7 ms·

The real solution? Send only information players are supposed to see. Sacrifice some speed for true cheat proof games.

by Aperocky 7y ago

dharmab 7y ago
That wouldn't prevent commonly used cheats such as aimbots.
- Aperocky 7y ago
  You’re right.
- AnthonyMouse 7y ago
  Maybe the solution there is to stop making games be about the things computers are better at. They're better at aiming, fine, let them be. Put the aimbot in the game from the start. Then instead you have a FPS with multiple weapons and it takes several seconds to switch between them. One has a high rate of fire but is useless against players with armor, another is the opposite. Cheaters don't know if the opponent is wearing armor (or what kind) because the server doesn't tell them that until they're in the same room, by which point it's too late. Make certain weapons better against certain defenses, so it's a matter of predicting what your opponent will do and being ready for it. Make the humans do the things that humans are still better at, instead of trying to prevent a computer from beating them at things computers are better at.
  - spathi_fwiffo 7y ago
    first person/multi-player rock paper scissors. would be an interesting game. probably family friendly. should make it.
    - AnthonyMouse 7y ago
      It doesn't even have to be that arbitrary. You've got an opponent with a defense against all your current weapons, but you can run faster than them, so you can escape but you still need to find something that can kill them. They need to find something that can make them faster so they can catch you, or take their chances by giving up the defense which is slowing them down. There can still be strategy in it without leaving any easy way to cheat.
  - Aperocky 7y ago
    We don’t even need those fancy things, just the angle of view. If someone come up behind without being seen they can still knife the player. Alternatively, the aiming would be slower if you are looking at the peripheral.
  - m11a 7y ago
    You can't tell people to play something else other than what they enjoy just because it'll "make cheating harder". Many of the very popular games now require good aim. Some people spend countless hundreds/thousands of hours just practicing their aim and muscle memory. It's part of the fun for them. CS:GO is a notable example. The problem isn't really that it's impossible to ban such cheating. ESEA do it pretty well in CS:GO - their anti-cheat is practically a RAT. It's that banning cheaters and making a bulletproof anticheat has two issues: 1) To do it properly you need to basically code a RAT, which has some privacy issues. If Valve do it, it could hurt activity, platform support, and be controversial. ESEA have done it since day one, but their player base is much smaller and more competitive. ESEA is also limited to Windows. 2) It's not profitable. That's the honest truth. They ban obvious public cheats and the cheaters will buy another account and come back, probably use another cheat again, get banned again, repeat. Some games don't have awful cheat protection because the company can't make a good one, they have awful cheat protection because it's not profitable to alienate all the cheaters in the game away from the platform (and their many purchases per person; legitimate players only buy once).
    - AnthonyMouse 7y ago
      It's somewhat worse than that, because anti-cheat and anti-anti-cheat are basically two things in ring-0 fighting each other. So it's a resource war, whoever spends more development time trying to win does. But the economies of scale favor the cheaters. If only one game has working anti-cheat, it may not be worth the cheater's time to break it because it's only one game, and they can go cheat in a different game. If all the games had it then more people would try to break it and more would succeed. And once they have a working method, it can be employed against any other game that uses the same kind of anti-cheating, even if it's not the same vendor's, as long as the mechanism is similar. But having all the vendors get together to pool resources and design one good anti-cheating software makes it even worse, because then it might be a little harder to defeat, but the prize for defeating it is being able to cheat in every game without even having to translate the cheating method to each new game. So you get small victories or nothing. There is no winning at scale because increasing scale only increases the number of adversaries and the gains from finding an effective way to cheat. It's basically the same problem as keeping the latest Hollywood releases off of pirate sites. They count success as occasionally having a week before a new release is on the pirate sites instead of being there the same day, and even that doesn't happen very often. It turns out it's more effective to make it so it doesn't matter. In their case by making it so easy and cheap to watch Netflix on your TV that it's not worth most peoples' time to figure out how to use BitTorrent even if it has 100% of their content. In this case by making more games that humans are better at than computers, so you can't use a computer to cheat.
  - meheleventyone 7y ago
    Many more people will play the fun game with cheating than will play the game you make to be “secure”. In your game design not only is it dull because you have a 50:50 chance of having the right gun equipped to take advantage of a situation even if no one is cheating. But it’s still very amenable to cheating both because the cheat would be able to switch weapon the instant the client was informed of the state of the opponent, beating a human where it actually matters and would still be able to have a mechanically perfect aim. Most videogames actually have the sort of human reaction prediction you are talking about and that’s why people cheat because they are bad at it. So make it up by exploiting information sent but not displayed or through mechanical assistance. The interesting case there is that mechanically good players that have a good read can beat cheaters. Stuff like the DeepMind successes playing Chess, Go and Starcraft ultimately show that even with an environment where things are fairly level that computers have the edge in games albeit with an unattainable for cheat makers level of technology.
    - AnthonyMouse 7y ago
      > In your game design not only is it dull because you have a 50:50 chance of having the right gun equipped to take advantage of a situation even if no one is cheating. Except that it's not random, because you can try to figure it out. You know what's in the map. If you go to where a particular weapon is and it's not there, now you can expect the opponent to have it, so you start looking for defenses against it etc. Or you know that you're playing against Bob and Bob always goes for the same weapon like a chump because he doesn't know what to do with any of the others. Or you get a weapon and then go around collecting every defense against it on the map so the opponent can't get them. > But it’s still very amenable to cheating both because the cheat would be able to switch weapon the instant the client was informed of the state of the opponent, beating a human where it actually matters and would still be able to have a mechanically perfect aim. Which is why it's not about switching faster, it's about making the right choice to begin with. The whole idea is if you have to switch at all you're probably already dead and switching a half second faster isn't going to save you.
      - meheleventyone 7y ago
- LoSboccacc 7y ago
  if your turn speed is limited and the server only send you precise enemy position information within your field of view aimbot can only help so much, having a drop in the enemy will enable better player to overcome aimbotters war thunder does it and it works well enough. it still has other problems, like p2w vehicles, but the system is solid
- drdaeman 7y ago
  > Then you have the mechanical helpers like aimbots I wonder if anyone had ever designed a completely external aimbot. A camera, a computer vision system and a hardware mouse or gamepad multiplexer/proxy that mixes real controls with the adjusting movements. That won't be fun to use (except, maybe, for PvE, if manual dexterity is just a irritating mechanical hurdle - there are games like that, too), but surely fun to hack. Heh. Sometimes I regret that I won't live long enough to see all the drama around cyborgs in sports and gaming...
spathi_fwiffo 7y ago
This would require a hardened OS and is outside the control of the game creator. Transparent walls, for example, can be done via the graphics driver (or by replacing your graphics driver).
- EamonnMR 7y ago
  I think what the parent means is that if a player is behind a wall, don't send that player's position at all. Of course you could still have a hack which, say, visualizes sound origins.
- mottosso 7y ago
  Unless the server truly sent only what the user is meant to see - i.e. video; no hacks can reveal more than what's in a grid of pixels.
- derefr 7y ago
  No, they’re talking about writing netcode in such a way that if you report to the server that you’re in position XYZ, then it’ll cull all player-events invisible to players at position XYZ out of the event stream being sent to your client.
  - spathi_fwiffo 7y ago
    I actually wonder why this isn't done. Pre-rendering the scenes on the server and then sending those is just not really possible today due to latency -- even google can't get it right with Stadia so far. But, if the game engine/server just crunched the line-of-sight numbers and didn't even send location data for rendering enemies that were out of LOS, then that would work and while it is heavier than doing nothing, it isn't really that heavy. That wouldn't stop an aimbot, but it would prevent some classes of exploits.
    - coldpie 7y ago
      It's not about server-side computation resources or bandwidth, it's about latency. If it takes 50ms to communicate your position to the server and 50ms for them to reply, that's 1/10 of a second for every position update, which is an eternity in high-level play. Netcode is really hard, anti-cheat is really hard, and mashing them together is really, really hard :)
SeanBoocock 7y ago
That’s already often done at some level, but doesn’t eliminate the ability to cheat via aim bots (and other similar tools). I have worked on several server authoritative multiplayer games and while they were mostly immune to common types of cheating, there are always things that people can exploit/automate to their advantage.
kgwxd 7y ago
Game and console companies would love it if users only had dumb terminals with DRM protected input and output, and they're working towards that dream: https://www.ign.com/articles/2019/03/19/e3-2018-the-rise-of-streamed-gaming-is-2018s-scariest-new-trend https://www.ign.com/articles/2019/03/19/e3-2018-the-rise-of-...
int_19h 7y ago
You have to send more than that for the game to be playable, if your game involves fast movement or turning (which is pretty much any FPS). Otherwise the player would see empty spaces after e.g. turning around, before the server can catch up with the movement. This problem is as old as competitive multiplayer itself, and various approaches have all been tried.

The real solution? Send only information players are supposed to see. Sacrifice some speed for true cheat proof games.

dharmab 7y ago

That wouldn't prevent commonly used cheats such as aimbots.

Aperocky 7y ago
You’re right.
AnthonyMouse 7y ago
Maybe the solution there is to stop making games be about the things computers are better at. They're better at aiming, fine, let them be. Put the aimbot in the game from the start. Then instead you have a FPS with multiple weapons and it takes several seconds to switch between them. One has a high rate of fire but is useless against players with armor, another is the opposite. Cheaters don't know if the opponent is wearing armor (or what kind) because the server doesn't tell them that until they're in the same room, by which point it's too late. Make certain weapons better against certain defenses, so it's a matter of predicting what your opponent will do and being ready for it. Make the humans do the things that humans are still better at, instead of trying to prevent a computer from beating them at things computers are better at.
- spathi_fwiffo 7y ago
  first person/multi-player rock paper scissors. would be an interesting game. probably family friendly. should make it.
  - AnthonyMouse 7y ago
    It doesn't even have to be that arbitrary. You've got an opponent with a defense against all your current weapons, but you can run faster than them, so you can escape but you still need to find something that can kill them. They need to find something that can make them faster so they can catch you, or take their chances by giving up the defense which is slowing them down. There can still be strategy in it without leaving any easy way to cheat.
- Aperocky 7y ago
  We don’t even need those fancy things, just the angle of view. If someone come up behind without being seen they can still knife the player. Alternatively, the aiming would be slower if you are looking at the peripheral.
- m11a 7y ago
  You can't tell people to play something else other than what they enjoy just because it'll "make cheating harder". Many of the very popular games now require good aim. Some people spend countless hundreds/thousands of hours just practicing their aim and muscle memory. It's part of the fun for them. CS:GO is a notable example. The problem isn't really that it's impossible to ban such cheating. ESEA do it pretty well in CS:GO - their anti-cheat is practically a RAT. It's that banning cheaters and making a bulletproof anticheat has two issues: 1) To do it properly you need to basically code a RAT, which has some privacy issues. If Valve do it, it could hurt activity, platform support, and be controversial. ESEA have done it since day one, but their player base is much smaller and more competitive. ESEA is also limited to Windows. 2) It's not profitable. That's the honest truth. They ban obvious public cheats and the cheaters will buy another account and come back, probably use another cheat again, get banned again, repeat. Some games don't have awful cheat protection because the company can't make a good one, they have awful cheat protection because it's not profitable to alienate all the cheaters in the game away from the platform (and their many purchases per person; legitimate players only buy once).
  - AnthonyMouse 7y ago
    It's somewhat worse than that, because anti-cheat and anti-anti-cheat are basically two things in ring-0 fighting each other. So it's a resource war, whoever spends more development time trying to win does. But the economies of scale favor the cheaters. If only one game has working anti-cheat, it may not be worth the cheater's time to break it because it's only one game, and they can go cheat in a different game. If all the games had it then more people would try to break it and more would succeed. And once they have a working method, it can be employed against any other game that uses the same kind of anti-cheating, even if it's not the same vendor's, as long as the mechanism is similar. But having all the vendors get together to pool resources and design one good anti-cheating software makes it even worse, because then it might be a little harder to defeat, but the prize for defeating it is being able to cheat in every game without even having to translate the cheating method to each new game. So you get small victories or nothing. There is no winning at scale because increasing scale only increases the number of adversaries and the gains from finding an effective way to cheat. It's basically the same problem as keeping the latest Hollywood releases off of pirate sites. They count success as occasionally having a week before a new release is on the pirate sites instead of being there the same day, and even that doesn't happen very often. It turns out it's more effective to make it so it doesn't matter. In their case by making it so easy and cheap to watch Netflix on your TV that it's not worth most peoples' time to figure out how to use BitTorrent even if it has 100% of their content. In this case by making more games that humans are better at than computers, so you can't use a computer to cheat.
- meheleventyone 7y ago
  Many more people will play the fun game with cheating than will play the game you make to be “secure”. In your game design not only is it dull because you have a 50:50 chance of having the right gun equipped to take advantage of a situation even if no one is cheating. But it’s still very amenable to cheating both because the cheat would be able to switch weapon the instant the client was informed of the state of the opponent, beating a human where it actually matters and would still be able to have a mechanically perfect aim. Most videogames actually have the sort of human reaction prediction you are talking about and that’s why people cheat because they are bad at it. So make it up by exploiting information sent but not displayed or through mechanical assistance. The interesting case there is that mechanically good players that have a good read can beat cheaters. Stuff like the DeepMind successes playing Chess, Go and Starcraft ultimately show that even with an environment where things are fairly level that computers have the edge in games albeit with an unattainable for cheat makers level of technology.
  - AnthonyMouse 7y ago
    > In your game design not only is it dull because you have a 50:50 chance of having the right gun equipped to take advantage of a situation even if no one is cheating. Except that it's not random, because you can try to figure it out. You know what's in the map. If you go to where a particular weapon is and it's not there, now you can expect the opponent to have it, so you start looking for defenses against it etc. Or you know that you're playing against Bob and Bob always goes for the same weapon like a chump because he doesn't know what to do with any of the others. Or you get a weapon and then go around collecting every defense against it on the map so the opponent can't get them. > But it’s still very amenable to cheating both because the cheat would be able to switch weapon the instant the client was informed of the state of the opponent, beating a human where it actually matters and would still be able to have a mechanically perfect aim. Which is why it's not about switching faster, it's about making the right choice to begin with. The whole idea is if you have to switch at all you're probably already dead and switching a half second faster isn't going to save you.
    - meheleventyone 7y ago
LoSboccacc 7y ago
if your turn speed is limited and the server only send you precise enemy position information within your field of view aimbot can only help so much, having a drop in the enemy will enable better player to overcome aimbotters war thunder does it and it works well enough. it still has other problems, like p2w vehicles, but the system is solid
drdaeman 7y ago
> Then you have the mechanical helpers like aimbots I wonder if anyone had ever designed a completely external aimbot. A camera, a computer vision system and a hardware mouse or gamepad multiplexer/proxy that mixes real controls with the adjusting movements. That won't be fun to use (except, maybe, for PvE, if manual dexterity is just a irritating mechanical hurdle - there are games like that, too), but surely fun to hack. Heh. Sometimes I regret that I won't live long enough to see all the drama around cyborgs in sports and gaming...

spathi_fwiffo 7y ago

This would require a hardened OS and is outside the control of the game creator. Transparent walls, for example, can be done via the graphics driver (or by replacing your graphics driver).

EamonnMR 7y ago
I think what the parent means is that if a player is behind a wall, don't send that player's position at all. Of course you could still have a hack which, say, visualizes sound origins.
mottosso 7y ago
Unless the server truly sent only what the user is meant to see - i.e. video; no hacks can reveal more than what's in a grid of pixels.
derefr 7y ago
No, they’re talking about writing netcode in such a way that if you report to the server that you’re in position XYZ, then it’ll cull all player-events invisible to players at position XYZ out of the event stream being sent to your client.
- spathi_fwiffo 7y ago
  I actually wonder why this isn't done. Pre-rendering the scenes on the server and then sending those is just not really possible today due to latency -- even google can't get it right with Stadia so far. But, if the game engine/server just crunched the line-of-sight numbers and didn't even send location data for rendering enemies that were out of LOS, then that would work and while it is heavier than doing nothing, it isn't really that heavy. That wouldn't stop an aimbot, but it would prevent some classes of exploits.
  - coldpie 7y ago
    It's not about server-side computation resources or bandwidth, it's about latency. If it takes 50ms to communicate your position to the server and 50ms for them to reply, that's 1/10 of a second for every position update, which is an eternity in high-level play. Netcode is really hard, anti-cheat is really hard, and mashing them together is really, really hard :)

SeanBoocock 7y ago

That’s already often done at some level, but doesn’t eliminate the ability to cheat via aim bots (and other similar tools). I have worked on several server authoritative multiplayer games and while they were mostly immune to common types of cheating, there are always things that people can exploit/automate to their advantage.

kgwxd 7y ago

Game and console companies would love it if users only had dumb terminals with DRM protected input and output, and they're working towards that dream: https://www.ign.com/articles/2019/03/19/e3-2018-the-rise-of-streamed-gaming-is-2018s-scariest-new-trend https://www.ign.com/articles/2019/03/19/e3-2018-the-rise-of-...

int_19h 7y ago

You have to send more than that for the game to be playable, if your game involves fast movement or turning (which is pretty much any FPS). Otherwise the player would see empty spaces after e.g. turning around, before the server can catch up with the movement. This problem is as old as competitive multiplayer itself, and various approaches have all been tried.