Top navigation Players Awards Media Partners About
Change skin White Black
Partners
Time:   23:37:16 CET   14:37:16 PST   17:37:16 EST   06:37:16 Seoul   05:37:16 Beijing

You are here: Home / Blogs / mauhiz
BLOGS
Online heroes

By Vincent 'mauhiz' P
Nov 21, 2007 16:23


I just read Harish's http://www.sk-gaming.com/blog/5861/ and I thought I'd make a complete answer.
The fact that 'online' league and players are scorned is not new and even if the disdained people have reasons to be onliners (too young, body handicap, living in a remote area, or always busy at weekends...) there is always a suspicion of cheat.
This is the case for most games - I believe that a poorly designed game like Half-Life (not bad! just lacking modern design practices) will never be cheat-free.
Nowadays everyone is silent-running online, and more and more people can do it without hte mouse wheel. What to do? No anticheat can beat it, like no anticheat could prevent people from bunnyhopping, fastscoping and many other features.
Not bugs, features. If you consider everything you can do without tampering with the game files (and using default GUI to bind mouse wheel to +duck can hardly be called tampering) as features, then you can relax, try to do it you too, and be the one who discovers the next possibility in this game.
A few years ago I tried everything I could find to boost my game performance - from cvar tweaking and aliasing to monitor color profiles, and there undoubtedly a lot can be done! But I stopped for two reasons : I play at lans too (yes yes) and I find I'm having less fun when having an unfair advantage.
My policy is to never comply about supposed wallhacks/16bits/cvars and only stop the match (or practice game) in case of undisputable evidence of cheating (rapidfire, speedhack...). Yes, keep cool, shut it, and focus on your own performance instead of what the opponents can do.
Hey, is not that just what mature people do?



A memo for Valve Developers

By Vincent 'mauhiz' P
Jan 1, 2007 19:03


As good Counter-Strike may be now, I believe it can be even better with little work.
A while ago, a cool guy named SwaT posted a wishlist topic on Steam forums. I followed the topic at that time, but CS:Source was released then, and it was stormed in by CS:S fanatics and in the end we never saw an answer from Valve.
Well, they don't really post much on their own forums, unlike Blizzard, but I think they read most suggestion topics. As a matter of facts, a few bugs were actually fixed, but this is as far as it goes.
I want to use this blog to sum up suggestions I approve (you can always post comments below if I forgot something). I am not expecting much from it but if there is a chance I can contribute to a better game... I am taking it!
Let's begin with THE most annoying issue.
The smoke problem
The smoke problem is dual : fps and visibility issues
First, my computer is quite recent, but I can't get 100 fps when there are several smokes. Smoke are said to be sprites, so they are not rendered textures neither fog. I ran several tests, and they proved that :
- the issue is almost not CPU-dependent (no difference between 1800+ and AMD64)
- the issue is not solved by tweaking cvars (like d_spriteskip or fastsprites)
- the issue is heavily GPU-dependent (almost no fps drop with top notch graphics cards)
- I can reduce the FPS drop by lowering my resolution and my color depth. I get steady FPS in 800*600 16bits.
This brings us to the second issue : 16 bits smoke have almost no effect, and it's terribly annoying, because it hinders the gameplay. Many pub servers have disabled the smoke grenade. Private clanwars can get very frustrating if you can't enforce bit depth. Third-party league anticheats have to take and upload screenshots, which is something that should not have to be done.
I sense something terribly ugly somewhere inside CS' smoke rendering, and this need to be fixed. Upgrading hardware won't atone for poor design.
The shield
I think the shield belongs more to CS:S than to CS. It was never totally debugged, especially its hitboxes. You cannot use your knife in scratch mode, only in stab mode. You have to drop your shield to use secondary functions of your Glock or your USP.
It is not used in tournaments because it is really bugged and imbalanced. Try to kill a shield equipped CT with a mp5, when other CTs behind him can strafe and hit you! The solution is to add a server-side cvar to disable the shield (for tournaments) and to debug it a little more for public use. I would say remove it but shield lovers would throw the first stone to me.
The hitboxes
I understand the need for different layers for display and hitboxes, but the current coordination between those two layers is very poor. Even in the static case.
For example, the new street lamps on de_inferno have hitboxes larger than they appear to, and I always get fooled while trying to throw a grenade close to it, or firing with usp... This won't be hard to fix, and it is very welcome.
On a static player, headshot from behind is also very hard, especially if you are in a lower place. Double difficulty if the player has alt-tabbed, I don't really know why.
More painful is the dynamic coordination bewteen display and hitboxes.
The netcode has to adapt to new broadband Internet connections. LPBs are now the rule. You don't want to use a 100ms default interpolation when your latency to the server is 10ms. You don't want to use a default low update rate or command rate when you can download a gigabyte of porn in two hours.
Steam prompts for connection type (you don't have much choice though), but only sets bit rate. Why do HPBs and low command rate users have to be so hard to hit?
There is a need for a global review of the netcode. A game is not only its graphics!
Silent bomb
If you plant the c4 on the side of a crate, it can be silent. Why ? Probably because the sound emitter is inside the crate. This is highly unrealistic, and could be fixed with simple things like a larger solid bomb (that would be oriented as the bomb poser intended to, and not always the same way) that would not fall from the crate if being planted over it.
Map bugs
Flying crates and flashbugs cause many disputes in tournaments.
A smart guy released a fixed de_nuke with same CRC, but it is not widely used, and not integrated by Valve. Why not?
There are flying crates/walls and flashbugs on de_train and de_dust2 for example. Fix it!
Skywalking on de_dust, roof trespassing (two player towering in a low room, then jumping, the higher can go through the roof) on de_nuke and as_oilrig are also unwanted features and should be taken care of.
HLTV and demo recording
As e-sport has grown faster than imagined, the demand for HLTVs and in-eyes demos has went beyond expectations.
But there are many bugs that plague these two features
First, when you join a HLTV, all players appear as spectators or have empty/null names. I smell a bug in the initialization...
I would also love a flashbang indicator to see if the spectated player is actually flashed.
I understand that HLTVs do not stream frequent updates of the game, which explains kills without visible shots and fuzzy movement. But this is a matter of bandwith, when in LAN you don't want to worry about bandwith! You want a fluid stream !
Even on the internet, why not allow the TV server to stream more data ? And let the administrator tune it.
For the coverage of a large event, I also thought that multicast could be used. I'll let more technical details to those who know them ;)
Secondly, demo recording is a feature that did not evolve from early Half-life time. The file format is very unoptimized (from what I know), as proves the fact that RARing a demo will shrink it to 1/5 of its original size.
There is also game experience hindrance when recording a demo (random FPS drops). This could be because of the many disk access, and related to the previous cited issue. As a consequence, most pro players do not record in-eyes demo anymore. The community is the victim in the end...
Handling demos is also something that can only be done with external tools like SK player or Rewelator. These tools fix the bugs and grab the info from the replays. Why not integrate a tool like this in Steam ? An official one this time. And fix the demo bugs by the way ;)
Server browser
The server browser sometimes prevents players from joining HLTV. This does not happen with the 'connect' command or external server browsers.
Some of the servers dissapear from the favorites from time to times.
The server browser queries all servers when refreshing instead of downloading a list from a master server. This comes with heavy bandwidth usage... I think that no 56k user ever saw the end of the list ;). Something like Gametiger would be better ?
The server browser does not know AMX slot reservation.
HLSW (my favourite server browser!) can do many things. Let's learn from it!
Features
When the game is paused, you can only talk once. Maybe the anti-flood need to differentiate between pause and not pause.
A large part of cheating in Counter-Strike will never be detected by VAC as it is only CVAR scripts. Using bugged cvars like m_pitch, or _special, these scripts can even perform loops without freezing the input. This was definitely not meant to be so. Third party software like WWCL and HLGuard were designed to block this but often fail. We need a reliable way to block cvar cheating...
Also, binding +jump to mouse wheel allows easy bunny jumping and should not be allowed.
Similarly, with +duck, the player can silent run.
Spray and Flashlight can be safely removed - they are only used to fool around.
Why do +use have to produce a sound when not using anything?
When two sounds are simultaneous, only one is played. This is not realistic, and allows silent defusing (flashlight + use) or silent falling (fire when hitting the ground).
Graphics
There is a bug with awp, when the hit mark is not really where the bullet went.
When alt-tabbing while loading the game, some letter sprites disappear and text display is bugged.
Why do we have to reload the game when adjusting brightness in the Video Options ?
Third party integration
This part is not realistic; there is no way a small company like Valve could handle all the bonus the community added to CS. But still if they read this...
At least they integrated RZE's buy aliases in CS 1.6, why would not they go further ?
Some third party maps are much more popular than official one; why not integrate some of these maps in next official release ? I'm thinking of de_cpl_ maps, de_forge, de_tuscan, and some fun maps like fy_iceworld or aim_map which have great success.
Metamod and plugins like AMX brought a lot to Counter-Strike; see the popularity of CS deathmatch (widely used by good players to warm-up before playing a game), StatsMe and Warcraft3 mod. Well there is no way Valve can support them all, but integrating and optimizing Metamod would be a good thing.
For the time being, these plugins usually overload the server; I have a strange feeling when playing on AMX servers...



Oh no, I'm dead

By Vincent 'mauhiz' P
Dec 29, 2006 14:31


You are playing Counter-Strike on a public server. It is your favourite map! de_dust2. You really enjoy rushing in the middle and owning every newbie who cannot buy their weapons as fast as you do. But this time, tough luck! A nasty awper snipes you as you cross the double door.
Oh no, I'm dead!
As frustrating as it can be, you should really not take off and get a snack. You should take this chance to observer better players and learn from their aim, their moves, their grenades, their timing... You don't have to be the best if you just play casually, but getting better won't hurt and only brings more fun.
I began gaming with UT and Quake3. I struggled a long time to learn playing, because my friends were better than I was and I strove to trade places. But, when I finally reached my goal, they all stopped playing...
I thought there were ways to get better more quickly. The first thing that came to my mind was to use replays of better players. I learned the basics for Starcraft and later Warcraft 3 by watching replays of top players, and then trying my strategies against random online players. Yes, in the beginning, I would get better faster. But this method capped my fun and made my learning boring : half my playtime was analysing the cookie-cutter strategy, and other half was practicing it...
This is wrong. A replay or two, why not? But you have to discover strategies by yourself to understand how, why and when to use them, in order to keep it fun and refreshing. This applies to most fun things in life...
I recently found in one of my oldies crates a game called Battle Realms. This game is one year older than Warcraft 3, and still has most of its concepts, with noticeable innovations for a 2001 rts : 3D, uncapped units grouping, automated production with rally points, four balanced races, four tiers... There are unique concepts : peasants are produced automatically, faster if you have few units. If you want soldiers, you have to get your peasants through training, and trainings can be combined. For example, if you put a peasant through archery and chemistry, you will get a cannoneer. There are also things that could have been better off copied by Blizzard : towers are capped (one per tier, rebuildable) and don't shoot unless you have a unit within it; you get less experience (yin/yang) if fighting in your base.
In short, it is a nice game, and I believe that if it wasn't for its non-existant marketing, it would have been a great success. But you have to learn it by yourself : there is absolutely no 'replay' function ;)
Back to Counter-Strike : you would never have thought what was actually making the game fun was getting shot dead every round ;)
« Previous1Next »

Partners