How the top scorers get those scores.

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

Moderators: Stardust@home Team, DustMods

stephenyavorsky
Posts: 2
Joined: Thu Aug 03, 2006 10:41 am

Re: Callibration movie feedback

Post by stephenyavorsky »

Congratulations on your contribution!
I am one of those who gets the messages about having gotten to a page by an unusual route, which means the work just done doesn't get counted. My concern is that more science could be done more quickly, even by those with slower computers and connections, if multiple tabs (as in Mozilla) could be used. The whole enterprise would benefit. It is frustrating to wait for a new movie to load when one could already be examining one in another tab. The work is easy enough, but waiting for it to arrive isn't.

= = = = =
Doug2a wrote:
Orion_0169 wrote:As someone who is in the top five, I should like to think that it isn't a matter of cheating. Ignoring the typically anti-intellectualism of assuming people "better" are cheating...

I was wondering about high scores too. And tried having more than one window logged in. And it works. I have three at the moment.

Why?

Efficiency. Not trying to clamour for high ranking. I might as well be viewing a movie while another loads. Makes more sense. With the tremendous number of images to be checked. I'm here to help. I would hope that I AM doing a good job of it to boot. And hope my numbers of mistakes and such are well below average.

You have to wonder though, if you don't care about rankings, why be concerned about how people have high ones?

On a more productive note, it would be nice if, when we got a callibration movie wrong, that we had a way of learning from the error.

I agree that feed back on misses on callibration movies would be helpful. We learn from our mistakes but only if we understand the mistake.

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

Post by Orion_0169 »

I was doing four (not today! won't let me) because the rhythm worked well as I was cycling through the four windows.

Just had three insanely small callibration tracks. I can see these ones being missed by lots of people and dragging their score down. I'm beginning to wonder if the system is set up to make you better. It should be!

Also, I wonder if the people running the show have thought to do rotations of images. So that positive hits come from different locations on the screen.
Je ne peux pas regarder la Mer sans me demander qui vit au-delà de cela.

Merovingian
Posts: 22
Joined: Wed Aug 02, 2006 11:33 am
Location: Juneau, Alaska, USA
Contact:

Post by Merovingian »

Orion_0169 wrote:Just had three insanely small callibration tracks. I can see these ones being missed by lots of people and dragging their score down. I'm beginning to wonder if the system is set up to make you better. It should be!
I think this makes sense. :) I seem to have noticed a general decrease in the frequency of large tracks in calibration movies.
In the dark of night, there lives the lighthouse keeper;
And the lone survivor of this world is he...

sibear
Posts: 3
Joined: Thu Aug 03, 2006 12:14 pm

Post by sibear »

Great info, using multiple windows.
One thing that I would have thought makes sense though us the way that possible finds are sorted. Ones made by highest scorers are checked first. However, being a high scorer doesn't necessarily mean that you are more accurate. Surely it would make sense to have a ratio system (correct movies : incorrect movies) for priority of checking possible hits.
Please note that I am in no way knocking anyone for putting in large amounts of work or being a high scorer. Every contribution is great.

barrett777
Posts: 5
Joined: Thu Aug 03, 2006 12:46 pm

What I did

Post by barrett777 »

I got up to 15th place on the rankings last night, I think I ended w/ 281 points, including 14 I missed. I'm the only person using my account, and I was going pretty fast. I'm on a cable connection, so the images loaded just about as fast as I could look at them.

When I finished, I was going probably 4-5 times as fast as I did when I started, and I also got more accurate.

I guess when you get more, you also get faster, if your internet can keep up.

Also, you can tell most of the trial movies from real ones, trial ones tend to be a little pixelated, if that's the right word :) I think the movies were made from different programs.

I'm sure it will be very hard to get up on the rankings in the coming days.
Since last night, I dropped to I think 94th place.
Last night, 100th place's score was 107 :)

Well, that's my story,
barrett777

sibear
Posts: 3
Joined: Thu Aug 03, 2006 12:14 pm

Post by sibear »

I think it's greta that so many people are putting in so much work.
It seems though that the servers are starting to show the strain. I'm on ADSL at 2meg and movies are taking ages to load at the mo. Earlier this evening they were instant :(.
Hopefully this is just a glitch.

agurney
Posts: 10
Joined: Thu Aug 03, 2006 1:07 pm

Post by agurney »

MaximVesuvius wrote: ...

Calibration Answered Correctly 501
Calibration Movies Answered Incorrectly 10
Your Overall Score: 473
Total Real Movies Viewed: 1394
Your Rank: 1 out of 3138
Specificity: 99%
Sensitivity: 90%
As I understand it, the only way you can achieve a score of 501 while only viewing 1394 real movies is if you skip the real movies and only click on the calibrations. It's easy done, but defeats the goals of the project.

Malc_UK
Posts: 3
Joined: Thu Aug 03, 2006 1:10 pm

Post by Malc_UK »

I was doing badly.. not seeing any tracks.. then I sneezed.. and hey presto! tracks all over :lol:

I just hope that never happens in the lab.... :shock:
We are Dyslexia of Borg. Futility is resistant. You will be ****-laminated

zuben el genubi
Posts: 19
Joined: Thu Aug 03, 2006 2:02 pm

Post by zuben el genubi »

Spend the last hours searching on my own,one window,one pc,with cable connection.
I'll need another 5,6 or 7 hours to catch #1,depending on how long this guy will go on.
Or I'll just have a couple of beers I guess :wink:


Overall Score: 660
Total Movies Viewed: 3054
Your Rank: 6 out of 5068
Specificity: 100%
Sensitivity: 87%

Calibration movies answered correctly 716
Calibration movies answered incorrectly 56

endust
Posts: 53
Joined: Wed Aug 02, 2006 9:42 am
Location: Kansas City

Post by endust »

Im a little concerned with everyone racing to hit the most movies. Several of us have found some anomolies that are very small specs that come into focus deeper in the gel. These may or may not be significant but I dont see how you could spot them going through the movies rapidly. If you can great, but when they say "expect the unexpected" they mean it and if you have a pre concieved idea of what your looking for you might miss something. I have no ranking to speak of but I have 100% on the calibration tests and to me thats more important than ranking. Good Luck all

nexguy
Posts: 38
Joined: Tue Aug 01, 2006 6:44 am
Location: Texas

Post by nexguy »

I agree. The more movies I do, the slower and more meticulous I seem to get. Now I spend at least 20 seconds on a movie that is not a test and has good focus.

I've viewed over 550 movies and have gotten all of my 135 calibration movies correct. I believe this is important because if I miss a few of those "easy" calibration tests, how could I count on my own observations of real movies?

zuben el genubi
Posts: 19
Joined: Thu Aug 03, 2006 2:02 pm

Post by zuben el genubi »

To me it has nothing to do about hitting the most movies,it's finding as much Stardust as possible.
And I just seem to spot them pretty fast.
I've seen over 2200 real movies and found what I believe are spots of Stardust on many of them.
I just go by my own tempo that's it.

MaximVesuvius
Posts: 20
Joined: Wed May 17, 2006 8:11 pm
Location: Boise, ID, USA, North America, Earth, Solar System, Milky Way

Post by MaximVesuvius »

agurney wrote:
MaximVesuvius wrote: ...

Calibration Answered Correctly 501
Calibration Movies Answered Incorrectly 10
Your Overall Score: 473
Total Real Movies Viewed: 1394
Your Rank: 1 out of 3138
Specificity: 99%
Sensitivity: 90%
As I understand it, the only way you can achieve a score of 501 while only viewing 1394 real movies is if you skip the real movies and only click on the calibrations. It's easy done, but defeats the goals of the project.
Then your understanding is quite pathetic. If I chose to skip all movies and just go for calibration, then why have 1394 movies? I should have had close to 2 grand by that time. Anyway, I take the movies as I get them; track, no track, or bad focus. Just take your complaint to the Admins, hell if I can release the data to what I did then I'd say take a look but that's up to them. Apparently the people who care so much about points spend time crying about it in the forums, just go back to work.

As far as I'm concerned I'm stepping back on this project until you have someone else to cry about.

agurney
Posts: 10
Joined: Thu Aug 03, 2006 1:07 pm

Post by agurney »

MaximVesuvius wrote:
agurney wrote:
MaximVesuvius wrote: ...

Calibration Answered Correctly 501
Calibration Movies Answered Incorrectly 10
Your Overall Score: 473
Total Real Movies Viewed: 1394
Your Rank: 1 out of 3138
Specificity: 99%
Sensitivity: 90%
As I understand it, the only way you can achieve a score of 501 while only viewing 1394 real movies is if you skip the real movies and only click on the calibrations. It's easy done, but defeats the goals of the project.
Then your understanding is quite pathetic.....
I agree, pass me some humble pie.

When I wrote that comment I was seeing something like 15-20% calibration movies, and assumed that would be the same across the board ... but recently I've been seeing more than 30% :(

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

Post by Orion_0169 »

Unfortunately, a little too late. Maxim's score has not budged since he said he was stepping away. That was precisely why I was being so adamant about the accusations. If he was someone just trying to do a good job, then he was bullied away. (I'm sure a bully would say something about him just sucking it up or something.)

It is unfortunate that this stuff starts in school and persists through life. We, as teachers, spend so much time talking the talk with students about bullying. But often those people don't walk the walk. I address these very things on a regular basis. And will likely use this example in my classes this year.

People need to spend more timing standing up for what's right. And supporting people with talent. Far too many kids shut themselves down because it just isn't cool to be good at things (unless it's hockey or football or...).

I've found myself easing up somewhat to reduce the likelihood of catching some flak. When I catch myself doing so, I remind myself not to allow others to drag me down to their level.

Kudos to you for saying what you just did. I hope he comes back. Projects like this need people with determination and dedication.

It is an admirable and hopeful thing to see so many people enthusiastic and interested in a scientific endeavour.

If you are still watching Maxim...

"Just keep swimming, just keep swimming."
Je ne peux pas regarder la Mer sans me demander qui vit au-delà de cela.

Post Reply