degree for ages; is going to be very off-putting, jesus, last time I remember cs feeling this way was in 1.6 and condition zero, but that might be my memory failing me. It shouldn't be down to the player to adapt to a flaw, it should be down to the developer to fix a flaw.
@@xSkittlesxNewbx Nah the problem is 64 tick subtick. In CS2 it has to cross reference multiple subticks to back calculate and determine game state. It's easier when you think of ticks as snapshots in time. The goal of subtick is to calculate precisely when the action took place between snapshots. In order to determine whose action took place sooner, the server has to back calculate based on incoming information to determine the game state between "snapshots", then feed that back to the user, all of which takes even more time. So while the game is figuring what just happened in that sub-tick moment, multiple ticks (snapshots in time) potentially have passed. If the server operated on 128tick, then the feedback loop can be even sooner. Think of 60hz vs 144hz. You don't have to calculate where gamestates were between ticks as much. You now have twice as many "snapshots". This is why 128tick feels so great, and that's why competitive players have been asking for (and competing on it) for many years. With 128tick, you don't really need subtick. For the brief moment where faceit allowed for 128tick CS2, pros went on about how much better it felt. Then Valve hardcoded 64tick into the game. The reason why Valve doesn't want to go 128 tick is because it's 2X the cloud server costs. You're processing twice as much data, so it will need twice the capacity to scale to the same number of players.
This video was so well done and it was genuinely entertaining to watch you suffer through all this. Should definitely do more stuff like this if the opportunity presents itself again. Thanks for the definitive answer.
I knew it wasn't just me and my sh**ty aim -- was goofing around with CS2 for like an hour, and something just felt off. Then someone asked me to join their CSGO matchmaking game and I was instantly hitting at least SOME of my shots (compared to none in CS2). Appreciate all the work you put into this.
Someone posted a Reddit thread yesterday saying that there is no client side response to shooting your gun in CS2, the game will wait for the server to respond before giving you the visual feedback that you shot, so even in offline you have an internal server that has to tick before you see your muzzle flash in CS2. Odds are the actual latency values are much lower, and that the reason you get 6ms sometimes and 20 every other time is because sometimes you are shooting right before the tick ends or right after it starts which will change the delay to your visual response.
@@EriXofficial No, supposed to be within the next couple of weeks as Valve said summer previously, and everyone has been expecting IEM Sydney next month to be on cs2. But there's still no confirmation and in it's current state having a million $ tournament played on cs2 would be disgraceful
I got a big suspicion that this is caused by the variable tick rate cs2 is using. The packets you send behave exactly like you described your input lag.
I knew it I'm a scout main hart and fast flicks are my play style, this got me so mad, this is really killing the cs vibe for me. What a good time it was R.I.P. CSGO ;(
While this shows very interesting results, it is important to point out that this is the visual response time for the firing anim/effects, I think it's important to consider it because movements such as the camera rotation (which really is the first cue for input lag at least for me) and the different firing effects are not handled in the same way, and while it's interesting to note the visual delay is higher for shooting in CS2, it may also not truly represent when the game is processing your input (just saying basing yourself on effects which are themselves triggered by what we're actually looking for isn't giving the straight answer, at least not without knowing for sure how the game handles it in the back, the game may give less priority to some effects/anims while it already handled the hit calculation prior, although here it for sure wouldn't be able to as the hit calculation has to happen on the server, I'm just trying to give an example of my point). CS2 and GO handles the shooting per tick (server ticks), you can see it by yourself by slowing down the gametime progression, the delay becomes very apparent as it has to wait for the next tick, the same is actually true for your character velocity, it updates per tick, but your camera rotation doesn't (thankfully), on the other hand the camera rotation is handled per frame and I suspect as such that its response time will probably be shorter on both games. I'd like to see the same test but with a camera rotation bind instead, I'm genuinely intrigued by the results you'd get compared to cs:go, and I think it's technically more accurate to test input latency this way as it eliminates any kind of potential "visual" only delay from any kind of optimisation shenanigans or potential issues specifically related to anims/effects.
ehh, maybe true in the past where this form of end-end latency measurement was uncommon. Csgo used to have the firing animation a frame late, but they synced up the animation about a year ago, (presumably) to facilitate testing with LDAT/reflex latency analyzer. Even then, in all the games I've recorded at high framerates it's a consistent delay (the frame of input, or the frame after usually) and doesn't introduce any variance of its own (outside of the expected variable render rate). The latency magnitude may be high by up to a frame, but the std. dev is very likely accurate. That represents the variation in frametime, more specifically the variance in how we experience client input. I see a possible misconception about how server ticks interact with asynchronous clients. The quick and dirty is that the client is showing input as soon as it can. shooting, movement, camera. our input displays without waiting for the go-ahead from the server. only if the client prediction is wrong (about an unpredictable event - for instance getting tagged) does the server force the client model to update its state and location. The firing animation does not wait for server authorization, and its display is solely dependent on the client processing/render speed. You might be interested in looking up client prediction models for server authoritative games, You might be interested in looking up client prediction models for server authoritative games. EDIT: Interesting www.reddit.com/r/GlobalOffensive/comments/16kgxg0/cs2_psa_why_shots_that_seem_to_land_behind/
@@laughinginthe90s im not sure where your assumptions about csgo and cs2 come from but you are incorrect, the shot animation on the client is started per tick as the logic for clientside shooting runs per tick aswell. "Csgo used to have the firing animation a frame late, but they synced up the animation about a year ago" is just plain wrong
Yes camera rotation is the way to test this, not shooting. Someone on reddit pointed out that animations like shooting are processed at ticks, so this test doesn't show that CS2 has more input lag, it shows that CS2 runs at 64 tick while the CSGO offline server that was tested runs on 128 tick. This is not the case for camera movements though
Also expecting the updated video since I still believe, like you, that something is off. I hardly believe it is related to frametime variance. VSync on seems to work like magic, but at the price of tanking my frame rate and adding unacceptable input lag.
I've run basically the same tests as you except on windows 11 but the numbers are the same, on my build csgo is way snappier than cs2, and I get the same results in csgo with win 10 and 11 btw, input latency isnt really a issue if u optimize well, cs2 is the issue here
Thank you for making this video. I felt the game was strange and there was some latency somewhere and I couldn't explain it. Also my score in cs2 is so bad compared to csgo... hope they will fix it. Can't understand how some ppl can actually play cs2 and be good at it.
This is not an issue of input lag. This is an issue of the subtick system where shots register ingame at the exact time you click, but the visualisation happens at least 1 tick later.
true, even my sensitivity feels so much different its like when Valorant came out and even u cant do a snappy aim unless u play like a really high sensitivity in that game. This is a sign that its gonna be semi rigged and EOMM to be introduced like valorant in order to make the game balance for every player skills (and to get revenue from more players). Say good bye aim gods and hello to minecraft players lol@@vaughan3974
For me the best settings by far has been reflex on+boost - fps_max 400 (normally my fps go 220-500 so in the middle) and use engine_low_latency_sleep_after_client_tick true in console/config. I honestly felt it was day and night when I changed these settings, but I don't have tools to try it so in case you wanna try more settings I would suggest that. PC for info: 3700x, 3060TI 16gb 3200 cl 16 and a nvme SSD used for everything.
Reflex on + boost, engine_low_latency_sleep_after_client_tick true and additionally windows game mode off, low latency + max fps 500 from nvidia control panel (not sure if those had an impact or not) Previously I had none of those options and this feels better, Thanks :) Ryzen 7 7700, rtx3060, 32gb ddr5 6000mhz
Really good Video! I absolutely had the same feeling after going back to CSGO, I also find the subtick system horrible atm. The amount of times I died behind a wall, is super frustrating.
It's now been confirmed that with the subticks on the server your shot will always be accurate to your click but the muzzle flash takes place on the next tick (60 a second or something), so there can be a time difference between the two.
got much better with the latest update of nvidia, also make sure to turn off HAGS in windows graphic settings, important if you use any kind of capture software or obs
There’s a video explaining subtick and the packets of data varies just like this input lag does. So it’s like the data packets are 1 big one then 1 small one. So could be related. I too instantly get something was wierd about cs2 just overall felt smooshy
subticks are fake, mostly. if you go on the official valve servers they are still running on 64 ticks while on faceit the servers run on more expensive hardware and run at 128 ticks but even with that it doesn't even do much
subtick isnt fake. its a new way they send information on the server so it doesnt matter if its 64 tick 128tick the packets of information are being sent differently so now the game feels different. if it can get evened out then great. if it cant then its a problem created trying to fix a problem that could of been fixed with better servers @@lywoh
@@marcedoreste I know but what can you do more? The one thing that we can make is to share the feedback and that`s it. I know Valve are not listening but we can give it a try
super glad someone made this video, and that the person making this video has a monster PC so there's no question abt hardware capability. the game is just not very well optimized for anything right now. I hope they get it sorted, there are times when the game is legitimately unplayable currently. CS requires far too much precision and accuracy for this kind of input lag to really be acceptable for competitive play.
Id love to see the results of what some of the 3rd party programs can do. Like custom timer resolution, msi mode, turning of hyperthreading for cs, high priority in taskmgr ect. My placebo addicted brain tells me that at least custom timer resolution should do something.
i hope this blows up. we noticed a delay of the voices when talking ingame and in teamspeak. there is a small delay in cs go too but it is bigger in cs2. Thank you for your work!
One guy in the comments wrote about the sound. Here's what I noticed, when I'm playing with a friend in discord, and the friend is talking in in-game chat, I can hear his voice in game with a big delay (compared to discord). In CS GO, the voice doesn't lag behind discord at all, I don't even hear an "echo".
Comparing this excellent analysis to the excellent post _CS2 PSA: why shots that seem to land behind players still hit (even with 0 ping),_ and I'm afraid the findings regarding shot animations there unfortunately kind of invalidate your test. You're measuring mouse klick input to *animation start*, not shot registered, and the post demonstrates how on CS2 subtick system, it's shot animations that are delayed to the next tick, but not shot registration. So the input lag might still be OK - just the shot animations aren't. A problem in its own regard, but more importantly, I think this invalidates the conclusion that mouse movement/aiming suffers from the same input inconsistencies, as that is not synced to ticks.
@@MegaMarcin981 ping is not real it's nearly impossible to have under 10 ping, CS just has a weird system where it drops it by 10 or so in the scoreboard.
I think your CS2 numbers make sense if you look at other videos on how the animations are delayed in relation to the click. There is a delay on the animation in comparison to when the shot is registered. When you click, the shot is registered in the subtick, then the animation plays on the next actual tick. At 64 tick, each tick is 16ms apart. So if shot registers at subtick, then animation plays at tick, your results may be explained. You would expect a 16ms difference between your min and max recorded values if this is the case. I think this makes sense with your CS2 numbers (mins of ~8ms and max of ~24ms)
Another thing with cs2 is that I often feel like a comp game in cs2 feels 10 times more laggy than private match in cs2 , any chance you could hop into a comp and take some time to check it?
You are THE greatest man! I formated my PC, installed windows 11(was w10) and installed the cs2 full fresh and clean. And now im feeling the game smoother, but i dont have resources to test it like you or data to verify if its just my mind or it really fixxed it... One suggestion: try to format your pc and make some tests again with a brand new fresh installed cs2, i think the old data of csgo or cs2 probably are making this inputlag. Or its just my mind/placebo lol.
Great content. I appreciate your efforts here. I also have an FPS stability issue. Where is in a local server it is ok somehow. but in the actual game, it is a disaster. I am not sure if it is really FPS or the input lag gets amplification.
This input lag problem bring memories back when CSGO was launched. Its def a problem with the new subtick problem, and it seems that only time will pulish the game as it should be. Wonderful job done
have you done frametime graphs? (nvidia frameview etc) I think it's straight up bad performance. the 1% lows are WAY worse in csgo than cs2. It's not so much an inputlag problem as it is a raw performance one. the heaviest frames are just too heavy. when I do frametime logging, I get like 500fps, but the slowest frames are like 90fps.
I feel like this is true too, same issues were happening to me in Tarkov where my FPS is high but 1% lows and frame times were so unusual so the game felt off. Though in case of Tarkov, NVidia Reflex helped.
Global and Faceit 10 player here. I'm so glad that I am not the only one. I hate cs2. It's like all the bots and new players from csgo suddenly became average-good players in cs2. Movement, shooting, sounds are so horrible. It's like Counter strike Valorant edition. I really hope they fix shooting and movement. Otherwise, I won't be playing cs2.
After 3 years of hiatus , i decided to play cs2 today. It feels like accuracy flicking does not apply very much, it's more advantageous to spray and control.
since 64 tick is hardcoded into the game, maybe thats why on cs2, you have around 15.6 ms of delay (1000/64 = 15.6) and maybe in csgo you had tickrate -128 (you said you disabled this after every test but oh well) and in testing, you had around 7ms i believe (1000/128 = 7.8) so maybe 64 tick is the problem? doesnt explain why playing mm in csgo felt better than in cs2
What is most noticeable to me is when I try to switch weapons or pull out a grenade, sometimes the grenade will show on my screen but when I click to throw it my rifle or pistol is instantly back in my hands firing...
It is true, the first few videos I watched you can literally notice it and everyone is talking about it. The input lag is crazy and it makes it just so much less crispy to play cs.
What a great work! Hope you've had a good rest after that whole day of data crunching. Hilarious seeing that cut to automatic logging into a csv file :D Worth pointing out is the nuance of standard deviation: comparing CSGO STDEV of 2.37 to best case scenario CS2 STDEV of 4.16, it wouldn't mean just about 75% increase. No, no, standard deviation is the square root of the variance. So the variance of the input lag in CS2 is actually *3 times* (+200%) of the input lag variance in CSGO in best case. Worst case makes an increase of input lag variance of 4.6 times (+360%). Something is very wrong about the game when timings jump around that much while staring at a dark corner in an offline match.
13:35 there actually is animation lag. the animations and "bullets" arent tied like they are in csgo. so shooting is subtick but the animation will play on the next server tick which is why youll shoot while flicking and the bullet will look like it when somewhere else. someone made a video on it but i cant remember who since a few people are making cs2 Lag, tick, input and shooting vids.
and a lot of people are literally ignoring if not deflecting this problem, i literally noticed it the first match i ever joined when i got into the cs2 beta.
Honestly I don’t even have cs2 access but what I can tell even from a 30fps UA-cam video is that there is a delay from taking a shot (headshot for example) and the headshot noise. 100% it’s like half a second that alone must be fixed asap!
@@m.n.s.s2825 Yeah right. Im not hating on valve here overall they are doing a fantastic job but they need to address this. Stuff like this is literally the core of all gameplay.
@@themomorain Fundamental gameplay flaws, Which wanna make you boil in rage wants to quit. Trust me I have few of those my own and I don't want to talk about them much, I know unless they get fixed, I won't play this game doesn't matter if the whole world loves it.
@@m.n.s.s2825 Yep also I watched a couple of videos and people discovered a lot of issues with bhop = random, mouse movement = random => Sprays and flickshots dont hit and input lag (which is also not constant 20ms or something but its random too). So yeah for any pro and high elo player this will be noticable. Some people will care more than others but I wont play if it stays like this aswell … I just cant be bothered to get frustrated by random events csgo has some issues like that aswell (example server and player side is always different from bullet holes to enemy’s when they are running => thats why you have the “feeling“ you dont hit moving targets even when you are exactly on them with your crosshair (even if you host your own server on your pc) But with faceit 128 tick this was still playable) now its even way worse than faceit csgo. So many random factors I cant stay it im sorry.
as soon as i joined cs2 i instantly told about this to my friend. even with my fps over my monitor refresh rate i felt something off with my aim and now finally things made sense
It's an everyone problem. I noticed the same thing, and when I went to GO, I noticed that I was missing the easiest shots because I was getting used to the compensation from CS2. It's really odd
The issue is that the shot fires on a sub tick but the muzzle flash still happens on the standard tick. So your hit registers real time but the gun fires on the tick. In essence, the game is much more accurate but the animation is not. In CSGO. Shot registering and muzzle flash happened on the same tick (either 64 or 128) so it felt more consistent… but the shot Reg was less accurate.
With the muzzle flash desync, mentally that is going to feel "off" but, theoretically. The "feel" should be the physical click. The problem is, I use the sound cues to control my sprays, and the variation is probably enough to make that completely unreliable now.
easy discussion me on CS2 : botfrag 2kills "rage on CS2 open CSGO and play Faceit...I also smash my head 3 times hitting my monitor..its 2023 style" me on CSGO FACEIT : 29kills with 2:9 comeback yes
I tried my best in cs2 and I couldnt frag more than 10 game. I got so frustrated I deleted the game from my 2 account and now I am peace. Let Valve cook and make a stable full release
Thank you!! CS2 had me going crazy as I have no way of testing the latency but something felt super off even though I kept trying all these different settings and tweaks.. placebo is annoying!!
THANKS TO GOD, I thought my mouse is floaty, but my friends said, its not possible, its just imagination. I STARTED TO BELIEVE AFTER 10k HOURS ON CSGO, that my sens is same on CS2. I´m so happy that u prove my feelings from first moments of playing CS2.
Of course, there is. What do you expect when new shaders and more lines of code are added? Valve still uses an OOP architecture instead of a data-oriented design with an entity component system, where every cache miss and CPU cycle is heavily optimized.
Omg, this video was something I was hoping for, I knew that it was an inupt consistency delay problem. Thanks for going for so much hassle just to show us how unplayable this game is. I really hope this gets posted everywhere so valve acknowledges
it's because of sub-tick. the shots are actually more crisp because they are recorded the actually moment you pressed the mouse button, unlike csgo which waits for the next tick. the thing that makes cs2 FEEL laggy is the animation still waiting for the next tick to happen. so your shot happened more accurately to reality than in csgo, but the animation firing at the next tick is where the laggy feel comes from. plus they capped tick rate in cs2, so you might be waiting longer for that animation. in csgo, the animation happened immediately, but the shot is recorded at the time of the next tick, which is why those trailing edge shots can miss, where in cs2, they never will. so hopefully they can change it so that the animation happens immediately like it did in csgo so we get the benefit of more time accurate shots of cs2, but with the crisp perceived feel of csgo
actually the shot doesn't occur when you click, it warps back in time and registers where you were aiming in the previous rendered frame (not joking) i've got a followup coming soon. the previous frame shot reg is... bad.
So now that this has been debunked when are you taking the video down.
im not, im making an updated version in another video.
cs2 does still have a mouse input problem, and the wider community agrees on that. I just didn't measure it correctly in this video. When this video was created, nobody knew about muzzle flash desync, and I'll be crafting better tests to investigate further - but that will take some time.
also the description is updated!
@@qkNorris Few questions. Are you using any "tweak" packs / input lag guides? Quite often they are placebo or can even make things worse. From what you have shown in the video, those changes are fine. I would recommend High Performance Power Plan in Windows. Make sure Reflex + Boost is enabled. We don't want any downclocking / power savings as those can create input lag.
Try capping the FPS to 200, then try 300 and play the game. I know in a **controlled** environment the input lag is a little higher but game stability is the more important factor here. Give it a real good shot and see if it happens in the same spots.
Also close any background software you have. RGB and anti-virus software has been notorious for creating issues.
I had similar issues in another game and spent years going down rabbit holes , but unfortunately I don't have access to the CS2 beta to test.
@@qkNorris why would you leave it up? it is an inaccurate video that you now know is inaccurate and you choose to leave it up anyway?
@@qkNorris They added a separate tickrate for mouse inputs and my theory is that its not synchronizing well with the actual tickrate.
@@qkNorrisI think it's best to edit the title to reflect this, maybe add "(Outdated, read description)" on it to avoid any future confusion
Not all heroes wear capes! Let's hope Valve takes notice and fixes this bs ASAP!
I think valve already know for a long time but they have no idea how to fix it.
Pls read the other comment concerning the muzzle flash, there is only a visual delay
@@oschwald9784It’s not really something you can just “fix”
Also this isn’t an issue do your research
valve needa an anti cheat, there no hero
You are the first one I've seen providing some detailed information for this problem. Good shit, I bet more ppl have been scratching their head about how shit cs2 feels to play atm.
Really appreciate all the hard work you went through. Great video!
I felt that in CS2, for whatever reason, my AK spray just felt all over the place.
I would literally go into CSGO and Yprac Recoil train, making sure i was getting consistent 170-230 scores, come back to CS2, and the spray is just not the same.
I also found that I was having trouble acquiring targets with precision, almost always overmoving the mouse past the target and then having to recalibrate, costing a LOT of gunfights considering my spray is starting off-target and the adjustments are almost a full bullet behind.
Ive also noticed that you can almost Shoot and drag across the enemies hitbox and get hits.
Ive gotten BLATANT missed shots registering as headshots on targets never near the original bullet.
Someone should record like 20 matches and download replays. Play em side to side to see if there is a difference. There is something horribly wrong sometimes server side because you can get 2 kills spraying and then have 0.5s head start on someone and lose it without registering a single hit when it looks like you blasted 10 shots through the enemy. Either antiaim is crazy good in cs2 or there is some sort of sync problem.
I saw some guy figured it out online, rhe first 9 ish bullets of the akrecoil in cs2 have far more vertical recoil than in csgo, test it out yourself, it feels off because you pull down harder and faster than you do in csgo, this easily accounts for so much of how I personally would miss in cs2 whereas in go I would be fine.
@@jepulis6674on god I feel the same way about my bullets not registering then I get random ass headshots aiming and shooting no where near a player. It’s probably the sub shit tick that everyone’s praising volvo for.
some recoils are different in cs2 not totally sure which yet i haven't played it
@@MontySlythonye i found out how its fully vertical with follow recoil, i would just blast bullets to the right of people based on muscle memory of csgo spray. Just adjusted my spray now its all good like csgo spray to me after using follow recoil to learn where ur actually missing
Don't know if the muzzle flash gives any accurate results on demonstrating input lag but i do know there is something wrong with the game. Any FPS enthusiast knows what i mean when i say comparing 300fps CS2 with 300fps CS:GO is nothing like. CS2 feels like the actions i take (shooting, peeking, turning around) are 60% me having 5K hours in the game and 40% someone grabbing and holding me back making everything feel floaty, unnecessarily light and smooth against my will. Like the game turning on mechanical motion blur and showing me a cinematic 65mm IMAX version of what i actually tried to achieve when peeking around the corner. Been enjoying the game so far, but i am faraway from grinding CS competitively again if this doesn't change.
ohh wow exactly what ive been trying to explain lmao
try "engine_low_latency_sleep_after_client_tick true" command. Add to your autoexec and restart.
300 fps in CS:GO vs 300 FPS in CS2 doesn't have the same input delay and never will. It's like landing on the moon and expecting gravity to be the same as earth.
Stop converting your sensitivity to the decimal point and go off of feeling. Muscle memory also isn't a thing, your brain does the work.
It's why you can't convert it to other games either and why it feels "off" in those games too.
the first problem of cs2 are the hackers, the other things are little things in a BETA, so it's obvious because the objective of the BETA is clean the wrongly behaviours of the game in a previous version.... tomorrow is a good day, will be release we are waiting the new vaclive!
Great video, and I appreciate the deep dive into the subject matter.
However, there's a potential issue with your test being conducted on a 128-tick server for CSGO because the data for the two games were not collected under the same tick rate.
Ideally, the tick rate should be the same for both games in order to make a fair comparison. Since CS2 was tested on a 64-tick server, then CSGO should also be tested on a 64-tick server.
Nonetheless, you've taken a step that no one else has, kudos!
yeah but they played matchmaking and it felt different in both games , even if he tested on 128 accidentally it's still sooooo visible on 64 ticks.
also cs2 uses like a sub tick thing idk about that one
Fair point I think. If the tickrate of the local CSGO server was 128 and the CS2 server 64 then the latency difference would be 7.8ms ((1000/64) - (1000/128)) - which is approximately what the difference in max latency would account for in your test settings.
I get your Point, but it shouldn´t make a difference, because the input lag and tickrate are 2 different things. Would be very weird if this changes anything, because the input lag is client-sided. I think the 128 tick debate is kinda exaggerated. It does not change everything. It will help to make a better game or atleast feeling, but it will not improve the client-sided issues. Appreciate your work too.
Also he needs to test without using the muzzle flash, there is a comment up above thats explains better. I don't have the tools to test it otherwise i would test it as well
I ran the same test and it looks like the disparity is tickrate related
i.imgur.com/n6MogBN.png
Something which I noticed when messing with settings between CS2 and CSGO was that if you turn off Raw Input in CSGO it might feel a lot more like CS2, but nothing I did could make it feel like CSGO. Even if there is nothing changing to the input, having raw input off made the aim feel slightly less instant and tactile. Might be worth an investigation
there is a setting in windows that you can disable mouse precision so that the raw input doesn't matter much besides mouse acceleration
@@antonioluiz3877 you are wrong, if raw input is turned off, then your mouse movement is tied to frame rate. So when u have lower fps u have higher input lag.
HI FLEWDUR
Remember that cs2 dont have raw input option in settings so i thinking that it is off since i move my mouse it is slower than csgo
@@o74769 I think your mouse movement is directly related to the framerate anyway lol.
I knew it! Apart from the lower frame rate, it almost feels like there's built in mouse acceleration or something. Another thing to note is that "voice_loopback 1" so you can hear your microphone, there seems to be a BIG delay, as well as built in noise supression, which is set far to agressively. You have to yell into the mic for it to pick up half your sounds, no more 3am whispering.
Cheers for the video, hope Valve addresses all the issues. CS2 is looking grim at the current state, much like CSGO was on beta.
I do like how the M4A1-S feels more like an M4A1-S should be. The metal finishes on weapons look fantastic (like the Deagle)
Bro... use full screen windowed turn on vsynch. If you have vsynch off you get tearing with 300 fps. Problem solved, no need to call Microsoft tech support. Oh an turn off msaa.
@@Puxi Full screen windowed takes the option to choose 144hz or higher away.
@@Puxiwhy are you even saying this?
like who asked?
voice_loopback 1 used to have a speech-jamming delay in CSGO too.
@@Puxilost you at vsync, also who tf are you even replying to?
Its not placebo, there is something off for sure. Going from a night of CS2 to CS:GO feels night and day and CS:GO INSTANTLY feels so much better and honestly feels so much easier.
and now consider its an obsolete game with 13ms input lag (valorant can go as low as 7ms, overwatch 7ms...)
@@schabigerlumpcs2 can go as low as 6ms and as high as 26, so, you know, better or worse?)
@@adisabyrbayev3080oh it’s absolutely worse😂
You need to run another test that doesn't use muzzle flash. It could be possible that cs2's muzzle flashes are inconsistent and might not have anything to do with input delay. Ideally something like camera movement, its unlikely that the simple act of looking around has changed while the muzzle flash very well could have since most visuals in cs2 have. maybe use a turn bind on left click and make the tool read a high contrast area? If this input is still as inconsistent in both games then there might actually be a proble
The turn bind is a great idea!
Pin this comment!
Yes, there was a reddit post some time ago, that showed the AK Spray was exactly the same (also input) but the muzzle flash was delayed. I think this video's test is just flawed.
hmmm i see
snart
On MrMaxims video "Is this why shooting feels off in CS2?" He explained this for CSGO: If an enemy runs by your crosshair between two ticks (for simplicity, let’s say right in the middle of two ticks: @tick 1.5 ). And you fire at exactly that point (1.5tick). The shot however will not be updated before the next tick (2 tick). If the enemy has passed your crosshair at that time it will register as a miss, even though you shot at the correct time (1,5tick).
In CS2 however the if you shoot at the head of an enemy at tick 1.5tick. The sub tick will register the shot fired, but not update the game before tick 2. Even though your enemy is out of your crosshair at tick 2, and it will still register the kill. The strange part is that the fire animation and the actual kill is not displayed before the next tick. 1s/64tick= 15,625ms. The time between you clicked the button on your mouse and the fire animation showing up on the screen in my example would be 7,81ms. Therefore, if you shot right after a tick you could get a delay approaching 15,625ms. This delay between the shot registration and the firing animation makes the shooting seem floaty. In CSGO you at least get the firing animation where your crosshair is at tick 2.
My take away from his video: The main thing you are testing here is the actual delays caused by settings and hardware, even though majority of the delay comes from the "shot registration to firing animation"-delay caused by sub tick. Unless I'm wrong, and the delay is only present while hitting a enemy target. Of course, the sub tick makes it fair, the one who shot the other first wins the fight. Not like in CS: GO, the one who still has his crosshair on the enemy by the time the tick updates, wins the fight. If they could process shot registration and firing animation on a sub tick level would be insane. I say make the fire animation, hit reg and bullet holes/blood splatter client sided, update to server at next tick. At this point we are just waiting for the server so we can see our fire animation on the shots we knew we hit.
ive watched this 300 times already and can confirm it is a Certified IConnerN hood CLASSIC
ah yes, now i can send this vid to people telling me that it's just my skill issue and not a laggy game issue
This happend when me and my friends came back to csgo after a full week of cs2 rank grind. While we goof a lot about how everything loks worse, we notice that every guns feels ver responsive compared to cs2, especially on deathmatch. Headshots feels easier, aiming feels smooth, whie pistols first shot accuracy wasn't as best as cs2, you notice the bullets actually hitting your enemy or not. Hope valve notice thsi since they will remove csgo completely
I tried playing csgo with a friend, it felt like the ground was made of butter, felt like a lamborghini flying around corners, got 16 kills in the end of that wingman. 😂
What rank do you get cs2?
@@brucieplays3720 Currently still 4 wins to go and i don't play much fps games in a week. Was MG1 in csgo and some of my friend got 3k-6k. I'm assuming i'll get around that since we always queue together.
@@brucieplays3720 I got around 9k as the first "Rank" in CS2. :)
They're removing csgo completely? Who said that? You can still play 1.6 and source online so what are you talking about mate
I knew I wasn’t hallucinating.. I feel is so off all the time. If I go back to CSGo I play like a God again
I unfortunatly don't have the tools to test it myself, but I can definitely agree i feel somethings off (Most likely input lag). Thank you for this test! I really admire your persistence to make this test so detailed and elaborate. Let's just hope Valve can manage to fix this for the full game release :D
“Crispness” is the exact same way I described going back to csgo from cs2. I’m glad someone with a social media following has finally brought this up.
The amount of work and dedication on this video is very impressive.
true
this is outdated and debunked. but @qkNorris is getting too many views from this. So the clown isnt taking the video down
I play left handed in csgo for years and now it's so uncomfortable to play cs2, just that visual stuff ruins my entire gameplay.
Thank you very much for some quantitative testing. What I would find interesting is the distribution of the measured input lag (a histogramm of these measurements), since if there is a spike in the histogram, this could point to even more of a problem. Additionally, as some others have said, this could still be a bug that happens with all animations. As a next step, looking at when bullet holes are appearing might be interesting.
Just got on CS2 for the first time today and it was extremely noticeable. We can’t go back to CSGO now either so we just have to twiddle our thumbs until Valve put out a fix… thanks for putting together the data. Hopefully Valve takes notice.
This is just subtick working as intended. The gun animation arrives at the next tick not on the client side the moment you click.
That's it. That's why there is variance. It varies based on where you clicked between ticks. If you click closer to the beginning of the next tick, the animation appears sooner.
Yup, but doesn't change the fact that while your action is based on subtick system (lame way of saying) while animation is based on an actual tick rate system (Again, lame way of saying) they do not align, and if there is a variance as you put it, that means there is a thing that needs to be fixed. And regardless how close you do it next to the tick (Assuming you're capable of tracking and calculating the ticks in cs, and if you can, jesus christ, apply for Nasa already, we need to find some aliens.) it will never be aligned with your actions; which in higher end players who didn't have to put up with something to this degree for ages; is going to be very off-putting, jesus, last time I remember cs feeling this way was in 1.6 and condition zero, but that might be my memory failing me.
It shouldn't be down to the player to adapt to a flaw, it should be down to the developer to fix a flaw.
@@xSkittlesxNewbx Nah the problem is 64 tick subtick. In CS2 it has to cross reference multiple subticks to back calculate and determine game state.
It's easier when you think of ticks as snapshots in time. The goal of subtick is to calculate precisely when the action took place between snapshots. In order to determine whose action took place sooner, the server has to back calculate based on incoming information to determine the game state between "snapshots", then feed that back to the user, all of which takes even more time. So while the game is figuring what just happened in that sub-tick moment, multiple ticks (snapshots in time) potentially have passed.
If the server operated on 128tick, then the feedback loop can be even sooner. Think of 60hz vs 144hz. You don't have to calculate where gamestates were between ticks as much. You now have twice as many "snapshots". This is why 128tick feels so great, and that's why competitive players have been asking for (and competing on it) for many years. With 128tick, you don't really need subtick.
For the brief moment where faceit allowed for 128tick CS2, pros went on about how much better it felt. Then Valve hardcoded 64tick into the game.
The reason why Valve doesn't want to go 128 tick is because it's 2X the cloud server costs. You're processing twice as much data, so it will need twice the capacity to scale to the same number of players.
still feeling this weird ass problem and it's driving me insane
This video was so well done and it was genuinely entertaining to watch you suffer through all this. Should definitely do more stuff like this if the opportunity presents itself again. Thanks for the definitive answer.
What do you mean definitive answer?
I knew it wasn't just me and my sh**ty aim -- was goofing around with CS2 for like an hour, and something just felt off. Then someone asked me to join their CSGO matchmaking game and I was instantly hitting at least SOME of my shots (compared to none in CS2). Appreciate all the work you put into this.
Someone posted a Reddit thread yesterday saying that there is no client side response to shooting your gun in CS2, the game will wait for the server to respond before giving you the visual feedback that you shot, so even in offline you have an internal server that has to tick before you see your muzzle flash in CS2.
Odds are the actual latency values are much lower, and that the reason you get 6ms sometimes and 20 every other time is because sometimes you are shooting right before the tick ends or right after it starts which will change the delay to your visual response.
Dont worry guys Valve still have 2 weeks to get this minor issue fixed and ready for release :D
is there an actual release date so far?
nope still says summer tho which ends next week@@EriXofficial
More like 1 week
@@EriXofficial No, supposed to be within the next couple of weeks as Valve said summer previously, and everyone has been expecting IEM Sydney next month to be on cs2. But there's still no confirmation and in it's current state having a million $ tournament played on cs2 would be disgraceful
I knew something felt wrong when I played a few match in CS2. Thanks for clarifying in this video!
I got a big suspicion that this is caused by the variable tick rate cs2 is using. The packets you send behave exactly like you described your input lag.
I knew it
I'm a scout main hart and fast flicks are my play style, this got me so mad, this is really killing the cs vibe for me.
What a good time it was R.I.P. CSGO ;(
While this shows very interesting results, it is important to point out that this is the visual response time for the firing anim/effects, I think it's important to consider it because movements such as the camera rotation (which really is the first cue for input lag at least for me) and the different firing effects are not handled in the same way, and while it's interesting to note the visual delay is higher for shooting in CS2, it may also not truly represent when the game is processing your input (just saying basing yourself on effects which are themselves triggered by what we're actually looking for isn't giving the straight answer, at least not without knowing for sure how the game handles it in the back, the game may give less priority to some effects/anims while it already handled the hit calculation prior, although here it for sure wouldn't be able to as the hit calculation has to happen on the server, I'm just trying to give an example of my point).
CS2 and GO handles the shooting per tick (server ticks), you can see it by yourself by slowing down the gametime progression, the delay becomes very apparent as it has to wait for the next tick, the same is actually true for your character velocity, it updates per tick, but your camera rotation doesn't (thankfully), on the other hand the camera rotation is handled per frame and I suspect as such that its response time will probably be shorter on both games.
I'd like to see the same test but with a camera rotation bind instead, I'm genuinely intrigued by the results you'd get compared to cs:go, and I think it's technically more accurate to test input latency this way as it eliminates any kind of potential "visual" only delay from any kind of optimisation shenanigans or potential issues specifically related to anims/effects.
ehh, maybe true in the past where this form of end-end latency measurement was uncommon. Csgo used to have the firing animation a frame late, but they synced up the animation about a year ago, (presumably) to facilitate testing with LDAT/reflex latency analyzer.
Even then, in all the games I've recorded at high framerates it's a consistent delay (the frame of input, or the frame after usually) and doesn't introduce any variance of its own (outside of the expected variable render rate). The latency magnitude may be high by up to a frame, but the std. dev is very likely accurate. That represents the variation in frametime, more specifically the variance in how we experience client input.
I see a possible misconception about how server ticks interact with asynchronous clients. The quick and dirty is that the client is showing input as soon as it can. shooting, movement, camera. our input displays without waiting for the go-ahead from the server. only if the client prediction is wrong (about an unpredictable event - for instance getting tagged) does the server force the client model to update its state and location. The firing animation does not wait for server authorization, and its display is solely dependent on the client processing/render speed.
You might be interested in looking up client prediction models for server authoritative games, You might be interested in looking up client prediction models for server authoritative games.
EDIT: Interesting
www.reddit.com/r/GlobalOffensive/comments/16kgxg0/cs2_psa_why_shots_that_seem_to_land_behind/
@@laughinginthe90s im not sure where your assumptions about csgo and cs2 come from but you are incorrect, the shot animation on the client is started per tick as the logic for clientside shooting runs per tick aswell. "Csgo used to have the firing animation a frame late, but they synced up the animation about a year ago" is just plain wrong
Hope he sees your comment and remake the test cause i'm curious as well
Yes camera rotation is the way to test this, not shooting. Someone on reddit pointed out that animations like shooting are processed at ticks, so this test doesn't show that CS2 has more input lag, it shows that CS2 runs at 64 tick while the CSGO offline server that was tested runs on 128 tick. This is not the case for camera movements though
Thank you for taking the time to investigate and for sharing this.
Also expecting the updated video since I still believe, like you, that something is off. I hardly believe it is related to frametime variance. VSync on seems to work like magic, but at the price of tanking my frame rate and adding unacceptable input lag.
I've run basically the same tests as you except on windows 11 but the numbers are the same, on my build csgo is way snappier than cs2, and I get the same results in csgo with win 10 and 11 btw, input latency isnt really a issue if u optimize well, cs2 is the issue here
Thank you for making this video. I felt the game was strange and there was some latency somewhere and I couldn't explain it. Also my score in cs2 is so bad compared to csgo... hope they will fix it. Can't understand how some ppl can actually play cs2 and be good at it.
This is not an issue of input lag. This is an issue of the subtick system where shots register ingame at the exact time you click, but the visualisation happens at least 1 tick later.
No its not, even the mouse sensitivity feels completely off
The beta has 64 tick no subtick
@@levin5365subtick and 64 tick are completely different thing
true, even my sensitivity feels so much different its like when Valorant came out and even u cant do a snappy aim unless u play like a really high sensitivity in that game. This is a sign that its gonna be semi rigged and EOMM to be introduced like valorant in order to make the game balance for every player skills (and to get revenue from more players). Say good bye aim gods and hello to minecraft players lol@@vaughan3974
the mouse response is definitely delayed in CS2
it happened to me today as well ... was playing cs2 for 2 weeks...when i went back to csgo...oh mann...it felt so good ..
For me the best settings by far has been reflex on+boost - fps_max 400 (normally my fps go 220-500 so in the middle) and use engine_low_latency_sleep_after_client_tick true in console/config.
I honestly felt it was day and night when I changed these settings, but I don't have tools to try it so in case you wanna try more settings I would suggest that.
PC for info: 3700x, 3060TI 16gb 3200 cl 16 and a nvme SSD used for everything.
What Graphics settings are you using?
@@ridwanahmed9135 native (1080p), and everything on low/off except reflex which is on+boost. Nvidia settings is just the usual towards performance.
Reflex on + boost, engine_low_latency_sleep_after_client_tick true and additionally windows game mode off, low latency + max fps 500 from nvidia control panel (not sure if those had an impact or not) Previously I had none of those options and this feels better, Thanks :) Ryzen 7 7700, rtx3060, 32gb ddr5 6000mhz
@@SneaaKs Good to hear!
Really good Video!
I absolutely had the same feeling after going back to CSGO, I also find the subtick system horrible atm. The amount of times I died behind a wall, is super frustrating.
Amazing video, thanks for putting so much time into your research! Hope Valve fixes all of this, Cs2 is miles away from the csgo experience!
It's now been confirmed that with the subticks on the server your shot will always be accurate to your click but the muzzle flash takes place on the next tick (60 a second or something), so there can be a time difference between the two.
I fixed this by playing CSGO instead.
yeah but with cs2 full release u will not be able to play csgo anymore
got much better with the latest update of nvidia, also make sure to turn off HAGS in windows graphic settings, important if you use any kind of capture software or obs
There’s a video explaining subtick and the packets of data varies just like this input lag does. So it’s like the data packets are 1 big one then 1 small one. So could be related.
I too instantly get something was wierd about cs2 just overall felt smooshy
subticks are fake, mostly.
if you go on the official valve servers they are still running on 64 ticks
while on faceit the servers run on more expensive hardware and run at 128 ticks
but even with that it doesn't even do much
subtick isnt fake. its a new way they send information on the server so it doesnt matter if its 64 tick 128tick the packets of information are being sent differently so now the game feels different. if it can get evened out then great. if it cant then its a problem created trying to fix a problem that could of been fixed with better servers
@@lywoh
Thank you for your work. Hopefully, Valve will see the community feedback and take it into consideration prior to the full version.
Keep praying bro Valve won't fix it pros been saying this for months
@@marcedoreste I know but what can you do more? The one thing that we can make is to share the feedback and that`s it. I know Valve are not listening but we can give it a try
super glad someone made this video, and that the person making this video has a monster PC so there's no question abt hardware capability. the game is just not very well optimized for anything right now. I hope they get it sorted, there are times when the game is legitimately unplayable currently. CS requires far too much precision and accuracy for this kind of input lag to really be acceptable for competitive play.
Id love to see the results of what some of the 3rd party programs can do. Like custom timer resolution, msi mode, turning of hyperthreading for cs, high priority in taskmgr ect. My placebo addicted brain tells me that at least custom timer resolution should do something.
Apreciate the work you've done for all these testings. Great job!
i hope this blows up. we noticed a delay of the voices when talking ingame and in teamspeak. there is a small delay in cs go too but it is bigger in cs2.
Thank you for your work!
One guy in the comments wrote about the sound. Here's what I noticed, when I'm playing with a friend in discord, and the friend is talking in in-game chat, I can hear his voice in game with a big delay (compared to discord). In CS GO, the voice doesn't lag behind discord at all, I don't even hear an "echo".
Comparing this excellent analysis to the excellent post _CS2 PSA: why shots that seem to land behind players still hit (even with 0 ping),_ and I'm afraid the findings regarding shot animations there unfortunately kind of invalidate your test. You're measuring mouse klick input to *animation start*, not shot registered, and the post demonstrates how on CS2 subtick system, it's shot animations that are delayed to the next tick, but not shot registration. So the input lag might still be OK - just the shot animations aren't. A problem in its own regard, but more importantly, I think this invalidates the conclusion that mouse movement/aiming suffers from the same input inconsistencies, as that is not synced to ticks.
There is no thing as 0 ping except you hosting the server btw
@@reptayN i played a match of cs2 with 2 randoms on my team having 0 ping the whole match. one guy had 1 ping at one point but still
@@MegaMarcin981 ping is not real it's nearly impossible to have under 10 ping, CS just has a weird system where it drops it by 10 or so in the scoreboard.
@@keatonwastaken as somebody who studied computer science, I understand that. But this just means Volvo is lying to us ;)
I think your CS2 numbers make sense if you look at other videos on how the animations are delayed in relation to the click. There is a delay on the animation in comparison to when the shot is registered. When you click, the shot is registered in the subtick, then the animation plays on the next actual tick. At 64 tick, each tick is 16ms apart. So if shot registers at subtick, then animation plays at tick, your results may be explained. You would expect a 16ms difference between your min and max recorded values if this is the case. I think this makes sense with your CS2 numbers (mins of ~8ms and max of ~24ms)
Another thing with cs2 is that I often feel like a comp game in cs2 feels 10 times more laggy than private match in cs2 , any chance you could hop into a comp and take some time to check it?
this isn't unusual, you probably have some connection issues if its noticeably different.
You are THE greatest man!
I formated my PC, installed windows 11(was w10) and installed the cs2 full fresh and clean. And now im feeling the game smoother, but i dont have resources to test it like you or data to verify if its just my mind or it really fixxed it...
One suggestion: try to format your pc and make some tests again with a brand new fresh installed cs2, i think the old data of csgo or cs2 probably are making this inputlag. Or its just my mind/placebo lol.
Great content. I appreciate your efforts here. I also have an FPS stability issue. Where is in a local server it is ok somehow. but in the actual game, it is a disaster. I am not sure if it is really FPS or the input lag gets amplification.
This input lag problem bring memories back when CSGO was launched. Its def a problem with the new subtick problem, and it seems that only time will pulish the game as it should be.
Wonderful job done
have you done frametime graphs? (nvidia frameview etc) I think it's straight up bad performance. the 1% lows are WAY worse in csgo than cs2. It's not so much an inputlag problem as it is a raw performance one. the heaviest frames are just too heavy. when I do frametime logging, I get like 500fps, but the slowest frames are like 90fps.
Can you explain in detail what frame time does and How it affects the game ? Thanks bro
google @@m.n.s.s2825
From what I know, frame time is how long it takes for the image to get to your screen@@m.n.s.s2825
I feel like this is true too, same issues were happening to me in Tarkov where my FPS is high but 1% lows and frame times were so unusual so the game felt off.
Though in case of Tarkov, NVidia Reflex helped.
In some games reflex made my frametimes even worse! so you should test before using it@@keatonwastaken
New to CS, the first thing I notice, is ''we are having fun'' = play streched, low rez, disable everything (turn the game to minecraft) to get 5% edge
Global and Faceit 10 player here. I'm so glad that I am not the only one. I hate cs2. It's like all the bots and new players from csgo suddenly became average-good players in cs2. Movement, shooting, sounds are so horrible. It's like Counter strike Valorant edition. I really hope they fix shooting and movement. Otherwise, I won't be playing cs2.
After 3 years of hiatus , i decided to play cs2 today. It feels like accuracy flicking does not apply very much, it's more advantageous to spray and control.
Youre doing gods work with that reasearch! This is so much more valuable feedback than just saying "the game feels off" or "its input lag"!
since 64 tick is hardcoded into the game, maybe thats why on cs2, you have around 15.6 ms of delay (1000/64 = 15.6) and maybe in csgo you had tickrate -128 (you said you disabled this after every test but oh well) and in testing, you had around 7ms i believe (1000/128 = 7.8) so maybe 64 tick is the problem? doesnt explain why playing mm in csgo felt better than in cs2
you mess tickrate with input lag, it has nothing to do with it.
@@schabigerlumpua-cam.com/video/4jfo6sWhLFs/v-deo.htmlsi=cU2YdGlIo8UsrUXD i was right
I was yelling about this all night last night, the shooting felt too weird. Awesome video!!
I tought I was crazy I searched for it everywhere and I didn't find anything... ggs. Great video!
I literally knew CS2 had something wrong. I could feel it man
Bro drinking beers causes YOU to have input lag
What is most noticeable to me is when I try to switch weapons or pull out a grenade, sometimes the grenade will show on my screen but when I click to throw it my rifle or pistol is instantly back in my hands firing...
It is true, the first few videos I watched you can literally notice it and everyone is talking about it. The input lag is crazy and it makes it just so much less crispy to play cs.
What a great work! Hope you've had a good rest after that whole day of data crunching. Hilarious seeing that cut to automatic logging into a csv file :D
Worth pointing out is the nuance of standard deviation: comparing CSGO STDEV of 2.37 to best case scenario CS2 STDEV of 4.16, it wouldn't mean just about 75% increase. No, no, standard deviation is the square root of the variance. So the variance of the input lag in CS2 is actually *3 times* (+200%) of the input lag variance in CSGO in best case. Worst case makes an increase of input lag variance of 4.6 times (+360%). Something is very wrong about the game when timings jump around that much while staring at a dark corner in an offline match.
The best video i seen so far about latency. So easy understand and you did a great work here!!!
On Aw2521h my low point goes to 4.4ms maxing at ~14ms (cs 2) capped at 360hz
13:35 there actually is animation lag. the animations and "bullets" arent tied like they are in csgo. so shooting is subtick but the animation will play on the next server tick which is why youll shoot while flicking and the bullet will look like it when somewhere else. someone made a video on it but i cant remember who since a few people are making cs2 Lag, tick, input and shooting vids.
finally an excuse for cs2 feeling bad! I've definitely been missing flicky awp shots for unknown reasons
Thanks for spending time on doing this data, had same feelings while I tried to play csgo after 1 week of cs2. Thanks again.
and a lot of people are literally ignoring if not deflecting this problem, i literally noticed it the first match i ever joined when i got into the cs2 beta.
Honestly I don’t even have cs2 access but what I can tell even from a 30fps UA-cam video is that there is a delay from taking a shot (headshot for example) and the headshot noise. 100% it’s like half a second that alone must be fixed asap!
It's so easily detectable, but some fangirls will try to convince it's placebo. The HS sound and HS killfeed coming delay you are right
@@m.n.s.s2825 Yeah right. Im not hating on valve here overall they are doing a fantastic job but they need to address this. Stuff like this is literally the core of all gameplay.
@@themomorain Fundamental gameplay flaws, Which wanna make you boil in rage wants to quit. Trust me I have few of those my own and I don't want to talk about them much, I know unless they get fixed, I won't play this game doesn't matter if the whole world loves it.
@@m.n.s.s2825 Yep also I watched a couple of videos and people discovered a lot of issues with bhop = random, mouse movement = random => Sprays and flickshots dont hit and input lag (which is also not constant 20ms or something but its random too).
So yeah for any pro and high elo player this will be noticable. Some people will care more than others but I wont play if it stays like this aswell …
I just cant be bothered to get frustrated by random events csgo has some issues like that aswell (example server and player side is always different from bullet holes to enemy’s when they are running => thats why you have the “feeling“ you dont hit moving targets even when you are exactly on them with your crosshair (even if you host your own server on your pc) But with faceit 128 tick this was still playable) now its even way worse than faceit csgo.
So many random factors I cant stay it im sorry.
Let's just remember, the next major is in roughly 6 months.
I don't see it fixed anytime soon...
I knew something was off. 2000 hours in csgo and my first game I said to my friend “do you not feel that input lag?”
as soon as i joined cs2 i instantly told about this to my friend. even with my fps over my monitor refresh rate i felt something off with my aim and now finally things made sense
It's an everyone problem. I noticed the same thing, and when I went to GO, I noticed that I was missing the easiest shots because I was getting used to the compensation from CS2. It's really odd
literally every game i feel like im backtracking people. I hit shots that I think I shouldn't hit and I miss shots that I think i shouldn't miss
this is definitely a thing in CS2, low ping advantage is a bit more noticeable, shots are more likely to land if you land them with low ping.
Played against and with you multiple times throughout the years, clicked on the video cause I recognized your name. Great video man :)
The variation in delay is from the animations waiting for the next tick to show the client side results.
The issue is that the shot fires on a sub tick but the muzzle flash still happens on the standard tick. So your hit registers real time but the gun fires on the tick. In essence, the game is much more accurate but the animation is not. In CSGO. Shot registering and muzzle flash happened on the same tick (either 64 or 128) so it felt more consistent… but the shot Reg was less accurate.
bro i thought i was going crazy my damn bullets felt off
With the muzzle flash desync, mentally that is going to feel "off" but, theoretically. The "feel" should be the physical click.
The problem is, I use the sound cues to control my sprays, and the variation is probably enough to make that completely unreliable now.
easy discussion
me on CS2 : botfrag 2kills
"rage on CS2 open CSGO and play Faceit...I also smash my head 3 times hitting my monitor..its 2023 style"
me on CSGO FACEIT : 29kills with 2:9 comeback
yes
I tried my best in cs2 and I couldnt frag more than 10 game. I got so frustrated I deleted the game from my 2 account and now I am peace. Let Valve cook and make a stable full release
@@m.n.s.s2825 that was the problem that I having also
I knew something was wrong, and Reddit told me it was a placebo
Bro! Thank you for taking the time to scientifically gather data and PROVE this issue.
Just wanna say you did fantastic job! Clear testing methods, good overall pace on the video. Short conclusion we live and breath placebo
Great video, thanks - hope they fix this issue soon I'm sure they will. I tend to play on LOW settings as just feels more responsive.
Instantly noticed when i joined CS2, that it has somekind of delay.
noticed this too. its very obvious how different the responsivness is in cs2
Subbed and appreciate your patience ,holy shit
Thank you!! CS2 had me going crazy as I have no way of testing the latency but something felt super off even though I kept trying all these different settings and tweaks.. placebo is annoying!!
only you have this problem, been playing cs2 since nuke launched and there was no problem with it
THANKS TO GOD, I thought my mouse is floaty, but my friends said, its not possible, its just imagination. I STARTED TO BELIEVE AFTER 10k HOURS ON CSGO, that my sens is same on CS2. I´m so happy that u prove my feelings from first moments of playing CS2.
Of course, there is. What do you expect when new shaders and more lines of code are added? Valve still uses an OOP architecture instead of a data-oriented design with an entity component system, where every cache miss and CPU cycle is heavily optimized.
Omg, this video was something I was hoping for, I knew that it was an inupt consistency delay problem. Thanks for going for so much hassle just to show us how unplayable this game is. I really hope this gets posted everywhere so valve acknowledges
it's because of sub-tick. the shots are actually more crisp because they are recorded the actually moment you pressed the mouse button, unlike csgo which waits for the next tick. the thing that makes cs2 FEEL laggy is the animation still waiting for the next tick to happen. so your shot happened more accurately to reality than in csgo, but the animation firing at the next tick is where the laggy feel comes from. plus they capped tick rate in cs2, so you might be waiting longer for that animation. in csgo, the animation happened immediately, but the shot is recorded at the time of the next tick, which is why those trailing edge shots can miss, where in cs2, they never will.
so hopefully they can change it so that the animation happens immediately like it did in csgo so we get the benefit of more time accurate shots of cs2, but with the crisp perceived feel of csgo
actually the shot doesn't occur when you click, it warps back in time and registers where you were aiming in the previous rendered frame (not joking)
i've got a followup coming soon. the previous frame shot reg is... bad.