Suggestions for improvements of CS:GO here on HLTV are not only spread chaotically, but the lack of a proper search engine also renders them pretty much impossible to find. I think if the HLTV.org community also wants to try and have a say in what changes Valve should consider in future updates, and not completely leave that to people frequenting reddit, we should have global threads about things people here think are worthwhile. I think this input is needed, as seemingly the site that was officially intended to submit and discuss suggestions (recently renamed to pwnsource.com) is mostly neglected, while popular demands on reddit have on numerous occasions found their way into the game.
The goal is not to randomly gather suggestions though. I would like you to make an effort to describe how and why you think the change would impact the game positively and also engage in discussion should someone question your thoughts. Although ultimately, I just want to have a pool of suggestions for the developers to draw from should they still be considering or ever again consider visiting HLTV.org for those purposes. So the blog turning into a pool of unopposed, uncommented or insufficiently well-described suggestions is not entirely unwanted either, - at least stuff stays easily found and central.
I'll start things off with my suggestions in the main post because they won't fit into a regular post:
- Aimpunch only occurring when shot in the head without helmet
Currently, the view of a player is also disrupted when shot in the upper body without armor. This disruption renders accurate aiming and recoil control practically impossible. Additionally, aimpunch is not considered in the server's lag compensating calculations. Meaning that even though on your screen you aimed and shot perfectly at one place, the server may already have registered you being aimpunched and set off your view, making your shot(s) miss. Removing aimpunch for body shots completely would solve both issues. Aimpunch from a headshot should be included in the lag compensation calculations. Recently Valve changed the animation of the model when shot in the head, because it used to cause the head hitbox to be forced into a different position, leading to perfectly placed bullets (most importantly on M4 2-bullet bursts) missing their target. Removing aimpunch caused by bodyshots is the next logical step, considering it serves the same purpose of removing randomness from duels.
- Increase tagging for bodyshots
The concept behind aimpunch, namely that of being put at a disadvantage or punished for being caught first is however desireable in a competitive shooter. It forces players to play smart and try not to get caught out of position. It also makes sure that instead of putting all of the emphasis in duels on aim, other skills can also be utilised to come out on top. Ultimately, the concept of being punished for being caught serves to create more a risk/reward-type gameplay and therefore enhances competition.
While aimpunch however is a rather severe punishment, tagging doesn't interfere with the other player still being able to react properly. Tagging as a mechanic by design also slows the game down, which is something that GO needs. The quick acceleration and fast movement speeds in the rather small maps allows players to almost always prevent being punished for a reckless move by flying back around the corner they earlier flew out of.
- Add viewscreen interpolation for recoil kickback, aimpunch, falling
The Source engine uses client frames to create smooth transition frames from one gamestate to the next. This way, the game always seems smooth, even when the amount of "real" gamestates is low (as per tickrate) or the flow of gamestate is interrupted (by latency for instance). Currently, the transition from one viewing angle to the next caused by recoil kickback, aimpunch, or falling is not interpolated. This means that, most noteworthily, the smoothness and therefore controllability of recoil depends on tickrate and can be disrupted by lag. Low tickrate and high ping causes choppy feedback. Additionally, since mouse movement is interpolated using client frames, the act of counter-maneuvering recoil with the mouse is applied differently from the recoil itself. Recoil control is not only an integral competitive asset, but an enjoyable and addictive skill and technique to master. Seeing as how recoil kickback is predicted client-side already, applying interpolation should be easy.
- Increase view_recoil_tracking cvar
This client variable determines the intensity at which the view of a player follows the recoil pattern of a gun that is being fired. For reasons mentioned above, proper recoil feedback is desirable. Its value currently sits at 0.45 and should be increased to at least 0.5, and could be experimented with further (preferably for all experimental changes, that are "up to debate" and not objectively undesireable, a test version of the game should be introduced for the developers to implement stuff more recklessly and players to try things out and give feedback).
- Remove auto-crouch when planting and adjust planting hitboxes
A feature that was later introduced into the game and for the same reasons that the planting sound was shortened for should be removed again. Not only does the forced crouch limit the possibilites of the planter, but it is now also easy to see for a CT with a quick peek whether a T is actually planting the bomb or faking it. Furthermore, after planting the bomb, the model is forced into a standing position again, even when holding the crouch key. The planter does not notice this, which for itself is badly implemented, but if he got to see what he would see from that standing position it would at least be somewhat of a compromise, because he then isn't a blind target when standing up (as can be often seen on de_train for example, when planting behind the barrels on outer bombsite).
This auto-crouch supposedly was introduced because players were hard to hit when they planted a bomb. But this was due to the actual hitbox already following a planting animation (that was present in CS:S), while the model didn't. Instead of making the model crouch, just don't change the hitboxes or model at all when planting, except for the arm hitboxes I guess.
- Crouching should be considered in the lag compensation calculations
Crouching like aimpunch is also not lag compensated for. On your screen, you might see someone standing, and shoot him while standing, but once this "+attack" command of yours reaches the server and the server in turn rolls back time calculatively to reproduce the gamestate you executed this command in, the standing player you were shooting in the head might already have crouched. If this is the case, the server will simply assume the player was already crouching on your side of things as well, making your shot miss. Especially the instant nature of crouching in this game makes this a rather game-breaking issue in online play, where crouch-spamming players are impossible to register on precisely.
- The client-side jumping animation should be synchronized with the server-side hitboxes
The client knows two jumping animations. When a player is standing still and jumps, he jumps with his legs spread out, his upper body straight. When walking or running and then jumping, the player model pulls its legs in and bends his upper body. The server only knows a hitbox for the latter kind of jumping animation. Meaning that if a player stands still and jumps, the hitbox the server will register shots for is not where the player models would indicate it to be. The server-side hitbox needs to know hitbox behaviour for both animations.
- Enemies should only visible on radar when spotted by you or a teammate, the bomb should never visible on the radar
I could not properly test this, but seemingly the radar tends to highlight player positions based on sound cues, and in some cases without the player model actually being visible for a player (as seen with the radar showing player through smoke). I am not too certain how exactly the radar functions in this regard, but the radar should not know more than the player can know himself. While the stuff the radar can do more than it used to in both 1.6 and Source may be useful to get newer players to learn how to read situations, it is not desirable that in competitive play the radar turns out more important than proper communication or interpretation of sound in some scenarios. If a team flashes and smokes, or rushes, or even just generally approaches an area, and the player holding that area makes a wrong call ("There are 4 guys here!" when really it's only two), his team should be punished for him not being able to read a situation and communicate properly instead of the team/player just being able to look at his radar and read everything from there. For the same reason, the bomb should not appear on the radar. Faking attacks, confusing enemies is a central strategical element of the game. Spotting the bomb as easily is too big a giveaway and further takes away from the game the necessity to use other tools to read situations. The bomb already emits a blinking light, that should be enough.
- The death cam should be removed
Along the same train of thought, the death cam is anti-competitive in the sense that you are not being punished for not making the smart team-oriented play. Even if you don't maneuveur yourself and play correctly to survive and gather information, but die immediately, you still get to see sometimes even more than you would have if you stayed alive. There used to be fade-to-black, and while I would wish for this to be back for tournament play (although instant black screen instead of still getting to see stuff beyond your death) since it also helps spectators seperating alive players from idling fallen ones (tournaments most of the time have dedicated screens for each player on the stage), I understand that this is not something a developer wants to enfore for every player since it can get quite boring staring at a black screen. Just make it so that after dying, players get to choose between either the still deathcam that shows who shot you from where that is already in the game (because I also agree that it can be frustrating to die and not even get to see how) or instantly switching to the next alive teammate.
- Improve sound system
The sound system currently lacks in directionality and suffers from inconsistencies and bugs such as spectators being able to hear more than alive players, and ghost sounds. Valve have recently harmonized the soundscape by logically deciding which sounds should be heard by other players and which should not be heard by other players and just the player himself. This is important to be able to inuitively know how to maneuver and act in order not to needlessly give information away. But still there are elements where the approach is questionable. Scoping and fire-mode switching should not be heard by other players. Sometimes players or more often spectators get to hear a sound they are not supposed to.
Another step in this direction is enhacing the directionality of sound. Whereas in both Source and 1.6 you were able to pinpoint players by the sound in terms of direction, distance and speed, in GO it's arguably even more confusing to even try to use sound to your advantage. Coming back to the competitive concept of being punished for bad plays, giving away valuable information in the form of sound should be punishable by a good sound representation. In 1.6, you could learn the soundscapes of the maps and the game itself to get good at judging even the number of players and their exact movement at a specific location.
The game is even faster and more random because players can simply run around without it being risky, sometimes it even being beneficial because the sound cues you are producing then serve to confuse the opponent more than they are giving your position and movement away.
- The M4A1-S and USP-S should benefit from not having the silencer attached
Currently, both of these weapons get worse if you detach the silencer in all regards. This is why they are non-existent throughout all layers of ranks and competition. Not only does it not make sense to introduce two completely new weapons and mess with the delicate competitive balance of being able to know which weapons your opponent is able to buy at a given point in the game by introducing loadout customizability, but doing so just to give people silencers alternatives when one could just have added detachable silencers to the existing guns is all the more questionable. Now that we have these weapons, they should at least be more than "silencer alternatives" from a weapon design point of view. Forcing people to leave the silencer on makes even less sense when deciding not to simply add silencers for the existing guns.
On top of that, if anything, the silencer should make the weapon worse, because the silencer itself is an advantageous feature (you are giving away less audible and visible cues). My suggestion is that the M4A1 and USP without their silencer attached should do more damage. The Five-Seven can neutralize a player with 1 shot in the head on close distance. The M4A1 unsilenced, on close distance, should be able to do that as well. That's up for debate though; generally how the M4A1 and USP unsilenced should be buffed is. It's just obvious that they need some kind of benefit from having the silencer detached.
- The desert eagle needs tweaking
The deagle is, not only comparatively, but then all the more, unviable. It is more expensive than more powerful pistols and as inaccurate as to making it a weapon dependent on luck, a factor which should not be at a competitor's disposal. It can stay as expensive, but should be a more viable alternative and while still staying a risk/reward-type of pistol, should be more controllable. The original behaviour was good really. Another option instead of reverting to before everything that has been done to the deagle would be to increase its general accuracy, decrease the vertical recoil, and have the innaccuracy caused by shooting recover faster so that consecutive shots are managable if paced correctly.
- The AWP is a shadow of its former self
Much like the deagle, the AWP is a classic CS weapon, that always was a favourite of the player community and a central element of the competitive mode. In GO, it has like the deagle lost much of its charm. Players do not excel to the extent they did in prior iterations with the gun, its impact on the game has decreased noticeably. And the few players people consider good AWPers all have one thing in common: They are nowhere near able to be consistent with it. Most of them are explosive and known for situations where they seemingly make the impossible possible, i. e. by hitting mind-boggling noscope flickshots and players rushing at them. On the other side, players that used to be known for their mastery with the weapon, and especially their ability to consistently have an impact by playing the weapon right, mostly defensively, have had a hard time doing the same in GO. This to me is not surprising when looking at how the AWP is made to function in GO.
First, it is not perfectly accurate when scoped in. Combine that with the quick movement of player models, their hitboxes being smaller than the model itself and it's easy to see how a defensively positioned AWP, i. e. where the sniper occupies a spot far away from the area he is defending, is likely to miss shots in the long run. Even a perfectly placed crosshair doesn't ensure a hit on long distances with the current amount of inaccuracy there is to the AWP when scoped in.
Then, if a defending AWPer is rushed, he does not have the possibility to consistently control the weapon to take down multiple targets in quick succession. This is because 1. unscoped, the AWP is too inaccurate to even have a chance of controlling it consistently, and 2. quick-scoping is impossible to control consistently either as there is a "charge-up" phase before the weapon becomes accurate fter scoping in and because the scope overlay kicks in before the gun is accurate, making it harder for players to know when they are able to trust the gun for the next shot. Alot of the grief of AWP shots not hitting can be attributed to this false indication of accuracy.
If we truly want to see the AWP have as big an impact again, and see players being able to master it to comparable degrees of consistent brilliance, it needs to have its accuracy increased to 100% for being scoped in, noscope accuracy increased by the amount it currently gets by crouching (and then crouching making it even more accurate), the "charge-up" time needed for it to reach scoped-in accuracy upon scoping in should be shortened by an amount that is up for experimenting, and the scope overlay should only kick in once scoped-in accuracy has been achieved. Although for this last part, there could be a cvar implemented that lets players themselves decide whether they want to have the overlay delayed or not.
- The CZ-75 and CT Autosniper are imbalanced
This has been discussed numerous times recently. The CZ-75 cheap, allowing players to buy it without risking serious economic drawbacks, while at the same time enabling them to, rather uncreatively, take out players that invested up to 20 times the amount of economy into the round. While the increased possibility to be successful in save rounds in GO is positive even from a competitive perspective, being successful without having to risk something in return, or without having to be creative or innovative is not. Also, the amount of won "save rounds" has been getting ridiculous ever since the implementation of the CZ-75. And mind you, being successful in save rounds doesn't automatically mean winning the round. Taking down 2-3 players and maybe picking up guns or planting the bomb was not a rarity to see with the P250 alone, given the skill and effort of the players wielding it. With the CZ-75 it seems, just having a teammate buying 1-2 flashes can decimate an entire rush in seconds by just pointing it at someone and holding MOUSE1. There's no real creative effort or skill of weapon handling necessary to do serious damage, and at least once in a while whinning the round, throwing the entire course of game off-balance.
The Autosniper at least is expensive and has a risk to it. And players can be better at using it by having more appropriate positioning. I consider it an enriching asset of competition. But the handling itself is rather dumb. Like with the CZ-75, you just have to hold down MOUSE1 while aiming in the general direction of something you see moving. Then there's the fact that it is a 2-shot kill weapon even when hitting a player in the foot. This even lead to Valve reconstructing de_train, because the small ramp entrance into the inner bombsite was a popular spot for an Autosniper to easily deny a whole team entrance.
While both teams can naturally use these weapons, inherently balancing the game, the effect it has on the effort and skill players have to put up to be successful is undesirable nonetheless. Both need adjustments to either their rate of fire or accuracy for consecutive shots.
- Smoke grenades need further tweaks
The recent update to prevent one-way-visibility of smoke grenades has made them even more impossible to be able to walk through alive. Your player model can be seen way before you yourself are able to see stuff again. While smoke grenades used to consitute a metagame of their own (not expecting someone to rush through them, the enemy team knowing that they are not expected to go through them, the other team in turn being aware of the enemy team knowing not to be expected to go through them, etc.), currently they are straight up walls. You can't make plays with smokes or use them innovatively, they simply deny an area of the map completely, stagnating gameplay.
There is a bug where looking at the ground at specific angles can cause smoke grenades to simply vanish and reveal the things behind it.
Both of these characteristics are undesireable. While one-way smokes in fact are not necessarily undesireable. They are predictable (keyword metagame), allow people to practice and master them, make plays with smokes of their own or even capitalize on opponents not placing their smokes correctly. Reverting to the old smoke behaviour therefore seems preferable.
Smoke grenades additionally should not puff through objects.
- Increase flashbang effective radius for players and make them longer lasting
Flashbangs currently are easy to turn away from, because they really only catch your vision if roughly 60° within your FOV center. ~90° effective radius seems preferable, because coordinated executions are then rewarded and wrongheaded, sloppy holds are punished. Also, flashbangs used to last infinitely longer than they do in GO. Being blind longer again serves to reward tactical play and slows the game down further. The exact radius and duration come down to debate and experimenting.
- Increase HE damage output radius, increase tagging for HE damage and fix the issue where minor obstacles prevent any damage done to a player
HE grenades really only cause notable damage when exactly placed beside or under a player. And every little object in the way of a player pretty much nullifies the effect of a HE. In the current form, HEs are pretty brute elements of the game that really serve no tactical purpose if not stacked somewhere. In fact, if a HE comes your way, you can either simply keep running or step slightly to the side behind some paper box to prevent serious damage. HEs could be much more - tactical tools to stop teams from mindlessly rushing, to clear areas and ensure map control, to increase your chances in a fire fight, enhacing duel play.
While smoke grenades should not puff through objects, HE grenades could. In 1.6 they had wall penetrability, this would not only solve the issue where every object, be it a paper box, cancels out all damage done to a player but at the same time increase the possibilites of using the HE grenade tactically. Of course adequate penetrations values would have to be found.
- Moltov/incendiary tweaks
The fire grenades have proven to become more of a competitive asset than people had initally assumed. There are many possiblities to use them offensively and defensively, set yourself up for surprising rushes through your own fire, obscure vision with the smoke of the fire. They are interesting and competition-enhancing additions to the game.
What I don't like is that they don't explode upon initial impact, but bounce off of walls. This allows players to throw them around corners without exposing themselves and have another easy way of denying entrance into a site without risking too much. Also, when used offensively, instead of players having to learn how to place them properly to "flush" out certain positions on sites players could be holding, they simply bounce them off the skybox.
Fire grenades should explode upon intial impact.
On top of this, fire grenades also explode in the air if they don't hit an object within a certain time span. But if they do, they do no damage. This should be changed. Either make it so that they only explode upon impact - this would yield new possibilities for the use of them, by being able to throw them further across the map. Or make it so that when they do explode in the air, that they do cause damage. This again would provide players with additional ways of utilising them. The amount of damage and radius are down to experimenting.
- Movement adjustments
GO is a faster paced game than CS used to be and that is not negative in and on itself. The problem with the fast acceleration is that the peeker's advantage is even greater, and that peeking players are impossible to react to in a manner where positionally smart play would be rewarded. On top of that, the fast acceleration makes maneuvering precisely rather difficult, as you slide beyond letting go of the movement key or after landing from a running jump. sv_accelerate should be reduced to 5.
Then there's the limited air strafing possibilites, preventing players from being able to pull off more complex movements and combinations of moves. sv_airaccelerate should be increased to 15-20, sv_staminajumpcost reduced to 0.050. I also suggest that instead of limiting the possibilites of accurate and complex movement, players should be punished when they are attempting to do those but fail. Like in previous iterations of CS, after landing a jump, there should be a stronger stamina punishment, slowing a player down (the higher the fall the more severe the slowdown), while timing consecutive jumps correctly should be rewarded by not slowing a player down upon impact.
Instead of having people slide, but giving them the possibility to quickly strafe into the opposite direction, make it so that you slide less, i. e. are more accurate when strafing, but strafe more slowly. sv_friction should be increased to 5.
- Competitive map workover with community input
The standared maps could need various adjustments inlcuding skybox approach, wallbangs, boosts, visuals, fog, modification of how certain areas can be used, etc. Preferably this should happen with the input of the general community.
- Consistent, intuitive wallbang approach
Players should be able see whether or not something is bangable without having to test it first. Certain textures of certain thicknesses should be globally penetrable and others shouldn't. Currently there's practically no rule or consistency to this. Also, penetrability of objects (damage througput) should be slightly increased on tactically limited spots on the maps. Another idea would be to use the windows that are present on most maps to implement limited strategical, but also expectable wallbanging. For this, windows on the other side of the walls would of course be needed, which is something that from a map design perspective should be desireable anyways though.
- Introduce alternate spraying patterns for crouching
Crouching is not a viable enough feature in duels in my opinion. It has its uses (especially online, as long as it is ignored by lag compensation) because it increases accuracy slightly and can set an opponent's aim off, but if crouching altered the recoil patterns of guns, it would make duels a much more versatile affair.
- Reintroduce hud_fastswitch 1/0
I don't like how the current loadout when set to cl_showloadout 0 gradually fades away. Cycling through grenades can also be a pain and if you click one time too many, you have to cycle through all of them again. hud_fastswitch 0 as it used to be in older versions with a loadout that vanishes instantly after selecting the object and cycles through grenades visually instead of physically is something I'd like to see back again as an option.
- Free some cvars from the cheat protection
Such as mat_postprocessing_enable, cl_wpn_sway_scale, cl_use_new_headbob, m_pitch.
- Add a nicer headshot animation where the model is forced in the opposite direction of where it has been shot from, the head being the main point of momentum.
- Add the possibility to adjust the dynamic behaviour of the crosshair so that we can decide whether it expands only showing the inaccuracy caused by moving, only the inaccuracy caused by shooting, or both.
- Increase the visual indication of bullet impacts (blood for bodyshots, dust, debris and bullet holes for walls and objects).