This is an archive of the mabination.com forums which were active from 2010 to 2018. You can not register, post or otherwise interact with the site other than browsing the content for historical purposes. The content is provided as-is, from the moment of the last backup taken of the database in 2019. Image and video embeds are disabled on purpose and represented textually since most of those links are dead.
To view other archive projects go to
https://archives.mabination.com
-
Kingofrunes wrote on 2012-11-29 12:44
Normal people don't lace their messages with lots of things like:
[, \\, |, |\/|,mil,gold,=,-,$ and so on and so forth
If gaming companies actually looked at all the bot strings I'm sure they could find similarities in them.
They need to look at the whole sentence and have a sentence rating system. The more of the characters above it sees, the more likely it's a bot message and should not be permitted. This should all be done server side of course so that the bot clients can't bypass it.
This could be placed on top of the usual filtering.
If the function is done correctly, it shouldn't have that much of an effect on players and stop 90% of bot spammers.
If bots still exist said implementation, collect the strings again and see what they are doing.
I'm only considering this because MMO's have been around for over 10 years now and not one company has found a foul proof way to stop bot spammers and here I am thinking of the most simplistic and easy solution that I could whip up in a day!
Computers/programming languages these days are good enough these days to easily quickly parse and loop through a string quickly and count up matching offending characters and if there's a certain amount of "offending" characters in the string, BAM! Problem Solved >.>
Should be easy to test as well. Look through the chat logs, and reports. Gather a list of sentences and run them each through the created function to see which one passes and which ones fails and adjust from there.
Maybe when I'm bored I'll data mine a bunch of chat from players, mail, party chats, and bots on mabi, guild wars 2, and runescape. Then create this function and test it. Put my theory into practice.
Thoughts?
-
Chiyuri wrote on 2012-11-29 13:42
Quote from Kingofrunes;988871:
Normal people don't lace their messages with lots of things like:
[, \\, |, |\/|,mil,gold,=,-,$ and so on and so forth
If gaming companies actually looked at all the bot strings I'm sure they could find similarities in them.
They need to look at the whole sentence and have a sentence rating system. The more of the characters above it sees, the more likely it's a bot message and should not be permitted. This should all be done server side of course so that the bot clients can't bypass it.
This could be placed on top of the usual filtering.
If the function is done correctly, it shouldn't have that much of an effect on players and stop 90% of bot spammers.
If bots still exist said implementation, collect the strings again and see what they are doing.
I'm only considering this because MMO's have been around for over 10 years now and not one company has found a foul proof way to stop bot spammers and here I am thinking of the most simplistic and easy solution that I could whip up in a day!
Computers/programming languages these days are good enough these days to easily quickly parse and loop through a string quickly and count up matching offending characters and if there's a certain amount of "offending" characters in the string, BAM! Problem Solved >.>
Should be easy to test as well. Look through the chat logs, and reports. Gather a list of sentences and run them each through the created function to see which one passes and which ones fails and adjust from there.
Maybe when I'm bored I'll data mine a bunch of chat from players, mail, party chats, and bots on mabi, guild wars 2, and runescape. Then create this function and test it. Put my theory into practice.
Thoughts?
good idea.. I would probably place that security inside the database. If it got too many signs of been a bot, it fails to add the new post and send an error message back to the client.
-
Lyrveil wrote on 2012-11-30 03:54
This is a pretty good idea in theory. It'll hinder bots until this become widespread enough. Then bots will adapt and we'll enter an endless process. Well it is better than silly censor and limiting message per minute and stuff. Also there's the situation where bot start using normal sentence (unlikely I know but it could happen)
-
Kingofrunes wrote on 2012-11-30 11:42
Fighting hackers and virus' is an endless process anyways. So it's not surprising it's the same for bots. That's what Internet Security is all about and what people get paid big money to do.
As for using normal sentences, they'd have to resort to telling them to google a certain phrase or something I'd think and then it's a matter of waiting for players to report it before adding that phrase into the censor list.
-
Lyrveil wrote on 2012-12-01 02:14
Quote from Kingofrunes;989425:
Fighting hackers and virus' is an endless process anyways. So it's not surprising it's the same for bots. That's what Internet Security is all about and what people get paid big money to do.
Fair point. Though I doubt mmo companies, especially free to play ones, are willing to hire security expert for that. I mean the good company would, but then again even good company who put a lot of resource into banning bots and stuff have a hard time with them.
I mean don't get me wrong, I like your suggestion and I think it'd work, but there has to be a reason why no one else tried it before. I'm sure at least someone in the world in the mmo industry thought about that solution. There might be reason why it was never implemented in a game before. Who knows.
-
Ninjam wrote on 2012-12-01 03:43
Foolproof method to defeat bots.
No talking until the player has finished the beginner tutorials and played actively for at least 10 hours, except on a specific beginner chat channel only seen by gm's and total level 500+ players who enable it in options(it disables itself after beginner tutorials are done and tells you).
After this, the game should also auto mute any player who is blacklisted or muted?(does mabi have mute or only blacklist? iono) over a total of 10 times by 10 different people within a specific amount of time. This number is dynamic and requires more people to blacklist to take effect the longer the target has been playing. After 10 people blacklist, the players name is sent to a *possible bots* list for gm's to see. To prevent abuse a good appeals system needs to be put into place, as well as an account wide limit of only counting 1 blacklist per account per bot/person, and also having people on your friends list or guild counting for less the more they blacklist a person too.
After a dynamic amount of blacklists by people in the same player group, the person is auto un muted and the guilds and friends that participated in the blacklisting will have their names sent to a gm list for investigation instead, to prevent any serious abuse.
This method would hopefully contain any spamming onto a beginner chat channel, which could have much stricter filters in place without disturbing the majority of the playerbase. It would allow for muting of bots that many people blacklist and hopefully will have enough penalties to prevent abuse of it.
the only major drawback is that this requires at least a few GM's to log on and help newbies on the begginer chat(if no higher level players have it enabled), as well as check the logs and ban any obvious bots.
-
RealityBreak wrote on 2012-12-01 19:24
Quote from Ninjam;989838:
Foolproof method to defeat bots.
No talking until the player has finished the beginner tutorials and played actively for at least 10 hours, except on a specific beginner chat channel only seen by gm's and total level 500+ players who enable it in options(it disables itself after beginner tutorials are done and tells you).
After this, the game should also auto mute any player who is blacklisted or muted?(does mabi have mute or only blacklist? iono) over a total of 10 times by 10 different people within a specific amount of time. This number is dynamic and requires more people to blacklist to take effect the longer the target has been playing. After 10 people blacklist, the players name is sent to a *possible bots* list for gm's to see. To prevent abuse a good appeals system needs to be put into place, as well as an account wide limit of only counting 1 blacklist per account per bot/person, and also having people on your friends list or guild counting for less the more they blacklist a person too.
After a dynamic amount of blacklists by people in the same player group, the person is auto un muted and the guilds and friends that participated in the blacklisting will have their names sent to a gm list for investigation instead, to prevent any serious abuse.
This method would hopefully contain any spamming onto a beginner chat channel, which could have much stricter filters in place without disturbing the majority of the playerbase. It would allow for muting of bots that many people blacklist and hopefully will have enough penalties to prevent abuse of it.
the only major drawback is that this requires at least a few GM's to log on and help newbies on the begginer chat(if no higher level players have it enabled), as well as check the logs and ban any obvious bots.
Or we could just mute all noobs permanently. No new characters can talk, ftw