Learning curve, and addiction

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

Moderators: Stardust@home Team, DustMods

Post Reply
blewis
Posts: 10
Joined: Wed Aug 16, 2006 11:01 am
Location: Evanston, Illinois
Contact:

Learning curve, and addiction

Post by blewis »

During the first few days of searching, my learning curve has been steep, in terms of my sensitivity values. Now, however, it seems to be flattening out, i.e., hovering around 77% (specificity is about 98%). I think I am missing the tiny ones, perhaps below the limits of my 72 year old bifocal vision. Using a magnifying glass does not help, since (according to my learned husband) its all in the pixels. Will it help to keep going? Is one's visual acuity a factor in identifying small tracks? By the way, I am becoming addicted to the search, and neglecting much of my other commitments. Perhaps a warning regarding addiction should be included in the description of the activity!
Pooh
jdrockster
Posts: 27
Joined: Wed Aug 16, 2006 11:14 am
Location: Northern California

addiction warning

Post by jdrockster »

:shock: Yes, an addiction warning should be issued! I too have some things that have been neglected since I started this project. It is fun though. I have been staying up much later to try to maintain a ranking around 330. It is difficult to gain on some of these "addicted" searchers. Good luck!
Howie
Posts: 331
Joined: Wed Jun 21, 2006 2:47 pm
Location: Florida

Addicted ??? Some Tips !!!

Post by Howie »

I no longer care about my score, I am doing the best I can and those guys up top may not even find the track that gets your name on it!!!

You could, so take your time!!!

Note ...

If you have the latest Winows 98x or XP and IE and all updated,,,
you may be able to use the zoom a page larger for viewing like I do!!!

Note ... I am using updated BETA 3, for XP, IE, and all the stuff that is included!!!

Hope this Helps,
Howie
benn_tom
Posts: 1
Joined: Wed Aug 16, 2006 5:34 pm
Location: North Wales, U.K.

Post by benn_tom »



Well i must (at the time of writing this anyhow) b the newest person to register, but i am hooked already

the more i think about the issues around this experiment the more interesting and exciting it is

just hope i can make my way to bed tonight! :)
StanPope
Posts: 7
Joined: Sat Aug 19, 2006 9:20 am
Location: Central Illinois
Contact:

Re: Learning curve, and addiction

Post by StanPope »

blewis wrote:During the first few days of searching, my learning curve has been steep, in terms of my sensitivity values. Now, however, it seems to be flattening out, i.e., hovering around 77% (specificity is about 98%). I think I am missing the tiny ones, perhaps below the limits of my 72 year old bifocal vision. Using a magnifying glass does not help, since (according to my learned husband) its all in the pixels. Will it help to keep going? Is one's visual acuity a factor in identifying small tracks?
I found similar problems early on ... even though I was sure that I was searching each movie carefully, I was missing enough of the test questions to keep sensitivity below 90%. It was frustrating because it appeared that I was making some systemmatic error, but unable to identify the cause. In retrospect, had the "test" movies answered wrong sent me to a focused training set ("This was the movie, this was the track. You missed it!" Or "This is a track like the one that you missed in the test!"), I'd have resolved the problem in hours rather than days.

I did several things to get on track:
1. Moved searching from an old CRT display with aging focus and brightness to a recently minted flat-screen. Still missed a few test movies.
2. Repeated the training movies on the new screen. Found that very small tracks were much easier to see on the good screen. I didn't remember that movie when I took the training sequence the first time (and scored 100% on the test).
3. To accelerate verification of the correction, I watched my proxy server's connections screen during a few dozen "test" movies, so that I would have a better idea of what I was doing wrong and what kind of tracks I was missing. Yes, indeed it was the very small tracks ... some that almost appeared to my old eyes as having no hole in the middle of the donut.

Since completing that process, I don't need to track the movie type and have reduced my current error rate to below 1%. Also, I apply the same review process to each movie, real or test, with much more confidence.

I suspect that I can go back to the old monitor now and maintain that low error rate ... but I have not proven that to myself yet.

Search on!

Stan
Post Reply