Shotty + Unlagged

[Nux]

Though I've complained about the idea behind unlagged in the past (degrading the use of dodge and such) this is not such a complaint.

This is a technical issue. I've noticed the shotgun doesn't work well with unlagged. Has anyone else? Unlike the other weapons, it seems to have a delay to when the hit is detected. More than this, it makes clear misses hits and clear hits misses from my point of view (something quite contrary to what unlagged is meant to achieve under hitscan).

Could it be that this is due to the simultaneous burst of many hitcan shots which can't be handled well by unlagged?

I will point out that I have noticed this many times over the months of playing on unlagged servers (which has resulted in me using shotty a lot less) and note that my ping on these servers is around 100 if that helps.

[Odin]

We get no such problem on PureTremulous. If anything, the Shotgun is even more deadly because it's suddenly more accurate because of Unlagged.

[Nux]

Please provide your pings with your accounts.

[Odin]

I get a stable ping of 70~ on PureTremulous 1.

[kevlarman]

the weapon firing animations are server events, so it looks like it takes time for the shot to fire. it's an illusion though, the shot fires instantly, and as you noticed you will get blood splatters in places where aliens no longer are. there is a work-around, but it requires 1) a client side mod and 2) letting the client predict a where the pellets (and all other weapons with spread) will hit, which means aimbots with perfect accuracy even when a weapon has spread.

[Nux]

kevlarman, useful comments please.

I can use this weapon on low ping servers without unlagged no problem. I know when a hit should have registered.

EDIT: ah, there's more to the post than "the weapon" i see

Well maybe the animation screwed up my timing... but I don't think I wait for the animation to do it's job..

I'll be back after a brief test.

ALSO EDIT: Well I noticed it again. Don't get me wrong, persistance meant that I still did well with the shotty, but the silly misses and silly hits are still there (and it's not due to animation). The blood splatter in the wrong place thing I'm quite familiar with, this is something different. This is about where I shoot and where it hits.

What else could this be.. suprisingly high spread shotgun? (I don't think these servers have modified shotgun), Me being too used to lagged servers? (I have lowping in that case, so I doubt it)

Also, those marks that appear when a hitscan shot hits a wall.. when/where are they calculated? What I want to know is (in an unlagged server) if an enemy can be hit by shots that were determined to have missed one target. I hope not as this would give every shot double potency.

Oh and off-topic (while I'm making with the questions), does anyone know of a delay built into an aimbot? I've seen confirmed botters aim just where the target used to be. Is this a way in which botters try to avoid detection?

[Drevo]

Even when unlagged was still common on servers in Europe, phenomena like the one you described happened to just about everybody, with pings in the 20-100 range. It's not just the shotgun, and unfortunately, not about 'clear hits' either. It was quite common (from my experience) to see players hitting dretches with MDs way, way off their hitbox (you could actually see the little green splash of blood appearing in thin air), behind corners, etc. Same goes for aliens hitting humans where they used to be (as opposed to where they actually are at the time of hit calculation), which is the reason why unlagged renders human dodging next to useless.

Another issue related to this is the apparent increase in alien range. Whether you were playing alien or human, you could clearly see that alien range is somehow 'augmented' past its non-unlagged norm. I believe this is not because of any actual modification to alien range, but rather a functional modification caused by questionable unlagged mechanics. Just imagine a situation where the human is backpedalling away from you, when you decide to attack. Just like in the example above, your position is not rolled back, but the opponent's position IS. Which means that the 'effective' distance between you and your target decreases, at least as far as unlagged code sees it. This might put your target within your range, very much in the same way as it put your target within the arc of your attack when he was dodging sideways.

Another problem with unlagged is, whether it actually benefits every player equally, regardless of that player's latency. If that's the case, the low-pingers' advantage over high-pingers is still retained, which doesn't make unlagged exactly helpful to what I would consider the target group of players - the ones with high enough latencies to make the game unplayable. Some time ago, there were a few attempts at making a latency-dependent modification of unlagged in Tremulous...

As for the aimbot question: what you're looking for is a simple horizontal offset setting, which can be readily changed. The aimbot's tracking still works normally, hence this could hardly make an aimbotter less likely to spot. You could make this setting (as well as the vertical one) random with every shot, but that would make the aimbot useless, plus it would certainly look very weird.

[Odin]

The backpedal issue you describe is due to the difference in ping between you and the other player. Unfortunately, Unlagged seems to favor players with a high ping, so they can hit you from much further away than you could normally hit someone because they are behind in server messages. So, because of this, the server shows you as closer to them on their client when they are further from you on your client. This is why lagging players on the alien team can hit "from 15 feet away."

[temple]

Unlagged has its flaws.  I play on other games that have unlagged built in and no one complains because its just a part of the game.  You have to take responsibility for being hit even when you think you shouldn't have.

The whole point is considering the alternative to unlagged. 

[Nux]

Though I appreciate the comments, I think you might be missing my point. What you are talking about is the problems with unlagged when it does what it's meant to, whereas I'm saying it's not doing what it's meant to.

Unlagged advertises Direct aim => Hit, in all hitscan cases

Is it just my ping doing this? Is 100 too much for consistent shotty?

To be clear: I'm glad for Unlagged. I wouldn't be able to play over on American servers without it. That said, I'd hate to have laggy players join an unlagged server I have low ping with. It ruins dodging.

[kevlarman]

Also, those marks that appear when a hitscan shot hits a wall.. when/where are they calculated? What I want to know is (in an unlagged server) if an enemy can be hit by shots that were determined to have missed one target. I hope not as this would give every shot double potency.

the marks are the same ones that would have been blood splatters if they hit something. the unlagged algorithm consists of looking up the time that the shot was taken, moving all players to the position they were at that time (player positions are recorded every 50ms, 500ms of positions are kept for this purpose, when the time isn't a multiple of 50 the server will linearly interpolate between two recorded positions), checking if the shot hit, and returning all players to their "real" positions. there is no way for a shot to hit something twice.

Oh and off-topic (while I'm making with the questions), does anyone know of a delay built into an aimbot? I've seen confirmed botters aim just where the target used to be. Is this a way in which botters try to avoid detection?

unless they're aimbotting at the datacenter, they're going to have a significant delay between where the server sees their opponents and where they see them.

[Lakitu7]

Regarding blood splatters out of nowhere, spurious hits on dretches:
You're aware that all hitboxes are rectangular, right? Thus, the corners of hitboxes can seem quite far off from the model.
Also, you know that the dretch hitbox in 1.1 is quite a bit larger than the model, right?

Both of these are the case regardless of unlagged, but when people are questioning every shot because they doubt unlagged then it becomes more noticable.

Also the wall decals from bullets are not changed by unlagged, I believe. Client-side components would have been needed to port that part of q3's unlagged.

[Nux]

Yet again I get the impression I haven't been understood.

I'm not asking whether a shot can hit twice. That's not what I mean by 'double potency'. I'm refering to a doubled likelyhood of hitting. As a simple test, can you tell me if a hitscan shot that you aimed in front of a target, so as to compensate for lag (leading) in unlagged, can hit said target.

This shotty problem has nothing to do with hitboxes and such.. I'm not a noob you realise.

The aimbot delay feature I was referring to is not related to lag. Can't you imagine that an aimbot might be made less noticeable by delaying the aim? It misses fast passing targets but hits them if they remain at one direction for any amount of time. Such a thing can be adjusted to suit cover and abilty needs.

[kevlarman]

Yet again I get the impression I haven't been understood.

I'm not asking whether a shot can hit twice. That's not what I mean by 'double potency'. I'm refering to a doubled likelyhood of hitting. As a simple test, can you tell me if a hitscan shot that you aimed in front of a target, so as to compensate for lag (leading) in unlagged, can hit said target.

if you lead you will miss.

[Lakitu7]

Kev already explained that there's not a double likelihood of hitting. The hit check is "rewound" to a previous time and done there. There's no double likelihood. It's not done twice, just at a different time. And as he's said again now, if you lead, you will miss (except for hitboxes etc.)

[Dracone]

True, there are still flaws, but they're very rare. Just watch some demos sometimes, because demos cancel out all noticeable effects of lag that you see in a normal game. It appears more as weird slowdowns during the times the demo'd player(s) lagged. But it accounts for the ping you had and everything is instantaneous, like everyone pings 0. However, I've noticed times where I shoot somewhere different and I do kill the dretch. But it's so rare, no one should really bother to care.

The point is, no matter how low your ping, you really need 0 ping to not see the shots connect afterwards in a different place from where the dretch died.

[Nux]

if you lead you will miss.

k.

Kev already explained that there's not a double likelihood of hitting. The hit check is "rewound" to a previous time and done there. There's no double likelihood. It's not done twice, just at a different time. And as he's said again now, if you lead, you will miss (except for hitboxes etc.)

He said 'there is no way for a shot to hit something twice' which told me he didn't see what I was saying. Double likelyhood doesn't mean double hit.

It's the 'rewinding' notion which brings questions to mind. Does every shot get rewound and checked? Or only the ones deemed likely to have hit? Does it rewind for hitting teammates? Does it rewind shots that were made after your retroactive death?

The reason I ask is because it seems like the multiple shot nature of the shotty is slowing things down or worse, causing error. Without knowing the answers to these questions, for all I know it's some exception handling mechanic which breaks consistency to cut down on run-time.

[Lakitu7]

Kev will correct me if I'm wrong, but I believe that all hitscan shots are "rewound." The game isn't going to calculate a likelihood. Doing the hit check at the previous time wouldn't be significantly more expensive than doing it at the current time, assuming that the previous time's data is all already there (which it is). Thus, computing a likelihood and only doing unlagged for "likely" hits wouldn't save anything and they're all (assumedly) checked. All shots of hitscan weapons being checked at the previous time would include those that hit teammates. The code really looks like

fire_weaponname{
 do_unlagged_rewinding()
 hit checks
 undo_unlagged_rewinding()
}

I would say that shots made after your retroactive death are applied, just based on experience. It's possible for a dretch and a shotty to both kill each other. It's somewhat fair in that the high-ping guy lives slightly longer because he doesn't know he's dead, but the low ping guy also lives longer because the high-ping guy's shots didn't hit him yet.

[kevlarman]

Kev will correct me if I'm wrong

yup

I would say that shots made after your retroactive death are applied, just based on experience. It's possible for a dretch and a shotty to both kill each other. It's somewhat fair in that the high-ping guy lives slightly longer because he doesn't know he's dead, but the low ping guy also lives longer because the high-ping guy's shots didn't hit him yet.

i don't quote understand why it's possible for two players to kill each other with hitscan weapons, but unlagged changes nothing here, you still have to be alive to get a chance to fire.

[Dracone]

Yeah, that's where ping comes into play. With unlagged, you can aim find with any ping, but with a high ping, it is still difficult to shoot, if you get my meaning.

If you and an opponent are at 1 hp and shoot/slash/bite/whatever at eachother at the same time and both are dead on, the person with the lower ping will win.

It's unfortunate, but the restriction of the speed packets come through is a law of the internet in itself, and will not be fixed until someone comes up with something that can be used widely and cheaply and makes EVERYTHING virtually instantaneous.

[kevlarman]

It's unfortunate, but the restriction of the speed packets come through is a law of the internet in itself, and will not be fixed until someone comes up with something that can be used widely and cheaply and makes EVERYTHING virtually instantaneous.

information can't travel faster than c

[Paradox]

lol@kev

[techhead]

Do as-yet-on-the-horizon quantum entanglement modems count?

EDIT
Retraction:
The prior statement is false, as the tech is not on the horizon, because it is impossible to exist.
Moral:
Research first, wise-crack second.

[blood2.0]

i sometimes get it where the shotgun doesn't make any sound its really annoying as a  marauder when your heath goes down by 50 points and you don't here anything. are you sure shotguns don't have a delay because i swear they take a second to fire once i click and that doesn't happen with any other weapon.

[Nux]

In my experience, if the shotgun doesn't make a sound that means you're dead. So I wouldn't worry about it too much.

As for this faster than light business, though this is true for local velocity, there may still be a way to bridge distances without breaking the universal speed limit. Think of your computer and my computer being partially inside of each other. >.> Plus the very small scale is only upwardly limited through tendency (as far as is measurable). So such limits may only hold due to unlikelihood rather than impossibility.

[Le Compilateur]

I can never hit anything with the shotty in lagged or unlagged, that could be because of the way the engine works or just because I have incredibly crappy aim.

[Dracone]

Well it's all about speed. There just needs to be a fast enough connection that everyone has a reasonable, good ping, no matter how far from the server you are. Services meeting such a standard are far from available commercially, not even CLOSE. It will be a long time before such things are available. The likelihood is that it will NEVER happen.

[techhead]

FTL communications violate causality, so disprove that before you build yourself an Ansible.
That said, due to refraction in optical fiber, light travels ~30% slower, giving a NYC-Australia ping of about .18s.
That means no pings dropping below 180 from across the world until we either thread optical fiber through the earth's core OR invent aforementioned FTL communications. If you want to poke at how it COULD be possible, look here.

[Nux]

Why dismiss something because of a little thing like causality >.> It's only generally believed to hold, not so easily verifiable.

You can invoke Novikov (Novikov self-consistency principle) or maybe even move that travelling backward in time doesn't necessarily mean traversing the same path (just look at Feynman diagrams).

I'm not saying we're living in a self sustained time loop or if we are we could use it in anyway. Neither am I saying that travelling as antimatter will be too enjoyable after that nasty collision. Hell, if you want a safe way to travel back, why don't we just look to CPT symmetry and have fun walking backward to confuse the antimatter people.

Oh damn.. I think I'm derailing my own thread.