Forums » Suggestions
Lag switch counter
My suggestion is to instead of making the ship jump when they lag spike (as in teleporting a few hundred meters in a different direction) make it so that the ship is ALWAYS where it was last updated. Like instead of the lag teleporting they don't move at all, and they continue on the course they were on to start with. So instead of the abuse of turning left, then making yourself lag then turning right, you continue left. Always staying on that course. I don't know how to completely deal with this or put it into words. But anything is better then this abuse of the ability to just teleport hundreds sometimes a thousand or two in a random direction and getting away. This makes it so much more difficult to hit or kill the player. Any other ideas on this are much welcome.
i think the devs didnt want to use server-enforced positioning, and explained in a previous thread here - I'll see if I can find it...
Ah okay, so let's just let the people with really bad ping be nearly invincible and impossible to chase. Gotcha
My position on this is that it should be contextual. The game could detect whether you're in proximity to other players and base the behavior it uses off of that. If you're just out doing PVE on a shitty connection, it would treat the client-reported position as canonical (within reason) so that you can have a nice experience. But if you encounter other players, it would switch behaviors and assign more weight to the server's notion of your position to avoid giving you an exploitable advantage.
For the sake of group PVE, it could also recognize if the people around you are in your group, guild, or buddy list so that they could be excepted from the "in proximity to others" test.
For the sake of group PVE, it could also recognize if the people around you are in your group, guild, or buddy list so that they could be excepted from the "in proximity to others" test.
Yup. I've previously responded in other threads with basically what Pizzasgood is describing.
The challenge is that the implementation requires not just significant changes to the protocol, but also a lot of heuristics that would need to be tuned and tested fairly extensively. It's a pretty big, time-consuming project to take on. The recent addition of fast-loading sectors, while a substantial change to the protocol itself, was much simpler than what this would entail.
Ah okay, so let's just let the people with really bad ping be nearly invincible and impossible to chase. Gotcha
Sarcastic, passive-aggressive responses like this do not belong on Suggestions. Luxen was just trying to offer the best of his recollection, while looking for my response. If you post something, expect feedback, and welcome it. Don't just spout straw-man nonsense.
If this were a trivial problem to fix, we would have done it already.
The challenge is that the implementation requires not just significant changes to the protocol, but also a lot of heuristics that would need to be tuned and tested fairly extensively. It's a pretty big, time-consuming project to take on. The recent addition of fast-loading sectors, while a substantial change to the protocol itself, was much simpler than what this would entail.
Ah okay, so let's just let the people with really bad ping be nearly invincible and impossible to chase. Gotcha
Sarcastic, passive-aggressive responses like this do not belong on Suggestions. Luxen was just trying to offer the best of his recollection, while looking for my response. If you post something, expect feedback, and welcome it. Don't just spout straw-man nonsense.
If this were a trivial problem to fix, we would have done it already.
This may be over simplifying things ,infact i'm sure it is, but could you not just set a limit on server ping
above which you can only damage bots
so if superdeadlylagfiend jumps in with 995 ping his shots dont do damage to players
above which you can only damage bots
so if superdeadlylagfiend jumps in with 995 ping his shots dont do damage to players
But the problem is generally not people with statically high latency. It's usually variable latency. Or latency and packet loss. Which gets into more heuristics about "how high is too high, and what does that mean?".
A 200ms average latency with high variability or packet loss is likely much worse, from a visual tracking standpoint, than someone with 400ms latency and great consistency.
Again, if this were trivial to fix, we would have done so a long time ago. The simplest solutions, like the well-intended OP, would actually make things much worse, for everyone, most of the time.
Understand, I'm not against continued ideas and feedback, that's all cool.. I'm not trying to shut down the conversation. I'm just mitigating expectations.
We do want to improve this, there are a lot of things I want to do with the Duel system, for instance, that really needs improvements in this area. But, well, for better worse I haven't been willing to jump off that particular cliff.
It's not that I don't think it's important.. "really important", even. But, like always, there are a lot of different "Really Important Things", depending on who you ask.
A 200ms average latency with high variability or packet loss is likely much worse, from a visual tracking standpoint, than someone with 400ms latency and great consistency.
Again, if this were trivial to fix, we would have done so a long time ago. The simplest solutions, like the well-intended OP, would actually make things much worse, for everyone, most of the time.
Understand, I'm not against continued ideas and feedback, that's all cool.. I'm not trying to shut down the conversation. I'm just mitigating expectations.
We do want to improve this, there are a lot of things I want to do with the Duel system, for instance, that really needs improvements in this area. But, well, for better worse I haven't been willing to jump off that particular cliff.
It's not that I don't think it's important.. "really important", even. But, like always, there are a lot of different "Really Important Things", depending on who you ask.
Tracking those who have *consistent* lag spikes just prior to teleporting out of a sector (i.e. under 5 seconds before jumping) seems less complicated. Lagging prior to jump is a classic lagswitch exploit that is, for me at least, the most annoying case of this.
If someone starts lagging out during a normal fight, big deal, just leave the fight.
Example:
If someone displays greater than a 100ms latency difference up to 5 seconds prior to, or during the jumping out of a sector in combination with being within 600m of another player at the same time, more than 3 times in a 30 minute period, then they should have their asses handed a penalty of some sort.
I would think that the consistent circumstances represented above are specific enough to weed out the accidental laggers with questionable internet service.
If someone starts lagging out during a normal fight, big deal, just leave the fight.
Example:
If someone displays greater than a 100ms latency difference up to 5 seconds prior to, or during the jumping out of a sector in combination with being within 600m of another player at the same time, more than 3 times in a 30 minute period, then they should have their asses handed a penalty of some sort.
I would think that the consistent circumstances represented above are specific enough to weed out the accidental laggers with questionable internet service.
"This may be over simplifying things ,infact i'm sure it is, but could you not just set a limit on server ping
above which you can only damage bots"
That misses the point. Nobody's worried about a laggy person hurting them. They're worried about being able to hurt the laggy person.
above which you can only damage bots"
That misses the point. Nobody's worried about a laggy person hurting them. They're worried about being able to hurt the laggy person.
If someone displays greater than a 100ms latency difference up to 5 seconds prior to, or during the jumping out of a sector in combination with being within 600m of another player at the same time, more than 3 times in a 30 minute period, then they should have their asses handed a penalty of some sort.
That's actually a substantial thing to implement, just for the possibility of improving an extremely narrow set of cases. For instance, this would do nothing to mitigate issues during duels, which is a blocker to my being able to build a robust duel-ladder and championship system.
The "false-positive" rate might also be a lot worse than you expect. 100ms of variation isn't much. Even US and European providers are overloaded right now, and people on mobile are often enduring a lot of transient events (especially those who are geographically far from the servers).
It also only addresses latency, and not packet loss, where the fundamental issue is retransmission time. What makes you think your target "cheaters" are just influencing packet latency?
What you're describing doesn't just implement minimized, general-purpose, context-based preference of server-predicted position (like Pizza and I have been referencing). You're talking about actually penalizing people in some external way, based on detection criteria that has a strong chance of being flawed.
Look, I'm not against the idea of Silly Hacks. We do Silly Hacks all day long, and obviously anyone who has seen our development or response to various issues has witnessed them countless times.
But this isn't a Silly Hack problem, or we would have done that already.
It needs to be approached carefully, and done with a lot of involved testing, and it needs to be the broadest possible solution, not just to address current cases that happen to bother people right now, but future issues that are blocking large areas of gameplay development. It's going to be time-consuming to fix.
If one doesn't approach this properly, it's easy to waste a ton of development time.. making the game worse.
That's actually a substantial thing to implement, just for the possibility of improving an extremely narrow set of cases. For instance, this would do nothing to mitigate issues during duels, which is a blocker to my being able to build a robust duel-ladder and championship system.
The "false-positive" rate might also be a lot worse than you expect. 100ms of variation isn't much. Even US and European providers are overloaded right now, and people on mobile are often enduring a lot of transient events (especially those who are geographically far from the servers).
It also only addresses latency, and not packet loss, where the fundamental issue is retransmission time. What makes you think your target "cheaters" are just influencing packet latency?
What you're describing doesn't just implement minimized, general-purpose, context-based preference of server-predicted position (like Pizza and I have been referencing). You're talking about actually penalizing people in some external way, based on detection criteria that has a strong chance of being flawed.
Look, I'm not against the idea of Silly Hacks. We do Silly Hacks all day long, and obviously anyone who has seen our development or response to various issues has witnessed them countless times.
But this isn't a Silly Hack problem, or we would have done that already.
It needs to be approached carefully, and done with a lot of involved testing, and it needs to be the broadest possible solution, not just to address current cases that happen to bother people right now, but future issues that are blocking large areas of gameplay development. It's going to be time-consuming to fix.
If one doesn't approach this properly, it's easy to waste a ton of development time.. making the game worse.