review of wrong answer
Moderators: Stardust@home Team, DustMods
-
- Posts: 1
- Joined: Tue Aug 01, 2006 4:25 am
- Location: Italy
review of wrong answer
Hi all
it would be very useful a review of the wrong answer movies, with arrows to show missed particles .... as some of them are very odd and/or little
it would be very useful a review of the wrong answer movies, with arrows to show missed particles .... as some of them are very odd and/or little
Nice timing! I just broke off ... got 2 in the last while.
Surely staff must recognize the statistical value of tracking which calibration movies folk are getting wrong. I realize this means adding to the database ... but we're reminded repeatedly that this is s'posed to be fun and right now I'm second guessing myself ... the creepy feeling that I'm making the same mistake again and again.
Addendum: I think I've located my systemic error. I gave my most recent movie a triple-close look and I spotted a teeny track ... 5 microns? less?
Distinct, and correct, and small!
Surely staff must recognize the statistical value of tracking which calibration movies folk are getting wrong. I realize this means adding to the database ... but we're reminded repeatedly that this is s'posed to be fun and right now I'm second guessing myself ... the creepy feeling that I'm making the same mistake again and again.
Addendum: I think I've located my systemic error. I gave my most recent movie a triple-close look and I spotted a teeny track ... 5 microns? less?
Distinct, and correct, and small!
When you look to see how the system works
Likely you will find that it does not.
Likely you will find that it does not.
Wow ... I just missed 10839A ... classic ... teeenie!BenTrem wrote:Addendum: I think I've located my systemic error. I gave my most recent movie a triple-close look and I spotted a teeny track ... 5 microns? less?
Distinct, and correct, and small!
If there wasn't a little track like this in Training I nominate this one.
When you look to see how the system works
Likely you will find that it does not.
Likely you will find that it does not.
-
- Posts: 2
- Joined: Thu Aug 03, 2006 2:23 pm
Ditto
http://stardustathome.ssl.berkeley.edu/ ... rds=?-2,-2
Movie ID = 10516A
Please fix this it's very frustrating!
Movie ID = 10516A
Please fix this it's very frustrating!
This one gives wrong answer. Theres a track but once you mark it you get incorrect answer!
Movie: 10494A
http://stardustathome.ssl.berkeley.edu/ ... rds=?-1,-1
Movie: 10494A
http://stardustathome.ssl.berkeley.edu/ ... rds=?-1,-1
-
- Posts: 9
- Joined: Fri Aug 04, 2006 1:05 am
I would assume that the stats for cal. movies are collected in just the same way as real movies. The difference is that these are secret - if we know the answers to the cal movies, then the cal movies become useless as a mechanism for testing user accuracy.
At the moment, there is not enough data to draw any real conclusions. My best real track has 6/33 agreements. If it was a cal, and had 27/33 agreements, is that really significant?
Sean
At the moment, there is not enough data to draw any real conclusions. My best real track has 6/33 agreements. If it was a cal, and had 27/33 agreements, is that really significant?
Sean
-
- Posts: 19
- Joined: Sun Aug 06, 2006 1:24 am
- Location: Belgium