New CM's

Discuss your experiences with and ideas about Stardust@home here.

Moderators: Stardust@home Team, DustMods

Post Reply
Star Gezzer
Posts: 86
Joined: Fri Jan 26, 2007 3:24 pm
Location: Newport Oregon

New CM's

Post by Star Gezzer »

I would like to thank who ever changed out the CM's. You leveled the playing field between the HONEST dusters and the programers. Now that you have caught them what will be done?

jsmaje
Posts: 616
Joined: Tue Aug 15, 2006 8:39 am
Location: Manchester UK

Re: New CM's

Post by jsmaje »

Star Gezzer wrote:I would like to thank who ever changed out the CM's. You leveled the playing field between the HONEST dusters and the programers. Now that you have caught them what will be done?
Star Gezzer, I don't quite understand what you're saying. What do you mean by "changed out"?

Are you referring to the CM movies themselves or, as I suspect, our scores for correct CMs? I did happen to think my own score jumped up by about 20 a few days ago, but put it down to forgetting about my own previous activity. Have I missed some official announcement, or have others noticed any similar unexpected increase?

Are you saying that the team have indeed increased our scores, having "caught" (and presumably reduced the scores of) "the programers"? If so, on what evidence, and to whom exactly are you referring? If our old pal KAC, I notice he is still sailing ahead at full steam, having amassed at least another 3400 points in the last 5 days!

Yours, intrigued :?
John

fjgiie
DustMod
Posts: 1253
Joined: Sat May 20, 2006 8:47 am
Location: Hampton, SC, US

Re: New CM's

Post by fjgiie »

Let me add something. Star Gezzer and I talked about the cm’s. We think that the team has added some challenging ones lately. These below may be old but show what challenging means.
This CM seems to have two tracks.
I clicked on the one on the left and got a correct answer. I think the one on the right looks more like a real track might look.

46003V1 Bottom left corner. That is what Star Gezzer calls small and washed out. A computer may miss that one altogether.

One more 394, 91

My glasses must be dirty and scratched up pretty bad as I have missed 40 CMs in just the last few days. I imagine a computer would have more trouble finding them than I do. :shock: Besides I don't really NEED glasses. I only use them when I want to see something!

fjgiie

jsmaje
Posts: 616
Joined: Tue Aug 15, 2006 8:39 am
Location: Manchester UK

Re: New CM's

Post by jsmaje »

Fjgiie, those CM examples are certainly challenging. I wonder what would happen if you clicked the one on the right in the first (can you please ask the team on our behalf)? The second seems to be yet another deliberate Moe-lookalike, and I admit needing your coordinates for the third. But how can you be sure they've only recently been introduced, other than there having been no postings before?

I guess a sufficiently sophisticated program could still find such tracks, simply because they are isolated features lasting several bars, regardless of how faint, fuzzy or small they may be. Sorting out such subtleties if they happen to be overlayed by the ubiquitous surface gunge and distortions of 'real' movies is an altogether greater problem, and one that I fear even the best Starduster's eyes or programs are unlikely to achieve with any reliability. If they neverthess did, such hits would surely be lost among the other hundreds of 'No track' votes, so is it likely the team could ever get round to examining them in our lifetime?

Given the prediction of only 45 IS particles having arrived over the whole collector, and the experience so far of how much of the surface is terminally obscured by distortion and debris, it should be possible to predict how many of those tracks, subtle or otherwise, are now likely to be detectable. I'd guess 20 at most; more likely fewer than 10 :shock: :?:

Meanwhile, I'd still like to know who Star Gezzer is referring to by the words "honest" and "caught".

John

Star Gezzer
Posts: 86
Joined: Fri Jan 26, 2007 3:24 pm
Location: Newport Oregon

Honest & caught

Post by Star Gezzer »

I would consider anyone honest that is not using a program to find CM'S. I would consider someone caught if they produce more points than resonably possible especially with the new CM"S. The examples given by Fjgiie are only a few of the new CM"S. I would guess that over two thirds of the CM'S were replaced that I have seen over and over again. All were made harder which I am ok with. It will take the programers a while to update their programs depending what they are doing. If they are using the movie ID to know which have CM"S they would slow down very little. One in particular is able to put out about 800 points in four hours just using rough percentages that is over 3200 views taking 3-4 seconds each if no mistakes are made with no breaks and they were putting out 1000 to 1500 points a day. I would be very interested to see the stats on three people. So if you are honest show me your stats! I am certain it would be very revealing. Anyone that conceals their stats is suspect. Whatever that means.

jelfving
Posts: 12
Joined: Tue Aug 08, 2006 2:34 pm

Post by jelfving »

I would say - in all friendliness: stop quibbling about CM:s and points, and concentrate on searching for stardust!

the moon
Posts: 177
Joined: Thu Aug 03, 2006 12:34 am

Post by the moon »

The CMs and points actually are important to the searching process. Especially now that the duster base is much smaller, if automated programs are accounting for a larger majority of total movies viewed it dilutes the real work dusters do because the programs are so poor at finding real tracks and only mediocre at finding CM tracks.

It could be the reason you see sometimes see people like elainekeefe just posted that he was the only person to click this obvious track out of 22 views.
http://stardustathome.ssl.berkeley.edu/ ... =9276534V1

Star Gezzer
Posts: 86
Joined: Fri Jan 26, 2007 3:24 pm
Location: Newport Oregon

Quiblying

Post by Star Gezzer »

There is a difference between a discusion and quiblying. Facts are facts


Calibration Movies Answered Correctly 130628
Calibration Movies Answered Incorrectly 1228
Your Overall Score: 129400
Total Real Movies Viewed: 515580
Your Rank: 2 out of 26237
Specificity: 99.96%
Sensitivity: 98.18%

I have looked at over 500000 movies this phase so that takes care of looking for star dust.

And if CM'S do not bother you in the next week or so your sensitivity will suffer. Just a friendly thought.

Nikita
DustMod
Posts: 994
Joined: Wed May 17, 2006 8:33 pm
Location: Indiana, USA

Post by Nikita »

I still don't see how one program can skew the data in a way that would invalidate our searches. If the program does a poor job, it will be noticed and the results weighed accordingly. This should be especially true with a smaller volunteer base, the team should be able to sift through the results quicker and determine who produces the most likely canidates. The program is designed to protect against inaccurate but high numbers.

On the other end, what if there is a classroom searching and posting a high percentage of movies viewed? School is beginning again for the folks in the north and I know of at least one classroom that is considering such a possibility.

We have gone over this before and the data seems to still show that dusters are the best resource. If the concern is about the results being messed up because of a program, I think we can relax. Especially with Phase III and a probable renewed interest in the project.
From dust we come

jsmaje
Posts: 616
Joined: Tue Aug 15, 2006 8:39 am
Location: Manchester UK

Post by jsmaje »

the moon wrote:It could be the reason you see sometimes see people like elainekeefe just posted that he was the only person to click this obvious track out of 22 views.
Sorry to add to the quibble quota, but I'm sure the last time Elaine checked 'he' found herself to be a 'she'! :)

elainekeefe
Posts: 190
Joined: Fri Aug 04, 2006 4:38 am
Location: Massachusetts, USA

Post by elainekeefe »

jsmaje wrote:
the moon wrote:It could be the reason you see sometimes see people like elainekeefe just posted that he was the only person to click this obvious track out of 22 views.
Sorry to add to the quibble quota, but I'm sure the last time Elaine checked 'he' found herself to be a 'she'! :)
Hey, my own personal gender defender! Thanks, John. :D

Star Gezzer
Posts: 86
Joined: Fri Jan 26, 2007 3:24 pm
Location: Newport Oregon

Gender defender

Post by Star Gezzer »

Gender defender Now that is funny. You better register that one.

bmendez
Stardust@home Team
Stardust@home Team
Posts: 530
Joined: Wed Apr 19, 2006 11:28 am
Location: UC Berkeley Space Sciences Lab
Contact:

Post by bmendez »

Hi All,

So far as I am aware, I don't believe any new CMs have been added to the system since we began Phase 2. If you are seeing ones that you hadn't before my first inclination is to think that it is just the random nature of randomly picking from the pool of thousands of CMs.

Also, if there is anyone who is using an automated program to dust, they have not told us (although we would like to know if you are).

We have always encouraged people to try their hand at creating an effective automated search tool. So we do not consider using such a program dishonest or cheating.

If, however, someone is exploiting the various ways of determining if a movie being viewed is a CM, whether with automated software or not, that does smack of dishonest. Though, as I have pointed out many times before, it really only cheats that person. So what if they are pumping up their scores. If they aren't really searching for stardust, they aren't going to be clicking on real movies. So they don't really effect the project too much other than to waste our bandwidth (which is annoying).

I should also point out, that we have never had any evidence of anyone doing this. So...there you go.

Happy dusting and thanks to everyone for their efforts,
-Bryan
"I am made from the dust of the stars, and the oceans flow in my veins"
- RUSH

bebop
Posts: 13
Joined: Wed Jan 30, 2008 3:16 pm
Location: Florence, Ita

Re: New CM's

Post by bebop »

Hi.
fjgiie wrote:These below may be old but show what challenging means. [...]
Here's a tricky one (got it right, but I wasn't sure untill I saw my score increased by one point): CM 3072225V1

Bye
b

Post Reply