Something very strange with the Test Range stats

[MikeH]

Something very strange is going on with the test range statistics (http://www.seventeenorbust.com/stats/rangeStatsEx.mhtml) .

Both 24737 and 55459 show the current test window max n as '9999999'. This is very strange (and maybe suspicious) because 9999999 is not in the current sob.dat file, i.e. we already have factors for these candidates.

So I guess there are two possibilities. Someone that doesn't know what they're doing has put these k/n pairs directly in the registry and is performing a pair of pointless tests. Or maybe it's to hide something else......;)

[Alien88]

| 389155 | Dropped | Unexpected | 1099171918 | 1099175530 | 42500 | NULL | NULL | 68.95.250.153 | 2.0 SSE2 | 5/1/2600/521 | 83 | 7239 | 10223 | 9999999 | 4.25e+09 | NULL | NULL |
| 389175 | Dropped | Unexpected | 1099174520 | NULL | 12500 | NULL | NULL | 68.95.250.153 | 2.0 SSE2 | 5/1/2600/521 | 83 | 7239 | 99999 | 9999999 | 0 | NULL | NULL |
| 389173 | Pending | Late | 1099174242 | 1099711622 | 1790000 | NULL | NULL | 68.95.250.153 | 1.2.5 | 5/0/2195/2054 | 83 | 7239 | 55459 | 9999999 | 1.79e+11 | NULL | NULL |
| 390944 | Pending | Unexpected | 1099528425 | 1099634780 | 1737500 | NULL | NULL | 68.95.250.153 | 2.0 SSE2 | 5/1/2600/521 | 83 | 7239 | 24737 | 9999999 | 1.7375e+11 | NULL | NULL |

[Alien88]

Apparently user 7239 (Beerknurd) is having fun with fake tests.. i've removed his tests for now.

[Beerknurd]

I was doing it to prove a point. Some other users didn't believe that it could be done. It is not my intention to keep this going. I was only pointing out a flaw in the system. If you can use the registry to change your k/n pairs then how many other people are doing this. My 1k and 9999999n tests are a little obvious to try and get away with. If I was trying to beat the system I would have used a number less obvious. Please reactivate my account and I will not run that anymore. All I was trying to do is prove a point.

Plus, I have only been running at 2 M. Who would beleive that I would aquire enough CPU power to push me over the 30 M barrier. That alone seems pretty suspicious.

[Alien88]

Originally posted by Beerknurd
I was doing it to prove a point. Some other users didn't believe that it could be done. It is not my intention to keep this going. I was only pointing out a flaw in the system. If you can use the registry to change your k/n pairs then how many other people are doing this. My 1k and 9999999n tests are a little obvious to try and get away with. If I was trying to beat the system I would have used a number less obvious. Please reactivate my account and I will not run that anymore. All I was trying to do is prove a point.

Plus, I have only been running at 2 M. Who would beleive that I would aquire enough CPU power to push me over the 30 M barrier. That alone seems pretty suspicious.

I'm sorry, but that mentality really pisses me off. If you are trying to prove a problem with our system, perhaps you should be kind enough to contact us privately instead of doing it to "prove a point".

[Beerknurd]

Well I apologize if I didn't bring it up to the right people. I'm sorry if the way I did it was wrong. It was not my intent to piss people off. So for what it's worth I'm sorry.

[vjs]

I'm sorry, but that mentality really pisses me off.

I can see your frustrations this problem is two fold.

First on the part of Beerknurd he is not really to blame, yes he has been submitting garbage to prove a point and this is just wrong. However I can see why he has done this.

Not trying to be devils advocate here...

Part of the problems with all DC stem from berkley's seti program. First, A fantstic proof of a DC projects ability in terms of performance, and second a rude awakining socially for ones belief in the possibility of ET by the shear number of participants.

ON the negatives, if you really look at the science it's a total sham. Going back to the higher archie of planatary scientists and their total disregard to the possibility of ET.

Berkley in not looking for ET they are looking at Hydrogen drift!!!

Users are very frustrated, cheating is rampent, SETI staff does nothing. Several years ago and even today the project co-ordinators do nothing about about cheating and the only way end-users were able to do anything was by malisously attacking a hole in the system to such a serious degree it was harmful.

Then and only then would co-ordinators even acknowledge the hole and potentially do something.

Now with SOB, cheating will happen from time to time, especially as k/n pairs become larger and the project more popular.

However what the end-users must understand is the project is more sensitive than seti to cheating. I would also hope that the end users who do or are cheating are smart enough not to harm the science of the project while doing so.

Second, end-users must understand that the project co-ordinators, are actively involved in this project, the forums, and are more than willing to help out with any problems and appreciate advise etc. This fact in its self is the only reason why I'm with SOB.

A lot of problems with DC stem from stats if co-ordinators can control awarded points cheating approaches 0.

With SOB this can be easily done, only give credit to a user for server assigned k/n pairs.

A users could still enter a k/n pair do the work etc, but he wouldn't be given credit for work done.

Second even if a K/N pair has a submitted residue, if that pair has not been assigned by the server it's not removed from the que.

Example,

I decide to test the time required to do k=67607 n=754000 (assuming this is a correct k/n pair), submit the reside. Even if this test is done 5 times by various people. First I on one get points, second the residue(s) are stored, third when the n value reaches n=754000 the server assigns the value to user ?, the user completes the pair like normal.

Now all residue could be stored for comparison to eliminate a double check, or tripple check in the case of supersecret, (if they match).

Thanks for your time.

[vjs]

Also,

Beerknurd, modification of the registry is not a problem, and people have been doing this for quite some time. Look at the thread sticky "How to fix a lost test", it actually explains how to do it.

The problem is that you were using it malisously, creating your own k/n pairs.

I personally have a program that works similar to SOB-Que (doesn't require a net connection), one can enter the k/n pairs of your chosing for testing and later submit the residues.

It's not difficult nor earth shattering.

I also propose that all users caught cheating by this or any other method have their stats reset to zero. Even for tests completed correctly.

[Beerknurd]

Well I regained access to the site. Thank you to who ever restored it. I will be crunching only k/n pairs assigned to me. But I will restate again it was not my intention to cheat, but to expose a problem. I am a very honest person and don't want to corrupt anything going on here. But it does make me feel better to know that there are admins that care and do look at the stats. So hopefully no one else is taking advantage of this bug. I am sorry for the problems I caused. If I find anything else wrong you can be assured that I will contact the correct people next time.

Thanks,

Beerknurd
:cheers:

[Beerknurd]

Originally posted by vjs


The problem is that you were using it malisously, creating your own k/n pairs.


That's a bit harsh. It was hardly malicious. I wasn't trying to beat the system for my own gain. Honestly, going from a 2M producer, to a 30M producer overnight is not someone who is trying to get away with it. I could have easily changed it to 3 or 4 M if I wanted to be sneaky about it.

All I wanted to do is expose a problem. I'm sorry if I went about it in the wrong way.

Beerknurd
:cheers: