In my opinion i think that there is a easy way to spot a cheater: track the players’ results. Is a statistic method but will work. The solution I am talking about is Ace-Guard. This software has been specifically designed for this task, for chess games but with few changes we can use it. Its operation is as follows: relying on behavioral analysis of the game processes and the gamer activities, the program analyzes the style of play and create a player profile as a reference and is able to analyze the matches in progress comparing to the “real player’s strength potential”.
Don’t forget that Tremulous is significantly more complex than chess.
Do you really think that? Have you ever tried to play?
“The game-tree complexity of chess was first calculated by Claude Shannon as 10^120” cit. wikipedia
Do you still think that?
I can’t even begin to estimate the game-tree complexity of Tremulous. How many legal positions are there on an average Tremulous map? How many legal moves each player can do each turn (e.g. in how many different directions can you shoot?)? The only thing I can reasonably estimate is the number of turns in an average Tremulous game: 10 players * 30 minutes * 20 turns (each player) / second = 720,000 (compared to ~80 turns in an average chess game). The complexity is incomprehensibly large.
btw this is OT, we are talking about fight the cheating. We can use this model for track players results, like kills por minute, deaths and the number of targets hit and how many missed. These are really FEW things unlike what happens in chess. I think you’re continuing to exhibit problems only to hide the fact that with such a software half of your friends should be banned. (And maybe you too!)
LOL, that escalated quickly.
I’m skeptical, because this is much, much harder to do in practice than you think. Look up VAC to see how a practical and efficient anti-cheat system is done.
nwahahahaha !
WRONG. it’s possible to use hacks to gain great advantages, while being humanly indistinguishable from good players. it’s kind of dubious — and generally loserishly negligent — to say that they’re few.
WRONG. depending on the IP ranges u choose to „subscribe to”, concealing identity may currently be as trivial as changing ur IPA, name, qkey and chat style (including chat binds).
how ? i bet that u’re clueless about the technical details of the workings and applicability of the said „solution”.
reference ?
simply those counts can only give u unusably crappy statistics.
note: i did not rule out the possibility of useful analysis of player play styles — however, it definitely won’t be trivial.
reference ?
how ? i bet that u’re clueless about the technical details of the workings and applicability of the said „solution”.
repubblica.it wrote it, is the italian version of BBC. Also fide.com use it! FIDE is the world chess federation.
simply those counts can only give u unusably crappy statistics.
Not true. Because, we can start banning all the users with 100% of targets hit. This is really simple. I can code it. We don’t need that this code work in tremulous we can run it out form the game with a tracker (we can grab this in the old tremulous community site), it’s the easiest way.
We can work with the speed of reaction parameters. = no more punchbots coz they have 0 reaction time and tiggerbots also. You can make huge progress with very little effort if we start to track players.
Online there are lots of tremulous trackers. Just copy the code and paste it. Then we collect data and the work is all done. Nothing impossible.
You know, hackers with infinitely short reaction time and perfect accuracy are pretty obvious and don’t need a complicated system to spot.
we can start banning all the users with 100% of targets hit
If I were a cheater and I knew this system is in place, I would randomly shoot sometimes to make my accuracy not 100%
You know, hackers with infinitely short reaction time and perfect accuracy are pretty obvious and don’t need a complicated system to spot.
I agree.
You know, hackers with infinitely short reaction time and perfect accuracy are pretty obvious and don’t need a complicated system to spot.
You know, you never played against eVo.
You know, hackers with infinitely short reaction time and perfect accuracy are pretty obvious and don’t need a complicated system to spot.
So do it! JUST DO IT!!
You talk and waste time but tell me how many cheaters have you caught? How many in banlist? I bet they’re 5 at most.
If I were a cheater and I knew this system is in place, I would randomly shoot sometimes to make my accuracy not 100%
It still be an improvement. But with other parameters this problem is solved.
You know, you never played against eVo.
Are you saying eVo players have infinitely short reaction times and perfect accuracy and somehow manage to hide it?
You talk and waste time but tell me how many cheaters have you caught?
I don’t catch cheaters very often because they’re either very obvious and get banned by admins immediately, or they hide their hacking well (so your statistical method won’t notice them either).
So do it! JUST DO IT!!
Interesting, I was just about to say the same thing. You said you’ve already implemented a similar system, so prove all skeptics wrong and demonstrate a Tremulous proof-of-concept.
Are you saying eVo players have infinitely short reaction times and perfect accuracy
At their best time, they were better than bots. So, you can’t understand by “watching someone” if he cheats or not (talking about players that are already known on Trem scene. How you can say that everyone that has short reaction cheat? Many good players had incredibles reaction, but this doesn’t mean that they were cheating.) @javateciava’s idea is not so bad (or impossible).
At their best time, they were better than bots.
What does this even mean? They hit more bullets that they shot?
you can’t understand by “watching someone” if he cheats or not
You just summed up my arguments against @javateciava’s idea.
You just summed up my arguments against @javateciava’s idea.
Statistical metod isn’t just “watching someone”. @javateciava’s idea is more is different from what you think.
What does this even mean?
You know already what I mean.
If you know a player well and played long enough with them, eventually it doesn’t take to long for you to identify them even when they try to hide. Sure things like how they chat can help, but also in how they play can give a player away.
It is true that if a bot plays “too perfectly” they can be more easily spotted, so that various attempts (automatic and/or manual) are made to introduce “imperfection” to hide that it is a bot, to the point where very skilled players might even do better than those bots. However, consider that if a particular bot exists, chances are that more than one instance of that bot exists. Also consider that the behavior of these bots can be even more “predictable” than the behavior of irl human players.
So it is plausible that the detection of particular instances of bots could be done by looking for their “signature behaviors”. Of course this approach would only work if the particular bot’s behavior is already known, and would not work on newly introduced bots with different “signature behaviour” not yet known.
We should just review all downloads by all Tremulous players. There’s not that many players so it shouldn’t take more than a couple hours a day. Then we can just ban anyone who downloaded a bot. Problem solved
We should just review all downloads by all Tremulous players. There’s not that many players so it shouldn’t take more than a couple hours a day. Then we can just ban anyone who downloaded a bot. Problem solved
I would prefer solutions that don’tt involve spyware/malware nor NSA surveillance nor invasion of privacy in general.