Member Discussions

My coder and I are dabbling in a bot to handle helps and other NPC actions in our MUD, but the concern has arisen that these powerful tools could be used for evil. No, not the dark-side of the force kind of evil, but mass-killing without any sort of effort.

How would we discriminate a perl bot, for example, from your average user? I don't want to have to snoop every suspected player...

And since these bots can use the exact same socket system as players, we can't simply deny everyone but local connections.

Any thoughts?

Ignoring the fact that I'm unsure as to why you seem to actively not want players, a few suggestions. First, most MUD's I've looked at have a simple immortal command along the lines of "connections", displaying a listing to you of all the players in the game with their IP's next to them. Just hit that, find the name of the entity in question and see if it's connecting from localhost or not. If for some reason you don't want to or can't do that, then it's all in the name. Make a script to generate names for the bots with predefined patterns that are easily recognizable, or just write out a list yourself that the bot script can choose from so that you know instantly whether a given entity is a bot from your shell or a player.

I'll try to clarify... We do want players, but we don't want players to abuse bots.

Upon seeing how easily my coder could create a user-side perlscript to connect to the MUD, I grew concerned over how a talented MUDder could whip up a script, and abuse the MUD. Since these bots appear as normal connections (and as long as the cheater isn't stupid enough to multichar), I'm asking if there is a tell-tale sign of a bot.

I don't want to have to be a gestapo simply because a player is levelling hard and not chatting... (also because I don't want to have to snoop everyone)

So, simply, am I just worrying over a rare scenario? Or is there a solution in the code to spot bots and/or abusers?

I gotcha now...I don't know how universal this is, but my experience has been that even "botting" players don't bother going that far. You don't really need knowledge of anything as in-depth as perl to realize your fears, as even the zMUD scripting language is plenty complex to create a full bot if you learn all of it. I've never, however, encountered anyone actually willing to do that...they'll write simple triggers to cast spells repeatedly and train their skill, or to stay in combat indefinitely against one opponent to train their weapons, but never anything complex enough to really worry about. I daresay that if you needed to worry about it, you would already have heard other MUD administrators and coders asking about ways to detect and eliminate bots.

you _HAVEN'T_ heard other admins mentioning this subject? I've heard it on a monthly basis for the entire 8 years i've been mudding.

5 years ago i wrote a set of zMud (and rhapscallion) triggers that could have an immortal be a full functional autoquestor with the ability to have up to 10 part quests and 5 seperate people on different quests at the same time. It also kept a que of up to 10 players waiting in line to recieve a quest.

It worked along the line of the vassago quest code. You find X item (100 possible items) at X location (lots.. just lots..) and retrieve it in X time dependant on how far away it was. It could assign you some 'outlaw' to hunt down and bring back a specific body part as proof of kill. Delivery services.. and a couple other special quests that were more game related than RP related.

The thing is it only took me under a week of casual tweaking to get it all working on rhapscallion (zMud was harder) and since i never told anyone what i was doing they assumed that i had coded an autoquestor in the code itself.

Scripts are really diverse and can be set up to do ANYTHING under ANY circumstances, within the limitations of the player. You can have preset RANDOM responses to people talking to you and even have maps scripted into your mud client so that you can go from a specified point to somewhere else and be active. Watch your damage.. heal self and rest when nessisary. Program in some randomized pauses and you CAN NOT tell a person is botting at all. At worst you would think they werebeing anti social or using basic path scripts instead of total botting.

Moral of the story is its NEVER worth it to try to stop botting. Only worth it so that if someone does bot its not an advantage. If a player is botting their skills that means you are either putting too much emphasis on them or you are making USING the skills too boring. From the perspective of a coder the only thing i think when a player bots is that I've personally let them down in some way as to make botting more interesting than playing.

Seronis
ICQ 43442042

We get bot players every now and then. Never have I seen that someone has constructed with perl. Ussaly just kids using something like zmud.... Ussually what happens is they hero or get high in level. And then rather then questing themselfs, they will write enough triggers to bot the auto quest mob.

Also seen gold run bots too, ussaly what we do is put agressive mobs in the way. Make maze's to things, make keys to treasures so you can only use them once. Then you have to go kill the big ubber mob or mobs again to get the key. Ussaly stops the bots cause it's hard to design one that can handle combat safly. And as most or boting players know. If your boting your character is noholds bar. You get pked and your corpse disolves and some dork walks by and takes everything, your SOL. Tends to be a good deterant.

My usual response to "problems" with botting is thus: If botting is a serious problem, your game is too boring. Or if nothing else, the aspect of the game that gets botted/scripted/triggered/whatever a lot is too boring. Or an action doesn't have a delay where it might or should, and players try to use a trigger and a high-speed connection to get an advantage.

Take, for instance, a mud where every time you cast a spell, under any circumstances, you get better at it. It's not going to take long before some player realizes, "Hey, if I cast this spell over and over, I'll get really good at it. Hey, I can make a trigger that'll cast it over and over!" There are better alternatives.

Another example would be PK triggers. "Bob DISARMS you!" and their trigger does "get sword;wield sword", with no delay, and they blithely continue fighting, thus making the disarm skill utterly useless. Of course, a better approach would be a spatially oriented room in which they would have to dodge past their opponant and go over and grab their weapon again, but...

The point being, if your game is reasonably complex and interesting enough that a bot would have to be equally complex to be particularly useful, you'll have a good deal fewer problems with them than a simplistic, stockish mud.

People will bot whether or not you make your perl script.

I had one of the most complex bots you'd ever see on a mud that was boring to level on which I liked to play. It was undetectable as well, and, at best, they would assume I was being antisocial, as was said, but not that I had been using any form of fastwalk. This bot cast spells on itself, walked around, killed stuff, blah blah blah. It raked in a lot of experience, and I didn't have any frame of reference to create it. And it was on zmud.

Point is, if someone is going to make an uberbot that is most definitely ABUSIVE, it most likely won't be made using your perl script.

from my experience in both administrating and writing bots is this: if you suspect someone of botting, a nonecho tran to a room will immediately identify a bot. echo'ing certain phrases to the room the player is in will work too.