|
On June 06 2012 04:07 mastergriggy wrote:It seems like after every shitstorm a brave community member comes forward with a great answer/fix. After MKP vs. Parting, we got the resume game from replay device, and now we have this  Nicely done. But nothing's been done ...
|
On June 06 2012 05:05 ZeromuS wrote:Show nested quote +On June 06 2012 01:19 Heh_ wrote:Sounds like a good plan, but there's something to consider about looking at fog of war: 1: If a player scans a base, briefly looks and it, and comes back a minute later to see what he might have missed/count buildings, then these instances should be highlighted but not scream "maphack". 2: Looking at base locations to shift-click a scouting worker shouldn't be highlighted. Some people prefer to click on a screen instead of the minimap. 3: Scouting for proxies in the normal spots. 4: What should definitely be highlighted is when the player camera goes over a random area that has "nothing", but is hiding a proxy pylon or tech. How to fix this? For point 1 I don't have a good idea. You just have to flag it up and analyze manually. For point 2 and 3, I would suggest to ignore the instance if it is followed up by a click (or shift click) within 2 seconds of vision in that area. That way, you won't have a million false positives showing up when a player decides to spam shift click for legitimate scouting. Edit: On June 06 2012 01:19 ZeromuS wrote: I posted this in the other thread but here it goes.
In MOHAA there was an anti hack that would screen cap randomly. If you were suspected of hacking you would be given 24 hours to email an admin the screenshot collection and if hacks were found you would be banned, if none no ban and if no Emil you would get banned as well. Worked pretty well and it can work for tournaments run online only for sc2 idf there was a similar application. You could run an overlay that's recording your game with "normal vision", so the hack can only be seen by you, making the screenshot worthless. Not necessarily, it doesn't have to screen cap sc2 but instead everything displayed on the desktop like hitting print screen on your keyboard.
There's something called privacy, i don't know if you could possibly do that and capture every single window the user has opened..
|
Do whatever it takes. It's up to the community to devise a solution. If Blizzard has a problem with it, then either it'll spur them to do something themselves, or give the community leverage to protest.
|
|
Basically it'll be awesome so long as you can get Blizzard's support behind it.
|
This sound great, as long as it can work and blizz doesn't herp derp over it <3 gogogogogoggo
|
Any CS majors in here?
Download many maphacks. Play a bunch of games on them using every effort to hide that you are, but still trying to win. Train a machine learning system on the maphack replays.
I wouldn't be shocked if there is some tipoff that could be 100% accurate since the maphack is interacting with the keys stores in the replay in order to do the view freezes and hide looking into the fog.
|
Would be cool to see this in daily and important online tournaments, as well as streams having a 5-10min delay.
|
I've developed a prototype application. It currently successfully detects all the existing hacks that I've tested it with. It also supports signature recognition of applications running on the system, which will allow blacklisting of hacks.
I've also developed a counterpart to the application which is an off-site service that performs analysis on replays post-game. The off-site service currently successfully detects Auto-Blink, Auto-Burrow, Auto-Inject and Screen blocking maphacks. If you know of other hacks that are evident in replays, and you have replays of them, please post them here or PM them to me.
Both players in a game running the application would ensure a hack free game. One player in a game running the application would still provide a decent detection of hacks through post-game analysis.
I plan to develop a user friendly version of these applications ASAP.
Thoughts?
|
On June 06 2012 06:20 JackDT wrote: Any CS majors in here?
Download many maphacks. Play a bunch of games on them using every effort to hide that you are, but still trying to win. Train a machine learning system on the maphack replays.
I wouldn't be shocked if there is some tipoff that could be 100% accurate since the maphack is interacting with the keys stores in the replay in order to do the view freezes and hide looking into the fog.
Very possible.
But I think a smarter idea would be, make a really amazing maphack. (LeMap) and have it harvest replays from the hackers. That way your information set would be all races, leagues, and thousands of wins and losses.
|
On June 10 2012 04:45 Veritas wrote:I've developed a prototype application. It currently successfully detects all the existing hacks that I've tested it with. It also supports signature recognition of applications running on the system, which will allow blacklisting of hacks. I've also developed a counterpart to the application which is an off-site service that performs analysis on replays post-game. The off-site service currently successfully detects Auto-Blink, Auto-Burrow, Auto-Inject and Screen blocking maphacks. If you know of other hacks that are evident in replays, and you have replays of them, please post them here or PM them to me. Both players in a game running the application would ensure a hack free game. One player in a game running the application would still provide a decent detection of hacks through post-game analysis. I plan to develop a user friendly version of these applications ASAP. Thoughts?  Can your program test replays, or is it only useful in-game?
|
Amazing work, OP. I and I'm sure many others would be willing to donate to this project if you create a tip jar or something. That's assuming Blizz won't object to it and tournaments will utilize it.
Keep it up.
|
On June 10 2012 05:44 Nightmarjoo wrote:Show nested quote +On June 10 2012 04:45 Veritas wrote:I've developed a prototype application. It currently successfully detects all the existing hacks that I've tested it with. It also supports signature recognition of applications running on the system, which will allow blacklisting of hacks. I've also developed a counterpart to the application which is an off-site service that performs analysis on replays post-game. The off-site service currently successfully detects Auto-Blink, Auto-Burrow, Auto-Inject and Screen blocking maphacks. If you know of other hacks that are evident in replays, and you have replays of them, please post them here or PM them to me. Both players in a game running the application would ensure a hack free game. One player in a game running the application would still provide a decent detection of hacks through post-game analysis. I plan to develop a user friendly version of these applications ASAP. Thoughts?  Can your program test replays, or is it only useful in-game?
The application will perform replay analysis by passing replays to the remote service after each game.
I could add replay processing to the application, but if it were to be cracked somehow then the detection patterns would be available to the hacker community.
On June 10 2012 06:13 Doodsmack wrote: Amazing work, OP. I and I'm sure many others would be willing to donate to this project if you create a tip jar or something. That's assuming Blizz won't object to it and tournaments will utilize it.
Keep it up.
Even if tournaments don't use it, I think it will still be useful for people playing in tournaments to have the option to be alerted instantly if their opponent is acting suspicious.
|
Well specifically I was asking to see if you could support or contradict the claim that Spades hacked in his showmatch against Lucifron with your program.
|
Let's say I send you a 25 replays package, with some people I know hacked in those and some legit players, you would be able to find who's legit and who's not without error? If so, can we try this out... and if that end's up working i'd need your help in pinpointing hackers if your willing to in here: http://www.teamliquid.net/forum/viewmessage.php?topic_id=340614
|
Blizzard should hire you to do their job. 
But seriously, thank you for being awesome OP. I hope this, in combination with stream delay, helps to legitimize online tournaments.
|
Great work keep it up. Thanks for putting the time and effort into finding a solution for this.
How can anybody even post something pessimistic before thanking the OP? If only the World was filled with all you sad sacks; nothing would get done.
|
Go for it, better anticheat is always a good thing.
On June 06 2012 00:51 Nightwatch wrote: Many people probably don't want to play tournaments if they have to download a software. 100% not true, many tournaments/ladders for games have you download programs that check for cheats or help log games. If this comes out and is decent, it will get picked up by any tournament wanting to improve their image. I'd say it would actually draw people to the tournament.
|
On June 10 2012 12:12 Nightmarjoo wrote: Well specifically I was asking to see if you could support or contradict the claim that Spades hacked in his showmatch against Lucifron with your program.
The anti-cheat monitor is the only fool-proof way of automatic detection. The replay analysis is meant to be a guide, nothing more. No automated replay analysis system will ever be without error. It is meant to inform users or admins that they may need to view a replay. Only significantly high counts of activity should be considered conclusive (e.g., 20+ suspected blink hack instances).
My current replay patterns only detect a few things. However, the Spades replays flagged more screen block warnings than 50 APM platinum replays. On average there were at least 8 points in each game where he stared at a single spot in his base for longer than 10 sconds. Beyond that I cannot say until it's more advanced.
On June 10 2012 13:34 ZweiGaming wrote:Let's say I send you a 25 replays package, with some people I know hacked in those and some legit players, you would be able to find who's legit and who's not without error? If so, can we try this out... and if that end's up working i'd need your help in pinpointing hackers if your willing to in here: http://www.teamliquid.net/forum/viewmessage.php?topic_id=340614
I'd be happy to give it a try. But only the things I mentioned a few posts ago are checked right now. If you have replays of hacks that can be detected via replay analysis then I'd love to have your help making pattern detectors for them.
|
On June 10 2012 14:16 Veritas wrote:Show nested quote +On June 10 2012 12:12 Nightmarjoo wrote: Well specifically I was asking to see if you could support or contradict the claim that Spades hacked in his showmatch against Lucifron with your program. The anti-cheat monitor is the only fool-proof way of automatic detection. The replay analysis is meant to be a guide, nothing more. No automated replay analysis system will ever be without error. It is meant to inform users or admins that they may need to view a replay. Only significantly high counts of activity should be considered conclusive (e.g., 20+ suspected blink hack instances). My current replay patterns only detect a few things. However, the Spades replays flagged more screen block warnings than 50 APM platinum replays. On average there were at least 8 points in each game where he stared at a single spot in his base for longer than 10 sconds. Beyond that I cannot say until it's more advanced. Show nested quote +On June 10 2012 13:34 ZweiGaming wrote:Let's say I send you a 25 replays package, with some people I know hacked in those and some legit players, you would be able to find who's legit and who's not without error? If so, can we try this out... and if that end's up working i'd need your help in pinpointing hackers if your willing to in here: http://www.teamliquid.net/forum/viewmessage.php?topic_id=340614 I'd be happy to give it a try. But only the things I mentioned a few posts ago are checked right now. If you have replays of hacks that can be detected via replay analysis then I'd love to have your help making pattern detectors for them.
Thanks a lot, i'll be sending you few replays of hackers and some other of people we suspect via PM!
|
|
|
|