ZDNet article regarding SETI - Now with link to a response from Berkeley

[FoBoT]

*edit* see Poof's link to a response from Berkeley on the Ars forum, about 60 posts down

anybody heard about this?

Cheats wreak havoc on SETI@home: participants

SETI@home administrators are allegedly ignoring claims that the project is being sabotaged by miscreants who are threatening to derail its reputation and that of many valuable Internet-based distributed computing projects.

here is another good quote

A bulletin-board posting allegedly from SETI Netherlands' team manager states cheating throughout the team is rampant and estimates that 41 per cent of the team's work is illegitimate.

:Q

Smokeball has provided a link down the thread abit to an open letter that many SETI teams are signing to send to the people that run SETI

 

[Robor]

This makes me friggin sick. I've spent nearly 3.5 years doing the SETI project and while I believe in the nature of the project I'll admit that the stats are a big part of it. For that matter, the UofB wouldn't even have individual, club, team, etc stats if they didn't think they were an important part. Why then are they ignoring these alleged cheaters? That the S@H1 project is nearing it's end is no excuse. There have been cheating allegations for a while now and apparently nothing has been done about it. More important, if UofB isn't going to stop cheating in SETI@Home 1 what makes us think they will do anything about it in SETI@Home 2? The simple fact is there is no hackproof piece of software. If it can be programmed it can and will be hacked, it's just a matter of when and how. UofB needs to do something or they will not get near the response they got in S@H1.

Maybe it's too late to fix the current SETI client but I hope UofB does something about the invalid results that have been turned in by these cheaters. It would be so nice to see the cheater accounts removed from the individual and team stats. All I know is I was climbing the overall charts doing 125 - 150 WU's per day and reached #152 overall at one point. Now, despite the fact that I'm producing nearly 200 WU's per day on average I've fallen back to the 169th position.

I don't think we have any cheaters on our team but looking at some of the numbers put up by some members on other teams I have to raise my eyebrow. If you're a cheater and you're reading this you are a loser. You're a pathetic loser. :|

EDIT: In all my venting I forgot to thank ya for the info. Thanks FoBoT!

 

[Shuxclams]

Its truly unfourtunete. I have seen those posts in the SETI@HOME forums and it is saddening. It really does detract from the "Fun" aspect when people have so little life that they have to resort to cheating in this to make them appear like they have something going for them in life.











SHUX

 

[Poof]

Alot of the background behind that story is in a few threads over on Ars, particularly here. Apparently the reporter was not entirely accurate with some things however if you all want to get the scoop and make up your own minds, see the above thread.

I know that Ars and Anandtech have had a very frosty relationship between each other for some time. However I hope that some of the links showing what prompted the suspicions, helps in bringing the DC community together in general. The Berkeley group just doesn't get it and numerous emails to them privately have generally been ignored.

These alleged cheats and cheaters impact Anandtech in a big way as well because TA is rapidly moving up the ranks and would stand to possibly become an even higher ranked team with legitimate results should those b0rked results on SETI@NL be removed at some point.

I think everyone agrees that none of the Top 10 teams condones cheating and sadly, SETI doesn't allow Team founders to remove accounts that might be suspect, so the team gets tarnished.

 

[Smoke]

This is most disconcerting. I have not publicized the fact but now is as good a time as any to make mention of it. When the cheating topic came up a couple of months (or more) ago, even though our stats appear to me to be all honest and true to the integrity of the project, I wrote S@H and ask them to please verify that the top producers of our TeAm were all legit so I could report the finding to the TeAm.

I do not know how much work would be involved doing such a verification but to date I have not received a response of any sort. So when I read the part where the author of this article wrote,

SETI@home administrators are allegedly ignoring claims that the project is being sabotaged by miscreants who are threatening to derail its reputation and that of many valuable Internet-based distributed computing projects.,

I can only report my own experience parallels his comment.

Team Lamb Chops - 40 days and 40 Nights (Until Second Place)

Some interesting clips:

Added by Crossroads
quote:

Yes, there are cheaters, probably also with us. SETI@home will scan the database afterwards, to clean up the stats. Because seti@home 1 will only last a few months. It is known that seti@netherlands has at least 8 cheaters. For example:

6) p.rullenmand 1726 1726 1726 1726 1726 1726 1726 280181 1726.0
7) Turbo Duck 1797 1797 1797 1797 1797 1798 1797 248239 1797.2

So we have to crunch a fast as we can, but we will only know afterwards if it is enough.

And

added by rickman2
quote:

Don't think we (SETI@Netherlands) haven't reported these accounts yet. The numbers posted by Crossroads are indeed shameless.

All suspicious accounts (more than these 2) have been reported to Eric Korpela AND Dave Anderson. Heck, I've reported 'm about 5 times right now: Names, WU's and E-mail adresses

They haven't replied and I'm getting pretty pissed at Berkeley right now. They just don't seem to care, and don't even take the time to answer.

At this rate our team will be #1 within a few months, but this is definatley not the way I want to win!

There is lots to read and I recommend everyone take the time to do so.

 

[networkman]

Well, I can tell you this: Much as I like the SETI project, I will not be continuing with the new incarnation(S@H2 or whatever) if these issues are not seriously addressed.

Cheater accounts should be zeroed, the password changed, and the account names replaced with "I cheated - I'm a loser" :|

 

[LastKnight]

Yes, there are cheaters, probably also with us. SETI@home will scan the database afterwards, to clean up the stats. Because seti@home 1 will only last a few months. It is known that seti@netherlands has at least 8 cheaters. For example:

6) p.rullenmand 1726 1726 1726 1726 1726 1726 1726 280181 1726.0
7) Turbo Duck 1797 1797 1797 1797 1797 1798 1797 248239 1797.2

So we have to crunch a fast as we can, but we will only know afterwards if it is enough.

That's absurd. I've got about 50 systems loaded with the client, maybe a few more, and my numbers fluctuate by sometimes 50 Wu's a day or more, depending on a ton of different variables. Somebody dumping 6 times as many work units as myself, on consistently the same number is more than a little suspicious.

 

[FoBoT]

Originally posted by: Robor
The simple fact is there is no hackproof piece of software. If it can be programmed it can and will be hacked, it's just a matter of when and how.

just to be clear, the cheating referred to by the source in the article has NOTHING to do with hacks, i am not going to describe the whole thing here, but you don't have to hack the program at all, its all procedural

 

[Shuxclams]

I have seen the way that it is done, or it was explained on the forums awhile back. Basically people have cached the Wu's and resubmit them over and over in a neverending rotation. Kind of pointless to have that data being sent to Berkeley if they have already gone thru that data 100 times already. What a waste of bandwidth for us guys who do it strait.










SHUX

 

[Wiz]

Several of us on this team and others have dedicated about 3.5 years cpu time on any and all the computers we could get our hands on to the Seti@Home project.
I have crunched every WU on my own personal machines, not even using those belonging to anyone else so I could say that all WU's done in my account were processed on my own machines.
Integrity on a project like this is paramount. Not just for the sake of the project itself, but also for the sake of those who are partaking / competing in the project and for the sake of the projects that come after.
If DC is to survive IMHO those who cheat must be discovered, their accounts zeroed, some kind of "scarlet letter' attached to them and ideally they should be prosecuted. If not prosecuted, then at least persecuted.

 

[Robor]

Originally posted by: FoBoT

Originally posted by: Robor
The simple fact is there is no hackproof piece of software. If it can be programmed it can and will be hacked, it's just a matter of when and how.

just to be clear, the cheating referred to by the source in the article has NOTHING to do with hacks, i am not going to describe the whole thing here, but you don't have to hack the program at all, its all procedural

I didn't mean the client was hacked, although it could have been. I mean the process in general. There should be checks in place to toss out these bogus results the cheaters are submitting. It taints the project and ruins the fun for legit users and teams. It also gives teams with cheaters a bad name. It's pretty obvious S@NL and Ars aren't happy about the cheaters but if UB isn't going to do something about it what are they supposed to do?

 

[Poof]

I think that one of the critical points is that the leaders/founders of a team have no way to take action on their own, to rid the team of problem accounts. I think every team would like the ability to handle problems "in house" , basically they want to take care of it themselves - no fuss no muss. The fact that a SETI Team Founder has administrative ability to see the email addresses of every member and were recently given the ability to remove accounts that have been inactive for at least 2 weeks, means that giving the Founder that extra ability to remove suspect accounts, shouldn't be an issue.

With this project being as large as it is (and although they continually tout "over 4 million members", we all know that only about 400,000 - 500,000 are actually actively submitting WUs), one would think that they would allow some modicum of self-policing, with the administrative ability to do so.

And although many have noted that these particular bogus WUs (ie., ones resubmitted thousands of times) may not affect the science, they DO affect the hardware (drives) and database that they're stored in, ie., they take up valuable disk space and utilize extra resources that may force the project to spend more money on what would turn out as unnecessary upgrades...

 

[Evadman]

I agree with Poof. Just let thee admins kick people who cheat. Assuming that the group is more trustworthy than the whole, it will work.

I think all cheaters should be dragged out into the street and shot in the kneecaps. :|

 

[Smoke]

Poof wrote

The fact that a SETI Team Founder has administrative ability to see the email addresses of every member and were recently given the ability to remove accounts that have been inactive for at least 2 weeks, means that giving the Founder that extra ability to remove suspect accounts, shouldn't be an issue.

I certainly wouldn't want to test it but it appears "any" member can be removed. Once a "founder" enters the edit mode there is a block next to each member's name that can be checked. Then all that is necessary is to click on the REMOVE button. Heck, there is even an option to DISBAND the TEAM. :Q

So it appears a founder can remove anyone. If that is the case I wonder why the "founder" of any team wouldn't just remove a cheater from the roster? Of course even though the option to remove any member appears to be there it may not actually work as Poof has indicated.

Just curious, Poof, where can I read about this?

 

[FoBoT]

if you want a test subject to see what happens, you can use my account

it just forces me off the team right? or will it not let me back on?
i guess somebody could create a new account, do one WU and then you could try it on that account

 

[Smoke]

FoBoT, you've got a PM.

 

[Smoke]

Without resorting to such dangerous activity I have found the answer. When I entered the "founders" mode I could only find the names of "inactive" members so Poof as usual is 100% correct.

If "founders" or those with "founders rights" were given the ability to remove "any" member, I believe the cheaters could be dealt with quite rapidly. :Q

 

[Poof]

Originally posted by: Smokeball
Without resorting to such dangerous activity I have found the answer. When I entered the "founders" mode I could only find the names of "inactive" members so Poof as usual is 100% correct.

If those "founders" or those with "founders rights" were given the ability to remove "any" member, I believe the cheaters could be dealt with quite rapidly. :Q

Sorry for the delay Smokeball... Things are getting hot and heavy on the SETI forums now... eg., see:

http://setiathome.ssl.berkeley.edu/bb/bb2/index.cgi and http://setiathome.ssl.berkeley.edu/bb/bb4/index.cgi.

I believe the thing about the 2 weeks inactive was posted by SETI too...

 

[Smoke]

Man, what a read. I'm not sure how many of you have the time or the inclination to read through all of the posts and replys on the Ars forum but here is the current situation as I see it.

There has been an "open letter" to Seti@Home drafted and you can sign it as sort of a petition. I do not see any signators from TeAm AnandTech SETI as yet and thought we might want to discuss this openly a little ... and of course as usual we are all free spirits to do as we wish.

The current "open letter" is as follows:

Open letter to Seti@Home, Berkeley

Dear friends,

We, the undersigned Seti@Home crunchers have been alarmed by recent appearance of a number of very suspicious user accounts with abnormally high daily output of completed work-units (not hundreds but thousands!). Although it's hard to be certain, everything indicates that those users are blatantly cheating by taking advantage of security holes on the Seti server.

We would like to ask you to kindly investigate these highly suspicious accounts and undertake an appropriate action.

We feel that cheating is not only ridiculing our previous efforts but also damaging the whole Seti@Home project. Furthermore, if the action to stop cheating is not taken many of us will permanently leave the Seti project.

Thank you.

It is now just a thread on the Ars Technica Forums. I have posted a question if it will be forwarded in some fashion to the S@H Officials.

If you want to sign the petition, GO HERE!

 

[Smoke]

The "open letter" will be forwarded with all signators to Seti@Home. IMHO, we need to show our support.

 

[CADsortaGUY]

:|Signed it :|

CkG

 

[networkman]

I signed the petition too!

 

[Wiz]

signed it

 

[Robor]

Signed and commented. I like the idea of letting founders kick potential cheaters as well.

 

[Evadman]

signed it too.