When Will UT3 For Linux Be Released?

[D0pamine]

Please read what I posted. A hit-scan ( correct term is ray-cast by the way ) yields a boolean result, this is correct, but the DIRECTION into which the ray-cast is done is random. Otherwise I would not have gotten this miserable hit ratios in my testings. Games like Q3A have no jittering on the ray-cast and therefore are always accurate. This is not the case here as otherwise I would have obtained a 100% hit rate which is not the case. It's a common trick done by developers to balance weapons by adding inaccuracy to strong ones or long reload times ( or small clips ). Here though they went too far turning the LG into a laugh-gun. At last in UT3 the sniper is a sniper again as it should be: powerful but slow.

2k3/2k4 sniper and lg are the exact same weapon but with different models/effects/sounds

the sniper is better but only because its harder to see where the shot comes from

[Svartalf]

There is no induced inaccuracy (which, for some bizarre reason, you've taken to calling "jitter") in the Lightning Gun as packaged in UT2003 and UT2004. I can assure you: everything you need to verify what I am saying is right out in the open and easily accessible without a commercial engine license. If you wish to continue to argue this point, you will need to unpack your files and show what code the entire development community has missed in the past five years. Really, though, the only explanation that isn't, "You simply cannot aim," is that you have high latency on your WAN connection.

The problem with your line of thinking is that "jitter" is the industry accepted term for introducing some randomization of the results into something, whether it's deliberate or not. Moreover, he's largly right- if you're in the same position and you don't get 100% results, there's a variance there.

You can say that it's not induced, but that would be inaccurate- as if there wasn't some random variances, due to a physics model that doesn't work consistently to a random-walk algorithm for in-flight behavior, you would get 100% of the shots 100% of the time if they're properly lined up. I've observed a bit of this myself in the game. "High latency on your wan connection" would also count as a cause of Jitter. And I don't seem to recall some of the other online deathmatch type games having quite so much issues with latency- moreover, I've noticed the issue in question from time to time and I don't HAVE a really high-latency connection and haven't had one for quite some time.

PS: "Hit-scan" is an accepted domain term of players, modders, and many developers and could not be called "incorrect" in the slightest.

Uh... I'd not heard the term before and I've been at doing game dev for a bit now.

[geamandura]

I don't remember exactly about this jitter thing, but it sounds plausible at least. Maybe one of you guys could post a 10 seconds clip somewhere on the web, with just this point demonstrated: an enemy standing still, and you taking a couple of shots at him, having him straight in the crosshairs? If there's induced inaccuracy on the laugh-gun then I guess the enemy should stand far away when shooting him, so that the inaccuracy can be amplified by the distance to target?

[Wyatt]

Please read what I posted.

Likewise.

There is no induced inaccuracy (which, for some bizarre reason, you've taken to calling "jitter") in the Lightning Gun as packaged in UT2003 and UT2004. I can assure you: everything you need to verify what I am saying is right out in the open and easily accessible without a commercial engine license. If you wish to continue to argue this point, you will need to unpack your files and show what code the entire development community has missed in the past five years. Really, though, the only explanation that isn't, "You simply cannot aim," is that you have high latency on your WAN connection.

PS: "Hit-scan" is an accepted domain term of players, modders, and many developers and could not be called "incorrect" in the slightest.

[Odin]

@Dragonlord
Of course, UT99 is the best one. ^^

btw. dropping a game mode is not a plus point since you weren't forced to play that mode.

[Dragonlord]

No, sorry, it is dead on. There's a lot of very good documentation directly from Epic on how the engine works as well as countless people that can corroborate this.

I don't remember the exact behaviour of the BFG, having never been a big fan of Quake, but when the projectile component of the Redeemer detonates, it iterates through the active pawns and tests to see if a trace can be made. If it can, damage is done in five waves of greater and greater radius for 250 a shot (less fall-off) with a short delay between.

But it isn't. I'm sorry, but there are no two ways about this. There are two major classes of weapon shipped in UT: projectile and hit-scan. Projectile weapons spawn discrete actors with their own physics and take time to reach their targets. Hit-scan weapons are a boolean trace to wherever you're pointing. With the exception of the high-ROF modes with the minigun, stinger, and enforcer, and the assault-rifle primary (which sucks), there is no induced inaccuracy on any hit-scan weapon.

If you were aiming at the enemy and fired, you hit: this isn't Battlefield 2. It's really that simple. :/

Please read what I posted. A hit-scan ( correct term is ray-cast by the way ) yields a boolean result, this is correct, but the DIRECTION into which the ray-cast is done is random. Otherwise I would not have gotten this miserable hit ratios in my testings. Games like Q3A have no jittering on the ray-cast and therefore are always accurate. This is not the case here as otherwise I would have obtained a 100% hit rate which is not the case. It's a common trick done by developers to balance weapons by adding inaccuracy to strong ones or long reload times ( or small clips ). Here though they went too far turning the LG into a laugh-gun. At last in UT3 the sniper is a sniper again as it should be: powerful but slow.

[Svartalf]

You can play UT3 with newest Cedega, they recently released version 7.1

You know... The big problem with that thinking is that you're telling Epic it's just fine to do what they did when you do that. Each and every copy bought under these circumstances means you don't want a Linux version of anything. Seriously.

Each copy of a Windows title bought is contributing to a Windows network effect- which tells the studios there's not a Linux market and there IS a Windows one.

[tball]

You can play UT3 with newest Cedega, they recently released version 7.1

I don't wanna touch that game as long a native linux client isn't released. And if it is released, it doesn't have my interest.

[zorroh]

You can play UT3 with newest Cedega, they recently released version 7.1

[Wyatt]

Sorry, this is not the case.

No, sorry, it is dead on. There's a lot of very good documentation directly from Epic on how the engine works as well as countless people that can corroborate this.

EDIT: What you describe sounds like the BFG from Quake. This one had such a system ( John-Woe I call this ). I can't remember a weapon in UT which uses this kind of splash damage. It's either a ray-hit or a sphere-hit ( Redeemer for example ).

I don't remember the exact behaviour of the BFG, having never been a big fan of Quake, but when the projectile component of the Redeemer detonates, it iterates through the active pawns and tests to see if a trace can be made. If it can, damage is done in five waves of greater and greater radius for 250 a shot (less fall-off) with a short delay between.

EDIT:EDIT: I think we also don't talk about the same. Looks to me as if you equate "hit-chance" to a probabilistic test. That's not the problem here, it's allways 100%. The problem is that the hit-ray is jittered around the shot direction resulting in misses ( it doesn't hit where you aim ).

But it isn't. I'm sorry, but there are no two ways about this. There are two major classes of weapon shipped in UT: projectile and hit-scan. Projectile weapons spawn discrete actors with their own physics and take time to reach their targets. Hit-scan weapons are a boolean trace to wherever you're pointing. With the exception of the high-ROF modes with the minigun, stinger, and enforcer, and the assault-rifle primary (which sucks), there is no induced inaccuracy on any hit-scan weapon.

If you were aiming at the enemy and fired, you hit: this isn't Battlefield 2. It's really that simple. :/