That's a good point. If someone finds memory addresses that these cheats alter, PLEASE REPORT THEM TO US. That can speed up the fixing process by a lot. We aren't so good at tracking these values down ourselves unfortunately.
That's a good point. If someone finds memory addresses that these cheats alter, PLEASE REPORT THEM TO US. That can speed up the fixing process by a lot. We aren't so good at tracking these values down ourselves unfortunately.
That doesn't surprise me in the slightest.
Just ban everyone you see using it (cheating) on your server, that's all you can do for the moment i think....
yes and when they get enough bans they will automatically receive a global ban of 30 days as well. Mind you this only works if you ban the person while he's on the host. If you apply a ban on someone while he/she is not on the host the ban will not count towards the global ban.
Hm.. didn't know that, but it's nice that a single host ban isn't totally useless for helping the rest getting rid of cheaters.
It is surely easy just to find the top speed of each car, and have an InSim app running to ban anyone who goes over that speed? Sure, there will be those who use the cheat in moderation, and will have to be dealt with manually...or else find a way to have acceleration calculated in the InSim app, taking into account lag and MCI intervals 
If the majority of servers were running this app, then pretty soon these "people" would have nowhere to go. Whether or not they get a global ban is irrelevant, being stuck with "Teh Dr1fT0r2" and the rest of the demo servers is just as bad
If the majority of servers were running this app, then pretty soon these "people" would have nowhere to go. Whether or not they get a global ban is irrelevant, being stuck with "Teh Dr1fT0r2" and the rest of the demo servers is just as bad
Rob, Its not easy atall. You cant just moniter top speed, Say for example they go on a track like aston with fast straights. That can be fast speed. Then they goto a track like South City. They will have complety different top speeds. There is no way ever you will get the same top speed on SO then on Aston. Acceleration is a partial factor, but again, Isnt easy. If they go down a hill, They will change again. It will be hard to make a solid app that doesnt get the wrong people and yet, still catchs cheaters.
True, never thought of that. I suck at maths 
(my understanding of physics is similar to Big Rigs lol...)
(my understanding of physics is similar to Big Rigs lol...)Acceleration is a rate of change, which you can calculate via InSim. The larger the rate of change of velocity (i.e. acceleration) the more likely it is to be caused by a cheat. Now obviously you need to take certain things into account - for instance what might be a good indicator for a cheat, for the XFG, will be totally impractical for the BF1...
I did do some brief testing before I got bored, and it sort of worked for the replay I had of a FBM cheater, but it went a bit weird for other vehicles. The other thing you have to watch out for is reliably detecting crashes and collision bumps, which was fine as long as the person spun. As soon as there was a linear collision the detection got triggered - which means you either need to resort to only logging or kicking, which makes it pointless, or using the position of the user relative to the walls of the track. But this doesn't apply for user placed objects. At that point I got really bored and fed up with pyinsim (purely because it was python).
The only realistic option, in my opinion, would be something that emulates UTpure (going old school there) or punkbuster, and sits at both the client side and at the server side. If there's no constant communication, or detection of a running memory editor, etc. then a kick occurs. But building something like that, which works on top of the LFS stack, is just a lot of effort for what is [currently] only a small issue.
If someone has a better idea, or can spot something that will reasonably work, or some sort of mistake/assumption I've made, then I'm happy to work on the concept with them in conjunction with the barricade (no, not had the time to finish it today so far, yes it's still coming).
I did do some brief testing before I got bored, and it sort of worked for the replay I had of a FBM cheater, but it went a bit weird for other vehicles. The other thing you have to watch out for is reliably detecting crashes and collision bumps, which was fine as long as the person spun. As soon as there was a linear collision the detection got triggered - which means you either need to resort to only logging or kicking, which makes it pointless, or using the position of the user relative to the walls of the track. But this doesn't apply for user placed objects. At that point I got really bored and fed up with pyinsim (purely because it was python).
The only realistic option, in my opinion, would be something that emulates UTpure (going old school there) or punkbuster, and sits at both the client side and at the server side. If there's no constant communication, or detection of a running memory editor, etc. then a kick occurs. But building something like that, which works on top of the LFS stack, is just a lot of effort for what is [currently] only a small issue.
If someone has a better idea, or can spot something that will reasonably work, or some sort of mistake/assumption I've made, then I'm happy to work on the concept with them in conjunction with the barricade (no, not had the time to finish it today so far, yes it's still coming).
I don't remember was it mcgas001 or dougie-lampkin who found out about these "hacks" (IIRC it was discussed on bugs). Did he found out how excatly these "hacks" are working?
But surely it's not that hard to ascertain the realistically feasable top speed of any car on any particular circuit?? From there you just need to add in a buffer, (say 10mph), over which you at least get flagged as a potential cheat. A bit like speed cameras have to allow for inaccuracies in speedometers etc plus a "good will" factor before you trigger the camera. Police guidlines are +10% + 3mph over the speed limit. Obviously there is no speedo inaccuracy in LFS but surely a figure of +10% over the max known feasable speed on any section of any particular car/circuit combo is fair. Cheaters aren't all that smart from what I've seen and the cheats aren't that sophisticate either. They nearly always exceed that limit.
Or maybe it would be possible for the devs, (or someone else in cooperation with info from the devs), to get a computer to calculate the theoretical perfect lap time based on the physics models of each car and track and just auto ban anyone that exceeds it. Surely that wouldn't be that hard?? Surely the model already knows what the limits of grip are, (max aceleration/deceleration and lateral g forces possible for each car), and a modern computer wouldn't take that long to run through even say the 1000 most likely fastest lines.
The thing is, As the_angry_angel says. Its just too much effort and not worth it.
Wow. TAA even had a go at it. I must had said something right.
Ok ..... You look at the acceleration yes, not the top speed of course. Now, if they are accelerating at a rate much faster than is possible for that vehicle then, bingo he is cheating.
Now the problems are:
Each car has a diffferent accelartion rate
>>Detect what car they are in and use a list of max accelartion rates. Possibly use Bob's VPHA for rates.
It's possilbe you might get rammed or hit a moon barrier and of course, break the acceleration rate.
>> BUT, this acceleration rate is only broken for a short period of time when crashing your car, you wont be going faster and faster and faster for a LONG period of time from hitting a barrier. You should only detect a NOS cheater if their acceration rate is sustained for for example more than 2 or 3 seconds, or if frequently used.
It's possible to be pushed along a straight by a BF1 causing your acceleration to be large and sustained for a long period.
>>So do some testing, what is the fastest you can bump draft the cars up to speed and there is your threashold value. This might restrict the detection system. But I'm thinking that the cheat goes much faster than a BF1 pushing a UF1?? no ?
It's possible someone has rediculous lag
>>So what happens here is that the packets arriving are far apart and so the inSim detects for example: 10kph 50kph 120kph 190kph over a period of time. You notice the change in velocity is quite huge. Which would appear as NOS'ing. So you must account for this as well, perhaps look how long it's been since the last speed value was known.
It's possible that someone might get spotted as cheating when in fact it was some unforseem event that gave a false positive. This means that using auto detection and auto perm banning would be possibly unfair to a couple of individuals. At this point you need to decide if collateral damage and potential /unbanning is worth the effort on glitches or dont perm ban and instead ban them until the patch is released.
Of course using this system would mean that the NOS'ers could just use spurts to avoid the detection, so you could keep a count of it while they are on the server and use a counter of times used or something. But they also wont know which servers might run it.
Well I hope this helps in anyway and might inspire someone.
Ok ..... You look at the acceleration yes, not the top speed of course. Now, if they are accelerating at a rate much faster than is possible for that vehicle then, bingo he is cheating.
Now the problems are:
Each car has a diffferent accelartion rate
>>Detect what car they are in and use a list of max accelartion rates. Possibly use Bob's VPHA for rates.
It's possilbe you might get rammed or hit a moon barrier and of course, break the acceleration rate.
>> BUT, this acceleration rate is only broken for a short period of time when crashing your car, you wont be going faster and faster and faster for a LONG period of time from hitting a barrier. You should only detect a NOS cheater if their acceration rate is sustained for for example more than 2 or 3 seconds, or if frequently used.
It's possible to be pushed along a straight by a BF1 causing your acceleration to be large and sustained for a long period.
>>So do some testing, what is the fastest you can bump draft the cars up to speed and there is your threashold value. This might restrict the detection system. But I'm thinking that the cheat goes much faster than a BF1 pushing a UF1?? no ?
It's possible someone has rediculous lag
>>So what happens here is that the packets arriving are far apart and so the inSim detects for example: 10kph 50kph 120kph 190kph over a period of time. You notice the change in velocity is quite huge. Which would appear as NOS'ing. So you must account for this as well, perhaps look how long it's been since the last speed value was known.
It's possible that someone might get spotted as cheating when in fact it was some unforseem event that gave a false positive. This means that using auto detection and auto perm banning would be possibly unfair to a couple of individuals. At this point you need to decide if collateral damage and potential /unbanning is worth the effort on glitches or dont perm ban and instead ban them until the patch is released.
Of course using this system would mean that the NOS'ers could just use spurts to avoid the detection, so you could keep a count of it while they are on the server and use a counter of times used or something. But they also wont know which servers might run it.
Well I hope this helps in anyway and might inspire someone.
Sorry to piss on your bonfire but this was a little while ago when a thread first appeared on the subject in the hosts subforum
The cheaters then learn to use the exploit in bursts. I'd say it would take about 4 kicks for someone to clever to figure that out. Or what about someone only using the exploit on short straights between corners? The the tool becomes immediately ineffective.
Border case. Sorry, but if you're not racing then I don't understand the need for anti-cheat tools on a server. For that reason I'd ignore this situation for a tool like this. I'd argue that a BF1 pushing a UF1 would be very similar to the cheat being used on a UF1, in comparison with a regular UF1 anyway - which would be the baseline in a sensible test.
Catching people blatantly cheating isn't too hard in my opinion (it would be a hell of a lot easier if things like throttle and braking were communicated via InSim, but there's no sense in wishing), if you ignore some of the edge cases and include a best efforts collision system (you'd get false positives, but that's something you'd have to live with), it's those who are being a little sneaky about it which are the real issue (and the dangerous ones imho).
Edit: Sorry if I'm come across as a bit of an arsehole or a know-it-all. I'm pretty knackered and having trouble thinking properly atm
at least im not the lowest
You are, on average, smart and stupid. Read a few more books and decrease your score!
Your Score Summary
Overall, you scored as follows:
37% scored higher (more stupid),
4% scored the same, and
59% scored lower (less stupid).
You are 59% stupid. This means...
Overall, you scored as follows:
37% scored higher (more stupid),
4% scored the same, and
59% scored lower (less stupid).
You are 59% stupid. This means...
You are, on average, smart and stupid. Read a few more books and decrease your score!
Well since I fail now:bananadea, can someone link me to this ?stupid test summary thingy? It's all over this thread and pretty much the topic now.
Oh wait I'll post my contibution submitted to me by a fan of our server.
I posted this before but included the name, so it was removed. Sorry, so here is the no_name version
(patch Y)
Oh wait I'll post my contibution submitted to me by a fan of our server.
I posted this before but included the name, so it was removed. Sorry, so here is the no_name version
(patch Y)
Don't be lazy.. search the thread.. It's not that many pages..
Sorry, I looked a few days ago, couldn't find it. I even used search and all, then gave up. >> http://www.stupidtester.com/
Great now the internet tells me I'm lazy, stupid and blind. The frogs confirmed that as well, I can't blooming see 'em. All I see is damn lines. You put frogs down and get less stupid.
Anyway sorry for being stupid. Least I'm not stupid enough to cheat. Anyway im 43% dumb and 100% honest. 
Great now the internet tells me I'm lazy, stupid and blind. The frogs confirmed that as well, I can't blooming see 'em. All I see is damn lines. You put frogs down and get less stupid.
Anyway sorry for being stupid. Least I'm not stupid enough to cheat.
Anyway im 43% dumb and 100% honest. I have a question. Seen as name and shame isnt allowed on this forum. If I make a website, with all the replays I find of the suspected hackers and name them on there, but give the replay to download aswell. Then put a link to it in my sig, Is that wrong? 
Sounds almost like wrecker barricade to me.
Sounds good though, like a public version of the barricade. It would also save every admin having to look through each replay manually, if the name(s) of the cheater(s) is posted. I know that's why the forum won't allow name and shame, but it's the best way to do it.
Please, please make that website
Your Score Summary
Overall, you scored as follows:
66% scored higher (more stupid),
4% scored the same, and
30% scored lower (less stupid).
You are 30% stupid. This means...
You are kinda' smart. Many have done better, but you did much better than half the other test-takers! Good for you.
Overall, you scored as follows:
66% scored higher (more stupid),
4% scored the same, and
30% scored lower (less stupid).
You are 30% stupid. This means...
You are kinda' smart. Many have done better, but you did much better than half the other test-takers! Good for you.
/me continues learning ASP.Net.
When people are sure that they have a replay from a cheater, they can post it in our forum as well.
We have a topic for the cheater
we did see online.http://www.r-racing.nl/forum/viewtopic.php?f=10&t=35
This thread is closed
cheat ??
(345 posts, closed, started )
FGED GREDG RDFGDR GSFDG