Am I still important?

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

Moderators: Stardust@home Team, DustMods

Post Reply
gebils
Posts: 62
Joined: Mon Jun 07, 2010 1:25 pm
Location: Colorado, USA

Am I still important?

Post by gebils »

I have been viewing now for over a year. Not one of my events has been processed. The last 19 of my events have been previously viewed by over 900 others. Am I contributing or wasting my time?
fjgiie
DustMod
Posts: 1253
Joined: Sat May 20, 2006 8:47 am
Location: Hampton, SC, US

Re: Am I still important?

Post by fjgiie »

gebils wrote:I have been viewing now for over a year. Not one of my events has been processed. The last 19 of my events have been previously viewed by over 900 others. Am I contributing or wasting my time?
Hi gebils,
Here is a copy of an email I just received from ERSTRS: (a high scoring duster)
_____________________________________________________________________________________

"Hey, Fred, I read the post by gebils, "Am I important?" I'd love to post an answer, but the I've "exceeded the maximum. . . ." message won't let me. I've tried every combo of name and PW that I've ever had for the Forums, and none will let me Post. However, I CAN still dust, so I'll await the new Phase 4, hoping that all of the tech problems will be solved by that time.


I would like to say to "gebils," that YES, we are important. But with few funds, and mostly all volunteers, there is little money or time for feedback to dusters. For instance we don't know how Orion has turned out---stardust or just plain old dust---and we are still harshly punished when we make a mistake with a CM. Very discouraging because of how long it takes to pick ourselves up off the floor and get points back when we do err. For instance, I've read more than eighty-five-thousand movies since I joined Phase 3. Missed 145 of planted Tracks in CM's. Yet my score is just 98.26%. I made the last error more than a week ago---yet I have not gained back even .01 point yet. I feel like I'm being punished "forever" for each mistake I make! Hugely discouraging!
Also, my Specificity is less than 100% because I took Stardust's admonition , "Expect the unexpected," seriously, and clicked on what I thought looked very much like an Orion Track in a 50micron CM. Thought it might have been missed, and it might be one of those "unexpected" things. OOPS! Wrong decision. Lost points. So, I resolved to never, ever, click on anything in a 50micron movie again, even if it looks exactly like an Orion Track! Therefore, it appears to me that something is badly wrong with the present program of offering "rewards" for finding Tracks in CM's. The present system leads me to consider my position in the Ranking more important than finding Stardust particles. I'm motivated toward keeping my position as number "20" more than anything else! And, I feel like I'll be punished if I'm an innovative or a free-thinking duster. Like you, I get sick and tired of all the repeated movies---have them all memorized in fact---so I just click them off as "No Track," as they pass by me daily. This has got to stop.


I'm awaiting Phase 4, hoping and trusting that the whole system will be revised, and that my motivation will return to finding Stardust----to heck with whatever Rank I happen to be in!
SO! I say all of the above to say this to gebils: Hang in there until Phase 4 arrives. I'm hopeful that it will be the end of movies that have been viewed hundreds of times by each duster, and that we'll connect with and get updates from the Stardust Team more often. So, YES, you and I are definitely "still important." I think the new Phase 4 will prove that to be the case.


Statistics for ERSTRS

Calibration Movies Answered Correctly 16161
Calibration Movies Answered Incorrectly 145
Your Overall Score: 16016
Total Real Movies Viewed: 84983
Your Rank: 20 out of 30178
Specificity: 99.99%
Sensitivity: 98.26%"
__________________________________________________________________________
jsmaje
Posts: 616
Joined: Tue Aug 15, 2006 8:39 am
Location: Manchester UK

Re: Am I still important?

Post by jsmaje »

I share gebils’ frustration and Evelyn’s take on matters. Assuming the team still think a distributed project to be the best way to achieve their scientific objectives and phase 4 ever materialises, I can only hope that its design takes more account of the psychological aspects of volunteer-dusting than hitherto (critically including ready and helpful feedback).

One indicator of how the project has progressed in terms of duster activity is the following sequence of graphs of Top 100’ dusters scores according to rank at the end of phases 1 & 2 and the present month of phase 3:

Image

Though regarding just the top 100 they exemplify how initial interest in the project was high during phase 1 when ~20,000 world-wide volunteers were recruited, then settled during phase 2 once temporary interest turned into commitment amongst a smaller but dedicated number of dusters, to the present (according to DanZ) fewer than 1000 still contributing.
Overall scoring last week was just 10% of the highest during phase 3. In fact only 34 of the top 100 dusters did anything last week and, while over 90% of scoring has usually been by just the top 5 enthusiasts, none of them have bothered either!
Do they know something I don’t, or perhaps all gone away on holiday together?

Meanwhile, for those interested in the mathematical forensics:
I first became curious about the shape of the graph during phase 2, and suggested it could be a good fit for a so-called ‘Zipf-Mandelbrot’ distribution, characterised by the log-log slope being close to minus 1.0 throughout its majority extent.
A surprising range of unrelated phenomena share this property, both natural (e.g. earthquake distribution) and human behaviours (e.g. word frequency, web-site visiting etc.), the only common factor seeming to be a large number of events happening for the same underlying reason(s) but operating without significant interaction, collaboration or competition.
As shown on the graphs, phase 1 ended in June 2007 with a slightly lower but not-too-different slope of –0.71, and phase 2 in March 2010 at a convincing –1.15. Phase 3 began soon after at –1.19, but has steadily drifted to the nowhere-near-ZM slope of –1.83.
We don't need such figures to appreciate that the Z-M premise of large numbers operating independently has not been maintained, but perhaps there's a behavioural analyst out there who may find them useful to plug into their model?


John
DanZ
Site Admin
Posts: 777
Joined: Fri Feb 27, 2009 2:44 pm
Location: Berkeley, CA
Contact:

Re: Am I still important?

Post by DanZ »

You are VERY important gebils. The numerous viewings per movie you're seeing has also been brought up by others recently and I'm trying to get to the bottom of it. Could just be a computer glitch, but I promise to have an answer for you soon. Regarding "Not one of my events has been processed" (and assuming you mean viewed/commented on by the Berkeley team), I'll make sure something happens there soon.

Regarding the negative scoring aspects of the current system - it is definitely one of the things we are trying to address in the new Phase 4. We just had a meeting with the team on Friday and are now shooting for a Phase 4 roll out in early July. Please be patient. One of the problems though is that one of our computer techs got assigned to a new project and has less time for Stardust.

Finally, I know it feels like we have "vanished" but it's not true, trust me. Nevertheless, thanks for the push. I sincerely hope we can accommodate!

Dan
PS - Look for science updates from Dr. Westphal soon.
gebils
Posts: 62
Joined: Mon Jun 07, 2010 1:25 pm
Location: Colorado, USA

Re: Am I still important?

Post by gebils »

I understand that a lot of projects are very underfunded but that does not help when I wonder if what I am doing is productive or a waste of time. Today, after seeing the previous postings I spent 4 hours dusting and located one very interesting track. After posting it I find that 1013 others ahead of me have posted it and nothing has be done to verify, test or whatever. I cannot accept the fact that my time is productive and I will suspend my efforts until the principals involved get their ducks in line. The project has so may potentials it is a shame to waste the time and effort of the dusters. I will continue to check back to see if I can once again find a reason to contribute to this effort.
DanZ
Site Admin
Posts: 777
Joined: Fri Feb 27, 2009 2:44 pm
Location: Berkeley, CA
Contact:

Re: Am I still important?

Post by DanZ »

gebils wrote:Today, after seeing the previous postings I spent 4 hours dusting and located one very interesting track. After posting it I find that 1013 others ahead of me have posted it and nothing has be done to verify, test or whatever. I cannot accept the fact that my time is productive and I will suspend my efforts until the principals involved get their ducks in line. The project has so may potentials it is a shame to waste the time and effort of the dusters. I will continue to check back to see if I can once again find a reason to contribute to this effort.
Thanks for bringing this up gebils. There is no doubt the Berkeley team has gotten behind in screening events, mainly due to recent lab analysis work and the amount of time spent trying to design Phase 4. However, Dr. Westphal has agreed to look at the high-fraction hits soon.

On the other hand, we want this to be fun/interesting for you, and we think Phase 4 will give everyone some new incentive to continue. Therefore, we hope you stay in touch and give Stardust a try again when we re-launch.

Best,

Dan
DanZ
Site Admin
Posts: 777
Joined: Fri Feb 27, 2009 2:44 pm
Location: Berkeley, CA
Contact:

Re: Am I still important?

Post by DanZ »

gebils wrote:Today, after seeing the previous postings I spent 4 hours dusting and located one very interesting track. After posting it I find that 1013 others ahead of me have posted it and nothing has be done to verify, test or whatever. I cannot accept the fact that my time is productive and I will suspend my efforts until the principals involved get their ducks in line. The project has so may potentials it is a shame to waste the time and effort of the dusters. I will continue to check back to see if I can once again find a reason to contribute to this effort.
Okay, more follow up. Dr. Westphal discovered that the movie you refer to, 2157740V1, has been viewed 1016 times but only clicked on 19 times. So, it's definitely a bug in the system that is showing this and other movies at such high number of viewings in the MyEvents pages (the number should be only 19). But we're still hunting for the reason why.

On a side note, we hadn't looked at it or commented on it yet because it didn't rise above our first trigger of about 15% of those who see it clicking on it -- the signal for this was ~1.9%. (And in case you are wondering, it is not a track, but is a surface feature that superficially resembles a track).

Otherwise, thank you so much for bringing this to our attention! We will continue to investigate.

Dan
DanZ
Site Admin
Posts: 777
Joined: Fri Feb 27, 2009 2:44 pm
Location: Berkeley, CA
Contact:

Re: Am I still important?

Post by DanZ »

DanZ wrote:
gebils wrote:Dr. Westphal discovered that the movie you refer to, 2157740V1, has been viewed 1016 times but only clicked on 19 times. So, it's definitely a bug in the system that is showing this and other movies at such high number of viewings in the MyEvents pages (the number should be only 19).
Okay, I got that wrong. The over 1,000 viewings for this and other movies is because, yes, a reduced subset of movies were being continually cycled through the VM. But we've fixed that and you should be seeing many more movies available for dusting right now. However, let me know if that's not the case!

Sorry for the bug and thanks for your patience on the fix.

Dan
PS - the "Number of Viewings" numbers of around 1,000 are thus correct for the movies that were being recylced.
tiburd
Posts: 24
Joined: Thu Aug 03, 2006 7:47 pm
Location: Wyoming

Re: Am I still important?

Post by tiburd »

I have been assuming that the team will eventually check out every movie flagged by a plausible number of dusters, including going back over the input from Phases I, II, and III. The core data are of sufficient importance that this is the only approach that would seem to make sense, even if it takes a good while. After all, the only reason to have dusters at all is to point out possible tracks, and ignoring that input would contradict the whole purpose of the project.

Having said that (which I assume is an affirmation, but if I'm wrong then it would be a stern rebuke), I think the raw % of dusters flagging a movie is a very crude screen for selecting which movies to check. A better (and easily programmed?) approach might be to weight the "flaggers" by some form of their score, such as (in the prior phases) selectivity. A duster who has never been wrong on a calibration movie (for example) is unlikely to be wrong when flagging a "real" track, even if he/she is the only one who saw it.
tiburd
gebils
Posts: 62
Joined: Mon Jun 07, 2010 1:25 pm
Location: Colorado, USA

Re: Am I still important?

Post by gebils »

tirburd,
I sorta got a feeling that what you suggest is indeed being practiced by the people involved with reviewing our results.
George
DanZ
Site Admin
Posts: 777
Joined: Fri Feb 27, 2009 2:44 pm
Location: Berkeley, CA
Contact:

Re: Am I still important?

Post by DanZ »

Well, not all high scorers are good at finding real tracks (in the past phases at least). In fact, some of the highest scorers missed real tracks that were presented to them on more than one occasion. But we also developed the "Red Team," who based on their skills in identifying tracks, constantly promote movies to the science team regardless of how many clicks. For instance, it was a member of the Red Team that promoted and strongly urged us to look at Orion, which then resulted in Phase 3 and the new search image training for so called "midnight tracks." Had that not happened, then this particle (and some that followed) probably would've taken much longer to get attention. Thus, there's a little bit of everything going on. But you're right, and I think "eventually," most anything that shows some promise in one way or another, will be looked at.

Dan
tiburd
Posts: 24
Joined: Thu Aug 03, 2006 7:47 pm
Location: Wyoming

Re: Am I still important?

Post by tiburd »

Super, I don't think anyone could ask for anything more.
tiburd
Post Reply