PDA

View Full Version : SRL Botwatch Competition - A new type of competition!



Hobbit
03-09-2014, 12:59 AM
SRL Botwatch Competition
March 8, 2014 - June 7, 2014

Overview
As a community that develops and uses ‘bot’ technology, it is easy to get caught up in speculation and theories regarding the feasible capabilities of various bot detection methods. Not only that, but misconceptions and misinformation spread and influence the development of various scripts and of SRL itself.

To combat this, the SRL staff would like to propose a challenge to the community, in order to encourage real research and a deeper understanding of the capabilities and limitations of modern technologies.

Your task is to design and justify a system to detect bots in the game Runescape. You’re free to use any and all resources at your disposal. The final product will consist of a paper that fully describes your approach, as well as a slideshow presentation to a panel of judges that will consist of SRL Staff members as well as real industry professionals. Not only will this provide you with a good knowledge of modern data technologies, but can also be added to your resume and used as a conversation topic in future interviews.

Schedule*

Competition Posted - March 8
Proposal (2 Weeks) - March 23 - Submission Form (https://docs.google.com/forms/d/1-Hg35PVRV0RttvbnLgZFC4oGE7FhRbW-q24PoRcplgM/viewform) (Plain text only, no links)
Eligibility Responses (5 Days) - March 23 - 29
Paper Draft (5 Weeks) - May 4 (Extended by 1 week from original April 27 deadline)
Paper Final (2 Weeks) - May 18
Presentations (2 Weeks) - May 19 - June 1 (Flexible with finals/schedules)
Results Announced (1 Week) - June 7


Note: Staff reserve the right to change or modify deadlines at any time for any reason.
*All dates imply 23:59 UTC unless otherwise noted

Eligibility

Submit short (half page) proposal
Ranks Jr. Member+ (Jr. must submit full page proposal)
Sporadic Targeted Interviews
Small groups allowed if all members qualify otherwise**
SSRL / Developer / Super Mod ranks automatically qualify


Note: Staff reserve the right to perform supplementary eligibility interviews as well as the right to allow or deny participation of anyone for any reason.
**Final submission should reflect that of a group (proportionally more research and depth)

Prizes

SRL Scholar Rank (new!) - More Details (http://villavu.com/forum/showthread.php?t=108341)
SRL Swag (Shirt/Mug/Etc)
More awesome prizes announced soon!

Requirements

Data

What data needs to be collected and how?
How will this data be stored?
How will this impact the end users?

Analysis

How will the data be analyzed?
What to look for?

System

What are the hardware and software requirements for the system?
What is the labour required to create/maintain the system (Dev, SA, Analyst, etc)?

Cost

What is the cost of the system?
What is the cost on human resources?
What are the costs of expansion?

Future

Growth/Expansion/Adaptability
Industry competition

Misc

Must justify all of the decisions made
Cite and reference all used sources



Note: This is a list of minimums, you are encouraged to go above and beyond what is in this list.

Restrictions

Max cost 1% of Jagex 2012 Earnings (£50 Million earned, GBP£500k cap (USD$837k))
Must use 100% currently existing technology
Must be able to be managed by a small team of employees (~1% of 450 total employees)

Evaluation Guidelines / Process

Paper

Recommended IEEE Publishing Template (http://www.ieee.org/conferences_events/conferences/publishing/templates.html)

Bonus points for LaTeX
MS Word/ODF documents must use the recommended template

Submitted as both PDF/PostScript and source document (.tex,.doc,.bst, etc)
Concrete judging rubric to be determined

Slides

Provide general overview, mostly as a tool for presentation

Presentation

Present slides to judging panel


Resources

http://en.wikipedia.org/wiki/Data_analysis
http://en.wikipedia.org/wiki/Data_mining
http://en.wikipedia.org/wiki/Big_data
http://en.wikipedia.org/wiki/Machine_learning
http://en.wikipedia.org/wiki/Time_series_analysis
http://en.wikipedia.org/wiki/Data_stream_mining
http://en.wikipedia.org/wiki/Sequence_mining

Clarity
03-09-2014, 01:58 AM
Wow, a damn thorough competition, I'm looking forward to reading the winning paper. Industry professional contacts are pretty intriguing.
Good luck to the entrants! This is way out of my league for now :p

Hobbit
03-09-2014, 02:06 AM
Wow, a damn thorough competition, I'm looking forward to reading the winning paper. Industry professional contacts are pretty intriguing.
Good luck to the entrants! This is way out of my league for now :p

Can't hurt to try :) We are by no means expecting a masters thesis to come out of this. We have realistic expectations of what we will receive ;p

And trust me, the additional prizes we are working on right now are very worth it.

rj
03-09-2014, 03:17 AM
To get this straight we are going up with a system and not trying to make one?

Hobbit
03-09-2014, 03:24 AM
Yes, you are just coming up with one, not implementing.

Brandon
03-09-2014, 04:40 AM
Your task is to design and justify a system to detect bots in the game Runescape.
Yes, you are just coming up with one, not implementing.


I suspect a lot of people will have this rank very soon.. Especially `if` teamwork is allowed.

Jake
03-09-2014, 04:49 AM
Seems interesting.
Might give it a shot, best of luck to others who choose the same.

NKN
03-09-2014, 04:50 AM
Who's up for working with me?
<3

If not, I guess I'll solo this black magic.

Jake
03-09-2014, 05:14 AM
pm me

Who's up for working with me?
<3

If not, I guess I'll solo this black magic.

Pakyakkistan
03-09-2014, 05:29 AM
Wow, a damn thorough competition, I'm looking forward to reading the winning paper. Industry professional contacts are pretty intriguing.
Good luck to the entrants! This is way out of my league for now :p
Couldn't have said it better myself. I may get my hands dirty and see what I can do tho! Like Hobbit said, definitely cannot hurt to try!

Hazzah
03-09-2014, 07:02 AM
Interesting stuff, good luck to everyone who enters, I expect the entries will be very few but hopefully those that do enter they are good quality.

slacky
03-09-2014, 08:39 AM
Sounds extremely boring, and very complicated.
I will be very shocked if someone turns in a decent paper.

Clarity
03-09-2014, 11:39 AM
What is the minimum or maximum complexity of the hypothetical bot to be detected? e.g. Static coords vs. Random waits vs. Humanized profiles

J J
03-09-2014, 12:31 PM
So SRL is teaming up with Jagex now? :norris:
Interesting subject for sure! I was thinking of testing something a while ago, but with this competition in mind I'll use it to write a small paper.

bg5
03-09-2014, 04:18 PM
Wow, professionalism required scared me out!

rj
03-09-2014, 04:41 PM
So the presentation is us like vidding ourselves presenting it or what? :s
Hobbit;

Neva Banned
03-09-2014, 07:48 PM
I dont have the experience to give a paper on this but I would recommend anyone trying this competition to get the most up to date rsps server up and create your own bot detection to give you the best real world result.

cause
03-09-2014, 08:51 PM
I'm totally in over my head on this one, but, to clarify: An example of HOW to collect the data (which then needs to be analyzed) would be to run a runescape private server, then have a client use a popular script for a few hours so you have some data...? Am I on the right track? Otherwise I have no idea how you would get data.
EDIT: lol it seems I have parroted nevabanned

Clarity
03-09-2014, 09:26 PM
I could be wrong but my impression was we would only need to describe a theoretical data collection and analysis process, not actually do it, although an actual simulation would provide for a better argument probably.

Neva Banned
03-09-2014, 09:38 PM
TBH why do it half a** if we can do a plausible simulation, why not prepare ourselves for the future.

Neva Banned
03-09-2014, 09:48 PM
I'm totally in over my head on this one, but, to clarify: An example of HOW to collect the data (which then needs to be analyzed) would be to run a runescape private server, then have a client use a popular script for a few hours so you have some data...? Am I on the right track? Otherwise I have no idea how you would get data.
EDIT: lol it seems I have parroted nevabanned

Depending on the rules of this, for the rsps you could use simba to bot on the rsps (quickest way) and make your own client (java based) to bot on it. You would then create a script run it for lets say 2 hours then do exactly what the bot is doing for 2 hours and compare. Now with this collection you have to create a way to identify the bot. Obviously for bot watch its doing it on a completely different scale.

Neva Banned
03-09-2014, 09:49 PM
If any one else needs help or idea's I have written some things up if you wish to take a look add me on skype.

Skype:credulite

Brandon
03-09-2014, 09:58 PM
Where do you submit?

Also, if no one has the rank.. who exactly is judging? Especially if it has difficult concepts..

Hobbit
03-09-2014, 10:09 PM
Link to proposal submission added to first post.

core
03-09-2014, 11:10 PM
I could be wrong but my impression was we would only need to describe a theoretical data collection and analysis process, not actually do it, although an actual simulation would provide for a better argument probably.

This is correct.


I dont have the experience to give a paper on this but I would recommend anyone trying this competition to get the most up to date rsps server up and create your own bot detection to give you the best real world result.

Depending on what your approach is, this may be very difficult. Actually implementing your solution will not give you any bonus or advantage in the judging. (But, obviously, we won't stop you if you're doing it for your own reasons)


Where do you submit?

Also, if no one has the rank.. who exactly is judging? Especially if it has difficult concepts..

"a panel of judges that will consist of SRL Staff members as well as real industry professionals"

weequ
03-09-2014, 11:28 PM
Damn now I miss I had completed the machine learning and data mining courses. I'm looking forward to see what you guys come up with.

Ashaman88
03-10-2014, 12:37 AM
pay wizzup? 800K to plant a function in simba which sends character info for people using simba to jagex. problem solved.


but really sounds like a cool deal - good luck to whoever does this!

Kevin
03-10-2014, 06:40 PM
My main question on the subject, is how do you plan on presentations to the judges working? A skype video call? Or something else? How will this be different/similar for groups as opposed to individuals?

And less important questions, but still curious on: will individuals be graded separately in some manner or exclusively on group merit? Would this allow someone undeserving to be carried to the rank simply because they are friends with more capable members?

Neva Banned
03-10-2014, 07:18 PM
My main question on the subject, is how do you plan on presentations to the judges working? A skype video call? Or something else? How will this be different/similar for groups as opposed to individuals?

And less important questions, but still curious on: will individuals be graded separately in some manner or exclusively on group merit? Would this allow someone undeserving to be carried to the rank simply because they are friends with more capable members?

Unless I am thinking about it wrong, as long as you are able to understand the basics of what a script is doing there in no advantage to someone who has been scripting for years. Someone that is majoring computer science(or in that realm) or is a bot client developer would have a huge advantage, if not then we all have an equal chance.

Brandon
03-10-2014, 07:29 PM
Unless I am thinking about it wrong, as long as you are able to understand the basics of what a script is doing there in no advantage to someone who has been scripting for years. Someone that is majoring computer science(or in that realm) or is a bot client developer would have a huge advantage, if not then we all have an equal chance.


Consider the requirements and analysis phases and the links provided..

"What data needs to be collected?"
"How will data be analysed?"

Then the evaluation process:

"MS Word/ODF documents must use the recommended template"

I'd say it looks pretty professional and rules out just bare scripters.. which is sort of unfair imo. However it does say groups are allowed so I guess it works out.. I'd say you might have a higher chance in a group even if you are pro and decide to solo it.

It is a bot system you're designing and explaining after all.

They are asking for a lot for developing a bot system though. And judging from the comments, I guess coding one doesn't help? Powerbot's requirements aren't even this high. You can develop a bot system without a lot of those requirements (specifically for windows) and it can be extremely effective. For example, being as aggressive as the Warden system or PunkBuster and monitor the user's mouse, memory and PE scanning will not even need sending of data or any flagging or analysis.. It's an instaban effective system. No need for collection of data unless your scans are unsure of what it finds depending on its heuristics.


For me, I simply modified a paper I had written a while back on reverse-engineering and submitted that (extremely dumbed down) with some code examples. I doubt it'd get me anywhere but meh.. worth a shot to see how they judge. I'll release what I submitted after it gets judged..

Again, just join a team.. I can think of a few people that would try this out and be willing to work with others. Maybe NKN being one of them (which he mentioned a couple posts back) or Benland if he's willing..

Sin
03-10-2014, 11:33 PM
297 words holeh.

P1nky
03-11-2014, 12:17 PM
Wish this competition was taken place during Summer :( Too much Uni stuff to worry about atm :(.

Clarity
03-13-2014, 12:15 AM
I find this pretty interesting regardless of competition results so I submitted a proposal :p
Wonder what those extra prizes are?

The Mayor
03-17-2014, 10:14 AM
Seems like an interesting competition, although I think there's only a handful of people on these forums capable of fulfilling the requirements.

shockage
03-19-2014, 12:04 AM
I have a SIMBA script that I was testing, and I was teleported. I received no repercussions, but it made me consider why it happened. I'll attach the script to this post when I get home.

I believe it occurred since I did not use boxes to outline areas for mouse click. After finding each point I wanted to click, and this colour finding never failed me to find a wrong object, but I would simply do point.rand(1). I was using SRL for mouse movements and clicks.

This occurred after running the script overnight with no fails. So the system must use statistics to find number of clicks relative to pixels... as a human wouldn't click in roughly the same area multiple of times. I believe the points are strictly related to the main-screen, and not the game's objects--the reason for this will be evident when I post the script, but it involves the minimap.

catsrback321
03-19-2014, 08:12 PM
This reminds me of when Rs buddy used to be around :p the site closed and the owner ran off to be a jagex employee haha
anyway good luck to all entrants :)
another thing that may be helpful for people to know is what type of bot they are catching? I'm no expert in any form of scripting but
I presume it is easier to catch an injection/ reflection script than it is to catch an ashaman colour script?

Clarity
03-19-2014, 08:26 PM
This reminds me of when Rs buddy used to be around :p the site closed and the owner ran off to be a jagex employee haha
anyway good luck to all entrants :)
another thing that may be helpful for people to know is what type of bot they are catching? I'm no expert in any form of scripting but
I presume it is easier to catch an injection/ reflection script than it is to catch an ashaman colour script?
For my proposal I assumed that the hypothetical system would have to detect every type of bot available on the market at the moment, but that may be too much, idk.

Kevin
03-19-2014, 09:06 PM
For my proposal I assumed that the hypothetical system would have to detect every type of bot available on the market at the moment, but that may be too much, idk.

Well... Can you even name every bot on the market at the moment? I'd be willing to believe there are private bots for the heavy duty gold-farmers (no, I do not mean private scripts).

Clarity
03-19-2014, 09:19 PM
Well... Can you even name every bot on the market at the moment? I'd be willing to believe there are private bots for the heavy duty gold-farmers (no, I do not mean private scripts).
On the market = every bot I can find available for use publicly.
i.e. RiD, SRL, Powerbot, Tribot, etc. but for my concept design it doesn't really matter what bot/type of bot I'm trying to detect, if it's not a legit player using the account, it should be detected immediately.
Whether this will be able to fit the budget constraint is the bigger issue :p but I think it can.

J J
03-22-2014, 12:08 AM
The proposal of Brandon and me has been send in. Looking forward to see how this competition is going to work out.

core
03-22-2014, 12:45 AM
Thanks for all the entries so far!

A reminder to those of you who haven't submitted your proposals yet, they are due in two days!

As clarification: you don't need to be some kind of expert to participate. The entire purpose of this competition is to provide motivation to explore and learn about a topic closely in line with the efforts of this community.

If you have any questions or concerns, or need something clarified, please don't hesitate to ask.

Looking forward to reading all of your proposals. Good luck!

Rich
03-25-2014, 11:54 AM
Would actually really enjoy taking part in this considering I can most definitely thank this forum for the career path I've taken. Suck that I've quite a bit of work to do at the moment then...Good luck to everyone involved though!

J J
03-29-2014, 01:12 PM
Will we get to know who has qualified or will that stay private?

core
03-30-2014, 02:29 AM
Eligibility PM's have now been sent out. Good luck going forward to those who qualified, all of the proposals had great ideas!

Please continue to discuss and ask questions here going forward.



Will we get to know who has qualified or will that stay private?

We have no intent to publish a list of those who have qualified at this time, however those who have or have not qualified are free to share that information if they so choose.

Xenobitz
04-03-2014, 07:56 PM
Eligibility PM's have now been sent out. Good luck going forward to those who qualified, all of the proposals had great ideas!

Please continue to discuss and ask questions here going forward.




We have no intent to publish a list of those who have qualified at this time, however those who have or have not qualified are free to share that information if they so choose.

Is it too late to submit?

rj
04-03-2014, 07:59 PM
Is it too late to submit?

yes

Dgby714
04-20-2014, 02:41 AM
I completely forgot about this >..>

Kevin
04-21-2014, 03:48 PM
I completely forgot about this >..>

I did as well. Oh well, there are still intelligent enough people working on it.

Dgby714
04-21-2014, 05:44 PM
I did as well. Oh well, there are still intelligent enough people working on it.

Hmm, core said I still have time to make a submission seeing SSRL don't have to do the preliminary stuff.

I prob still wont be able to submit anything though, finals are in 2 weeks and I have like 5 essays and 3 test to do...

NKN
04-21-2014, 06:22 PM
Hmm, core said I still have time to make a submission seeing SSRL don't have to do the preliminary stuff.

I prob still wont be able to submit anything though, finals are in 2 weeks and I have like 5 essays and 3 test to do...

Dang. Atleast you got the chance, I completely forgot to get the proposal done in time.

Kevin
04-21-2014, 08:10 PM
Dang. Atleast you got the chance, I completely forgot to get the proposal done in time.

True partners we were.

NKN
04-21-2014, 08:35 PM
True partners we were.
The best were were.

Clarity
04-25-2014, 05:34 PM
Yeah, the deadline for the first draft is unfortunately in the middle of exam season for I'm guessing most people (I have 3 papers due/4 exams this/next week).

If enough people agree, is it possible to push back the first draft deadline or increase its flexibility, but keep all the other deadlines the same? I'm still very interested in this project, but I just can't prioritize it over schoolwork. I don't think I'm alone in this situation.

Kyle Undefined
04-26-2014, 07:55 AM
Man, I wish I would have seen this before. Sounds like fun.

core
04-26-2014, 05:57 PM
Yeah, the deadline for the first draft is unfortunately in the middle of exam season for I'm guessing most people (I have 3 papers due/4 exams this/next week).

If enough people agree, is it possible to push back the first draft deadline or increase its flexibility, but keep all the other deadlines the same? I'm still very interested in this project, but I just can't prioritize it over schoolwork. I don't think I'm alone in this situation.

I agree, and I'm actually in the same situation as well. The goal of the first draft deadline is so that staff can ensure that people are on the right track and to provide some preliminary feedback so that people can revise before the final submission, as well as motivate people to not wait until the last second to start.

I think it would be fair to make the draft submission more flexible, and pushing it back by 1 week seems to be the way to make that happen. I can't push it back more than that, because that would leave only 1 week between the draft and final submissions, and staff will not be able to read the drafts and provide feedback in time for anyone to make any changes, making the draft entirely worthless.

Also remember that this is just a draft. It doesn't need to be perfect. I welcome feedback/comments on these changes as well.


Man, I wish I would have seen this before. Sounds like fun.

I wish you'd've seen this before, too.

Maybe you'll bust out a quick draft in the next week anyway? ;)

Kyle Undefined
04-26-2014, 07:33 PM
I wish you'd've seen this before, too.

Maybe you'll bust out a quick draft in the next week anyway? ;)

Psh, I wish. I'm too OCD to throw something together lol.

Eejit
04-28-2014, 10:40 PM
Although its a bit late, this may be of some use to participants: http://asp.eurasipjournals.com/content/pdf/1687-6180-2009-797159.pdf

The Mayor
04-28-2014, 11:14 PM
Lol there is actually quite a bit of literature on runescape and bots in general, even some peer-reviewed stuff.

http://cs.ucsb.edu/~chris/research/doc/spmagazine09_gamebots.pdf

https://wiki.umn.edu/pub/UmmCSciSeniorSeminar/Fall2013PapersAndTalks/Presentation_Slides_Phou_Lee.pdf

core
04-29-2014, 07:05 PM
Lol there is actually quite a bit of literature on runescape and bots in general, even some peer-reviewed stuff.

Good find! This competition and large prize would not exist if this wasn't a real problem with no published 'modern' solution ;). A good submission here could easily be a foundation for future work, such as an undergraduate research project. The knowledge gained in the process would also be a great topic of discussion in job interviews and applications.

The linked projects are good starting points, but their approaches stop short of ideal. They describe reactive systems that process collected data after the fact. They stop short of building on their techniques to flex the capabilities of various tools that are available today.

An ideal bot detection system would flag the bot as quickly as possible. Then it should cut the time from detection to verification and need minimal human interaction. To do this, the system needs to analyze the data as it comes in in real time, and not introducing much (if any) overhead for the client.

In other industries, these types of systems exist as solutions to many different problems. In many cases, they operate on much more data than could be generated by the whole of Runescape. From this perspective, it is interesting to see these solutions not perforating this field as much as I would expect them to. Right now there is only one major game developer that I know of that is doing any serious work on this idea.

rj
04-29-2014, 09:30 PM
one of the prizes should be a shirt signed by all the cool admins :)

Krazy_Meerkat
05-15-2014, 07:38 AM
one of the prizes should be a shirt signed by all the cool admins :)

Or a job at Jagex :p It's good to see a competition which will get people thinking about how bot-detection works against some of the different types of bots.. I don't see much specification on the type of bot it's being designed for but I assume people are submitting a bot detection paper specifically targeting one kind of bot.

rj
05-18-2014, 02:19 PM
Have presentations started yet?

Cheddy
05-18-2014, 04:27 PM
Easy, prevention before removal; cripple injection/reflection bots by creating a launcher for the game which randomises the gamepack completely upon launch every time. That way the only java bots that would work are ones that include the updater in the bot and the updaters would have to be 100% flawless. Then spam the game with random events that are near to impossible for scrapers to complete. Alternatively, rewrite the whole of RS in C using a DirectX/mantle engine; that way performance would be increased whilst also making it far harder to inject into. Then just pwn the scrapers again with randoms

rj
05-18-2014, 06:43 PM
Easy, prevention before removal; cripple injection/reflection bots by creating a launcher for the game which randomises the gamepack completely upon launch every time. That way the only java bots that would work are ones that include the updater in the bot and the updaters would have to be 100% flawless. Then spam the game with random events that are near to impossible for scrapers to complete. Alternatively, rewrite the whole of RS in C using a DirectX/mantle engine; that way performance would be increased whilst also making it far harder to inject into. Then just pwn the scrapers again with randoms

Rewriting the whole game in C is completely impractical, the whole reason Jagex keeps Java is so it runs in the browser and to keep cross platform support (Easily)(not to mention the amount of time it takes to write bigger programs in C vs Java)

And then players would suffer from annoying random events..

Cheddy
05-18-2014, 06:52 PM
Rewriting the whole game in C is completely impractical, the whole reason Jagex keeps Java is so it runs in the browser and to keep cross platform support (Easily)(not to mention the amount of time it takes to write nigger programs in C vs Java)

And then players would suffer from annoying random events..
In which case, they will never truly get rid of bots; as time goes on bots will become cleverer, maybe even self aware. I know that at the minute I, personally, am training a self learning mouse for my java bot, next stop is self learning auto-responder. Once bots are self aware they will be impossible to differentiate from humans

rj
05-18-2014, 06:56 PM
In which case, they will never truly get rid of bots; as time goes on bots will become cleverer, maybe even self aware. I know that at the minute I, personally, am training a self learning mouse for my java bot, next stop is self learning auto-responder. Once bots are self aware they will be impossible to differentiate from humans

I don't think they will reach that point anytime soon. And it has been stated many times that Jagex will never get rid of bots ever.

NKN
05-18-2014, 07:43 PM
Easy, prevention before removal; cripple injection/reflection bots by creating a launcher for the game which randomises the gamepack completely upon launch every time. That way the only java bots that would work are ones that include the updater in the bot and the updaters would have to be 100% flawless. Then spam the game with random events that are near to impossible for scrapers to complete. Alternatively, rewrite the whole of RS in C using a DirectX/mantle engine; that way performance would be increased whilst also making it far harder to inject into. Then just pwn the scrapers again with randoms

What if I had it detect a changed gamepack, upload it to my server, the update does its magic, and outputs the hook file.

Cheddy
05-18-2014, 08:09 PM
What if I had it detect a changed gamepack, upload it to my server, the update does its magic, and outputs the hook file.
Ahh no. I mean the user only comes into contact with the gamepack the launcher has spat out. Which obfuscates it randomly every time it is launched, hence the updater would have to be flawless. But yes, you could take a gamepack and make your bot run off that instead. But that would be completely illegal as you would be distributing Jagex's game ;)

Krazy_Meerkat
05-25-2014, 01:44 PM
I think they are trying to be realistic with this competition fish, as in, something Jagex would evaluate and approve financially. After all, programming time is money.. I don't think a game re-design is in the cards as a potential botwatch proposal.
Multipliers and other application padding hasn't had much of an effect on the target and has only provoked elitism within the remaining java-bot coders. It also wastes cpu time and bandwidth, which does eventually cancel out any benefits gained by stopping bots (when you start scaling up)..
Also collecting information and building profiles about player behaviour has to be very lean when you think that thousands of accounts will be online at the same time..

Clarity
05-26-2014, 05:09 AM
Or a job at Jagex :p It's good to see a competition which will get people thinking about how bot-detection works against some of the different types of bots.. I don't see much specification on the type of bot it's being designed for but I assume people are submitting a bot detection paper specifically targeting one kind of bot.

My paper will be late since I had exams but my proposed system detects any bot, including complex/human-like ones such as RobotzInDisguise.

rj
06-11-2014, 10:55 PM
Alright 4 days past due date!
Hobbit;

EvilCabbage
06-30-2014, 04:48 PM
Who's up for working with me?
<3

If not, I guess I'll solo this black magic.

Hey NKN! :)

On topic: This is out of my league for sure but I would love to read the outcome!

Kevin
06-30-2014, 04:57 PM
Hey NKN! :)

On topic: This is out of my league for sure but I would love to read the outcome!

NKN and I were gonna... but then we didn't write the paper in time. Guess we preferred actually programming the results over writing a paper on how we'd do that :p

As for the outcome... we're still waiting on the results for almost 3 weeks now...

EvilCabbage
06-30-2014, 05:05 PM
NKN and I were gonna... but then we didn't write the paper in time. Guess we preferred actually programming the results over writing a paper on how we'd do that :p

As for the outcome... we're still waiting on the results for almost 3 weeks now...

Ohh, I'll check out your scripts then instead hehe

Kevin
06-30-2014, 05:15 PM
Ohh, I'll check out your scripts then instead hehe

Oh god no. I haven't written or fixed a script in... I wanna say since last year. There are nice concepts in some of them, but NONE of them will work if you wanna use any of them.

core
06-30-2014, 06:05 PM
The few people that made it to the final stages did not submit the final paper or presentation, so this competition is now closed with no winner.

Sorry to disappoint the many of you who were looking forward to the results of this competition.

We have received a lot of feedback from participants and other members that will help us host better competitions in the future.

EvilCabbage
06-30-2014, 06:18 PM
Oh god no. I haven't written or fixed a script in... I wanna say since last year. There are nice concepts in some of them, but NONE of them will work if you wanna use any of them.

:( <-- The face of dissapointment

cosmasjdz
06-30-2014, 08:21 PM
I am not so experienced in programming. But taking in account jagex monitors mouse movements and color they clicking ,id look for similar patterns maybe in pixel colors or in actions.Porbably the less people use same bot the less likely ban is going to happen. Also during fixing switchtoworld fucntions and when i had forgot failsafes with miner i noticed that random world option has higher change of logging you in in area with people already in. I think jagex use this in hope poeple would spot bot and report it. This suggest their bot detection system isnt so great they are taking about, because otherwise they wouldnt need this. Still only my observation But in the weekends i am gona to test this theory.

masterBB
06-30-2014, 09:40 PM
I am not so experienced in programming. But taking in account jagex monitors mouse movements and color they clicking ,id look for similar patterns maybe in pixel colors or in actions.Porbably the less people use same bot the less likely ban is going to happen. Also during fixing switchtoworld fucntions and when i had forgot failsafes with miner i noticed that random world option has higher change of logging you in in area with people already in. I think jagex use this in hope poeple would spot bot and report it. This suggest their bot detection system isnt so great they are taking about, because otherwise they wouldnt need this. Still only my observation But in the weekends i am gona to test this theory.

Nothing specific though. Using mousepaths? Using colour clicks?

All those things, we already tried to randomize.

rj
07-01-2014, 01:51 AM
The few people that made it to the final stages did not submit the final paper or presentation, so this competition is now closed with no winner.

Sorry to disappoint the many of you who were looking forward to the results of this competition.

We have received a lot of feedback from participants and other members that will help us host better competitions in the future.

is apps for this rank going to open up?

Krazy_Meerkat
07-09-2014, 08:15 AM
is apps for this rank going to open up?

I'd also like to know what's going on with this.. The competition was a fail, but I hope that future competitions are more focused on smart-botting rather than bot-detection and anti-botting..

Brandon
07-09-2014, 01:41 PM
I'd also like to know what's going on with this.. The competition was a fail, but I hope that future competitions are more focused on smart-botting rather than bot-detection and anti-botting..


The competition itself was actually a neat idea but way too many requirements and fiddling around with useless stuff.. They tried too hard to make it too professional.

Even powerbot's scholar system is easier. I'd rather code than write a paper for this. But that's just me.

JJ had the best idea and it was codes but we didn't have time to write a paper. :l
We preferred testing it out rather than theories.


I wonder who made it to the finals and dropped.. My money goes on NKN..

Kevin
07-09-2014, 05:51 PM
I wonder who made it to the finals and dropped.. My money goes on NKN..

Not NKN, we started it together, coded up a few things, tested them out, then never wrote the papers...

Brandon
07-09-2014, 08:22 PM
Not NKN, we started it together, coded up a few things, tested them out, then never wrote the papers...

Are you saying I lost my money? -__- Dammit NKN. Why you no write paper?!

Meh.. Expected though. I didn't really see anyone wanting to write a paper. What'd you guys code? ;)

NKN
07-09-2014, 10:26 PM
Are you saying I lost my money? -__- Dammit NKN. Why you no write paper?!

Meh.. Expected though. I didn't really see anyone wanting to write a paper. What'd you guys code? ;)

The game of laziness is a dangerous one indeed.

Novite666
07-31-2014, 02:03 PM
The few people that made it to the final stages did not submit the final paper or presentation, so this competition is now closed with no winner.

Sorry to disappoint the many of you who were looking forward to the results of this competition.

We have received a lot of feedback from participants and other members that will help us host better competitions in the future.

Heartbroken tbh.