seven people thought so... don't know why though

Archived here are older posts which are no longer relevant or were redundant.

Moderators: Stardust@home Team, DustMods

Orion_0169
Posts: 153
Joined: Wed Aug 02, 2006 8:17 pm

seven people thought so... don't know why though

Post by Orion_0169 »

http://stardustathome.ssl.berkeley.edu/ ... e_id=52935

So, I flagged this one, just to see what kind of responses it might have. This movie, to me, is just coming into focus in the last layer or two. I would select it as bad focus, since I can't get below the surface. But I flagged it to see...

Seven other people flagged it. That puts it in the top ten of my events for #oa.

Sigh...
Je ne peux pas regarder la Mer sans me demander qui vit au-delà de cela.

mwhiz
Posts: 95
Joined: Tue May 23, 2006 3:58 pm
Location: Seattle, Washington, USA

Post by mwhiz »

golly. that's a bad sign. that's either a lot of misclicks or... i dunno what. i hope the stardust team can statistically weed stuff like that out. golly. :shock:
"The Earth is the cradle of mankind, but one cannot live in the cradle forever."
~Konstantin Tsiolkovsky

Orion_0169
Posts: 153
Joined: Wed Aug 02, 2006 8:17 pm

Post by Orion_0169 »

I've been hinting at it for a while now. Based on the number of threads that people are starting, it wouldn't surprise me if every movie has been flagged by now. And right now, it's only going to get worse, I suspect, as whatever issues exist are just being perpetuated. The top 100 people alone, as of early this morning have gone through maybe 300 000 movies. And now there are over 8000 people going at it. If people are flagging movies like this...
Je ne peux pas regarder la Mer sans me demander qui vit au-delà de cela.

mwhiz
Posts: 95
Joined: Tue May 23, 2006 3:58 pm
Location: Seattle, Washington, USA

Post by mwhiz »

well, i did some calculations myself... i don't think its toooo bad, remember, they don't have to look at ones with only a few hits. i mean, i have plenty with more then 15 hits, so i think 7 is pretty far down the list. also, i think the top 100 have gone through about 150,000.
"The Earth is the cradle of mankind, but one cannot live in the cradle forever."
~Konstantin Tsiolkovsky

Orion_0169
Posts: 153
Joined: Wed Aug 02, 2006 8:17 pm

Post by Orion_0169 »

That movie I mention has 7 out of only 21 views. That's one third of the people that have seen it. That's huge.

As for the top one hundred. As of early this morning the combined scores of teh top one hundred was 99 582. Assuming a 1:3 ratio for callibrated to real, multiply that number by three to get 298 746 real movies.

Just the fact that you have "plenty" with 15+ is telling, given how few streaks there are expected to be.
Je ne peux pas regarder la Mer sans me demander qui vit au-delà de cela.

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

Post by the moon »

I can see why people click that one. The blob in the middle at the bottom looks like the tracks in the examples except it's not under the surface of course. I'm guessing a lot of people are missing the concept that tracks always extend under the surface.

And yes, probably every movie that's in focus has been flagged, but since they start with the most often flagged, if there is any real dust in the 40,000 or so movies, I'm sure it will be in the top 2000 movies flagged. So we've still saved them a lot of time.

Orion_0169
Posts: 153
Joined: Wed Aug 02, 2006 8:17 pm

Post by Orion_0169 »

http://stardustathome.ssl.berkeley.edu/ ... e_id=19603

Another. Same thing.

The problem is not enough people knowing what they are doing. Or rather, far too many that don't. Too many don't understand that a ring getting smaller, coming to a point, then getting bigger, is precisely what to expect when a point is taken in and out of focus. As opposed to finding things IN focus.
Je ne peux pas regarder la Mer sans me demander qui vit au-delà de cela.

Belinda
Posts: 14
Joined: Sun Aug 06, 2006 4:08 am

Post by Belinda »

You have relieved a little worry of mine, I am also flagging as 'bad focus' those movies where the surface is too close to the bottom of the range.
I think it will all come out in the wash though, wont it? if people are incorrectly flagging a movie it just means it will get viewed by an expert?

Something which would be good though is a 'probably not a track, but something weird to check out' button! maybe some people are flagging the movie as strange rather than potentially containing a track, because the organisers have advised us to do that.
Some method of differentiating the clicks would be good, maybe even a 'comments' box (optional of course) for each movie?

Cheers
Belinda

Orion_0169
Posts: 153
Joined: Wed Aug 02, 2006 8:17 pm

Post by Orion_0169 »

Thought about that actually.

A button for...

Dots.
Fuzzy rings.
Neat artifacts.
And ballerinas of course.

;-)
Je ne peux pas regarder la Mer sans me demander qui vit au-delà de cela.

Driven1
Posts: 20
Joined: Sat Aug 05, 2006 7:27 am

It's a result of insufficient training...

Post by Driven1 »

that stuff like this happens. Face it, probably most of the people doing this have never seen a microscope since biology class in High School.

There are tons of poorly focused movies and some people don't seem to be able to identify what movies are usable and which aren't from a search standpoint.

In the Tutorial, the stress was put on finding a track using doctored images of what the team kind of expected to see. Those of us who have found real suspected tracks are well aware that the tracks so far, look nothing like those in the training in most cases. It was also not stressed hard enough that there were many different flaws with movies that could/would make them unusable.

The movie that you gave an example of is definately poor focus. You're just at the surface at the end of the Focus Range leaving no range to "drill down" deeper to find actual tracks. I've hit the "Poor Focus" button on tons of these. I've even hit it when there is less than 35 to 40% of usable focus range from the surface focus.

There are movies like this one. There are movies that were shot slightly off kilter giving a sort of "focus wave" effect, there are movies of pieces of Aerogel with uneven surfaces that are unusable, there are some that are just flat out out of focus, there are some where two different movies are combined!

I think the Team should start throwing some of these flawed movies in as test movies as well as some real movies with known verified tracks as tests. The test movies are soooo obvious.

If nothing else, it might better train those who may not be that familiar with the use of a microscope or what is usable and what is not and get more accurate results for actual tracks.

I will admit that I did flag one that was (still is) questionable. No one alse has agreed with it and that's fine. It shows what appears to be a large horizontal track. The only reason I flagged it was because of it's shape. It is almost a perfect ballistic shape as if something large came in from the right side, slowed down, and stopped.

My 2 Cents.

mwhiz
Posts: 95
Joined: Tue May 23, 2006 3:58 pm
Location: Seattle, Washington, USA

Post by mwhiz »

@orion: yeah yeah. i don't see what can be done about it though... except make the team aware, like you've done. perhaps a more rigorous test? i'm not sure. hopefully they can just eliminate "the noise" statistically.
"The Earth is the cradle of mankind, but one cannot live in the cradle forever."
~Konstantin Tsiolkovsky

Orion_0169
Posts: 153
Joined: Wed Aug 02, 2006 8:17 pm

Post by Orion_0169 »

A start would be the feedback for people and maybe that is best achieved by suspending operations for us for a while. On time. Off time. Sort of thing. Time where they aren't getting peppered by our enthusiasm, and a chance to address the many issues cropping up.

I'm REALLY hoping that the movies will be coming in batches. So for example, soon, these 40 000 will be taken completely off the books and replaced by another set. That way each batch gets even odds at multiple looks rather than having this early stuff lasting indefinitely.

Plus, as has been mentioned in other places, taking piles of THESE real movies and turning them into callibration movies as well.

Problem is, the sheer volume of movies that some of us are going through makes a large sample size of... well, samples necessary.

I try not to pay attention to some of the more obvious callibration movies (partly because some of the features that people keep saying are so obvious DO show up in real movies) but mostly because I'm making a VERY concerted and conscious effort not to become lazy. Easy to do right now.

And for the scoring, if the movies were to be released in batches, it would be neat to see scores reset for every batch. It would be a lot more encouraging for those people taht find themselves "way behind" Running totals could be kept on each leg for those people who've done so much work, but... ah... just musing...
Je ne peux pas regarder la Mer sans me demander qui vit au-delà de cela.

mwhiz
Posts: 95
Joined: Tue May 23, 2006 3:58 pm
Location: Seattle, Washington, USA

Post by mwhiz »

hmm... interesting. batches. could be fun. my guess is that movies will just be slid out of circulation as they get to 100 views and new ones slipped in as they are imaged. there might be quite a loss of momentum if there are breaks in the searching (of course, there might be too much momentum right now :shock: )
"The Earth is the cradle of mankind, but one cannot live in the cradle forever."
~Konstantin Tsiolkovsky

Betelgeuze
Posts: 19
Joined: Sun Aug 06, 2006 1:24 am
Location: Belgium

Post by Betelgeuze »

Yeah the stats dont make a lot of sence, some things that are nothing get a lot of marks, and this one for example only got only 9 marks after 53 viewings while I clearly see a track:
http://stardustathome.ssl.berkeley.edu/ ... e_id=47394

mwhiz
Posts: 95
Joined: Tue May 23, 2006 3:58 pm
Location: Seattle, Washington, USA

Post by mwhiz »

@betelgeuze: :shock: hard to miss that one. golly.
"The Earth is the cradle of mankind, but one cannot live in the cradle forever."
~Konstantin Tsiolkovsky

Post Reply