Welcome!

By registering with us, you'll be able to discuss, share and private message with other members of our community.

Sign up now!

Questions - From One Programmer to Another

Joined
Mar 6, 2019
Messages
12
Note prior to reading: this is not trying to rag on or belittle any current scriptt writers / bot producers, more so questioning the lack of variability I've noticed using these bots and where nearly every bot could be improved from a structural sense which would improve all sequences within the bot and reliability for anti-ban.
--------------------------------------------------------------------------------------------------

Although I'm not savvy in Java or anything useful for writing scriptts and producing bots on any RS platform, I'm curious as to why I see a lot of bots fail to have basic variance in a lot of the commands performed throughout the sequence.

A few things that come to mind is click area variance, timing variance, intentional misclicks, checking exp rates, going to the skills level page, speed of the bot, etc, just to name a few.

Click area variance - Some bots that display the cursor are clear signs that it clicks in the same exact spot, or within a few pixels nearly 100% of the time. Isn't it possible to write a scriptt that when called throughout the sequence it randomly generates an offset for a set area? So if a clickable area is a 100x100 pixel box (say a position in your backpack or bank), with the center being 50 up and 50 in, you generate a random number between +/- 1 to 45 that guarantees your next click will not be identical.


Timing variance - same concept, except set a 0-XXX ms limit on a randomly generated number which in turns always leaves you clicking the next spot at a different time than the last. This is particularly useful when dropping an inventory of logs, fish, ore, etc where you run up and down your rows in a identical stroke from position to position. There is a mining bot I'm not using anymore due to this issue, definitely asking to get banned considering every click is nearly identical.


Intentional Misclicks - what more human can you get than mis-clicking off your next item once or twice an item every inventory or two. Running through dropping an entire inventory but missing one or two due to the variance in click that might have put one of them slightly outside the "click zone", then after finishing the inventory at the bottom it loops back through the scriptt, checks the inventory and goes back to drop position "4" and "11" since an item still remains present from the misclick.


Checking skill table - I forget all the time what level I just reached when AFK'ing or considering my next skill to train when sitting there idle while an inventory is finishing up creating leather bodies or fighting a NPC. Once or twice an hour, to switch over to it and hover the skill you're training or something that is record-able on Jag's end that shows a little more human like action. Maybe this one isn't that beneficial, I'm not sure what is recorded on Jag's end.


Speed of the bot - the majority of bots on this site are way too fast and too perfect to ever justify it as "human-like" whipping through a bank, or through menus. Allowing users to change the speed on the fly in some of the key places with an adjustable variable would be really neat.


On a total side note; does java not give a result when the gameplay screen is set to adjustable and the viewing field is bigger than on the standard small screen? Obviously the safe route is assume everyone will stick to the default loaded size but if I'm using a larger adjustable area and a bot is still using the mini map to click when a clickable area is within reach I'm curious as to how or why this has never been addressed.


Curious what a Java savvy person's take on this is. Have considered learning java but it's just not in the time I have currently to learn and perfect something that is relatively cumbersome.

As another side note - if bots were more advanced user friendly to allow adjustability, I would ALWAYS pay hourly for the most reliable bot that never imitated another user of the same bot identically.


Thanks,
Steve

P.S: I don't like the auto correcting of my proper use of "scrip-t"....
 
Last edited:
Community Manager
Joined
Apr 7, 2015
Messages
1,394
Most of the things you bring up here will either just create more patterns or are just useless in terms of what Jagex detects.

For the “never imitating another user” part, read up on PlayerSense.

geng
 
Java Warlord
Joined
Nov 17, 2014
Messages
4,906
Yh I mean the botting scene is pretty damn old, none of what you listed has not been tried or experiemented with before.
Most of this is what people usually address as plain, old, ineffective, voodoo "antiban". It was very popular a decade ago when people started to add things like these wherever it was possible, and because bot detection back then was so bad, people assumed that it was actually very effective.
 
Joined
Mar 6, 2019
Messages
12
Most of the things you bring up here will either just create more patterns or are just useless in terms of what Jagex detects.

For the “never imitating another user” part, read up on PlayerSense.

geng

I understand some of it may not be as useful or help in even a small way but I still don't understand why there isn't more variance in every script.

Have been reading on PlayerSense; am not experienced enough with Java and it's structure to understand how the swapping of keys in the "Sense" element changes anything. I can run two bots on two different accounts on two different computers side by side and they appear to click the same spots, all the time. Unless it's back end stuff, being green to java puts me as a disadvantage to get in depth.


Yh I mean the botting scene is pretty damn old, none of what you listed has not been tried or experiemented with before.
Most of this is what people usually address as plain, old, ineffective, voodoo "antiban". It was very popular a decade ago when people started to add things like these wherever it was possible, and because bot detection back then was so bad, people assumed that it was actually very effective.

I've been around RS since 02. I've also been programming in other languages for 6-7 years, not with Java, but have the general knowledge a programmer skilled in other languages would know to maybe try to get to the bottom of my questions.

Bot detection was extremely poor back then, let me set the premise that I'm not some nooby with programming and have zero understanding of how any of this works. Clicking on a field is clicking on a field. Clearly bots are extremely detectable when the bot clicks on the same damn area every single time.
 
Java Warlord
Joined
Nov 17, 2014
Messages
4,906
Clearly bots are extremely detectable when the bot clicks on the same damn area every single time.
no. i dont know why you're advocating random variance with everything.. humans are not random, bots shouldnt be either.
Human click positions within a certain clickbox are predictable.
 
Joined
Mar 6, 2019
Messages
12
no. i dont know why you're advocating random variance with everything.. humans are not random, bots shouldnt be either.
Human click positions within a certain clickbox are predictable.

Humans are random, unintentionally from non-precise muscle movements.

When I hold shift and run down to drop an inventory of stuff, I can record my path 1,000,000 different times. It will NEVER be the same, I guarantee not one path from A to B or Y to Z will ever be the same pixel to pixel in regards to mouse movement or click speed/duration.

Random variance within a designated click space and/or mis-clicking is more human like than the same exact pixel getting clicked.
Change my mind.
 
Joined
Mar 6, 2019
Messages
12
and thats exactly whats being done, just not an even distribution

Unless the 25 bots I've used aren't using PlayerSense or whatever is generating different mouse and click speeds, I can't see a difference in the least if they are.

Are the generated variation of numbers only with a +/- 5 difference in ms or pixel? Just with my eyes and not actually inspecting true delay in the code or sequences via other code is enough to see there is very little variability.

Also, what do you mean by "just not an even distribution"? Just not even distribution in what context, delays and speeds?
 
Java Warlord
Joined
Nov 17, 2014
Messages
4,906
Unless the 25 bots I've used aren't using PlayerSense or whatever is generating different mouse and click speeds, I can't see a difference in the least if they are.

Are the generated variation of numbers only with a +/- 5 difference in ms or pixel? Just with my eyes and not actually inspecting true delay in the code or sequences via other code is enough to see there is very little variability.

Also, what do you mean by not an even distribution?
i cant tell you the exact deviation, anyway the clicking is implemented at api level so all bots should behave the same behavior given the same playersense context.

also we've been discussing this for close to an hour without even thinking about whether or not this whole click position thing is actually substantial to bot detection. you can see how thats a problem right?

authors who believe in antiban spend way too much time on coming up with weird mechanics to make their bot seem more human like without the slightest clue whether it has any effect.

All these fundamental aspects like click position, speed, and variance is handled at api level based on actual research and collected human data. And runemate is doing quite well with bans, there are private bots that have been used extensively for years with relatively low ban rate.
If any of this was important it would be noticeable.
 
Joined
Mar 6, 2019
Messages
12
i cant tell you the exact deviation, anyway the clicking is implemented at api level so all bots should behave the same behavior given the same playersense context.

also we've been discussing this for close to an hour without even thinking about whether or not this whole click position thing is actually substantial to bot detection. you can see how thats a problem right?

authors who believe in antiban spend way too much time on coming up with weird mechanics to make their bot seem more human like without the slightest clue whether it has any effect.

All these fundamental aspects like click position, speed, and variance is handled at api level based on actual research and collected human data. And runemate is doing quite well with bans, there are private bots that have been used extensively for years with relatively low ban rate.
If any of this was important it would be noticeable.


Maybe that could have / should have been the first thing you mentioned then, LOL.

I don't know exactly what Jagex looks for but just in the programming line of work it would appear that blasting through menus too quickly and dropping a full inventory is 1.5 seconds is a little unhumanlike. If you say it isn't substantial or they don't detect for it, then I'll take your word for it.
It does sound odd that they wouldn't care as much about click speed, duration of the click, or time between clicks though compared to other things aside from actually penetrating their server and controlling characters from the actual server side (which is safe to assume is not happening).

What do we know to be substantial at this point? Do we have any concrete evidence behind their bot busting system?


Also what determines the PlayerSense context? Is it simply a variable or randomly generated through each client or instance? In theory you should see slightly different result with every account and every bot I run, no?
 
Top