Possible calibration movie error

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

Moderators: Stardust@home Team, DustMods

Locked
smandla
Posts: 6
Joined: Wed Aug 02, 2006 1:26 pm

Possible calibration movie error

Post by smandla »

The following ID number corresponds to a calibration movie: 10516A. I believe it has a track in it, yet the system said it does not. Can someone verify this and make sure it isn't an error? Thanks.

mstetson
Posts: 4
Joined: Fri Aug 04, 2006 12:33 pm

Post by mstetson »

Absolutely! In fact I just got dinged on this one. There's an obvious track just to the right of centerline about a third of the way down from the top. I clicked on it and was "rewarded" with an "incorrect" score.

-Mike

smandla
Posts: 6
Joined: Wed Aug 02, 2006 1:26 pm

calibration movie error

Post by smandla »

It sure stings when that happens, doesn't it? Well, I'm glad to know it's not me. By the way, I did some digging and found another reference to the same movie and the same problem from "AK47".
We are not alone in our suffering...
;-)

Jeff
Posts: 2
Joined: Fri Aug 04, 2006 11:55 am

Post by Jeff »

Just registered on the forums because of this one. ;)

Shamgar
Posts: 2
Joined: Fri Aug 04, 2006 3:45 pm

Post by Shamgar »

I just had the same problem but with another Calabration movie: 11629B
There's totally a /classic/ example of a trail on the left side, just like the ones from the tutorials, but it dinged me anyways. :!:

Tis rather frustrating...

smandla
Posts: 6
Joined: Wed Aug 02, 2006 1:26 pm

10516A rears its ugly head... AGAIN!

Post by smandla »

I can't believe this happened again... with the same calibration movie!! I think we should start a hit list of all the incorrectly classified calibration movies - i.e. movies that have a track but are scored as not having one and vice versa. I've noticed a lot of postings describing this problem.
I'm going to post this message with "the powers that be" in admin, in the hopes that this problem can be corrected.

ecohen
Posts: 2
Joined: Tue Aug 01, 2006 3:33 pm

YABCM (Yet Another Bad Calibration Movie)

Post by ecohen »

This is frustrating. I got a Calibration movie wrong, but I think that the movie that I saw was almost certainly a track. Here's a screenshot
Image

Here's the movie that is called up with the viewer.

http://stardustathome.ssl.berkeley.edu/ ... e_id=10635

They are not the same. I do not know what the movie is that I saw, but the whole movie was consistent. Not like some where the movie obviously shifts midway down, this one was cohesive.

In another thread they were calling for a list of failed calibration movie efforts, so that theycan be reviewed and learned from or challenged.

DiamondGirl
Posts: 77
Joined: Tue Jun 13, 2006 12:16 pm
Location: Syracuse, NY

Post by DiamondGirl »

I'm so glad to find out it's not just me! I went back and checked in dismay (okay, it's silly, but I was SO proud of my perfect score!) and there is a definite track, but I still got dinged. <<sigh>> Any word on whether we'll be able to have them reviewed? Get our standings fixed?

Also, if I divide my total number of correct calibrations by the total number given, it gives a higher percentage than what my score is reflecting. ??? Anyone else having this problem?

JosephR
Posts: 1
Joined: Sat Aug 05, 2006 12:44 am

Post by JosephR »

Same here: there was clearly a track in 11309B, so I clicked on it... and my perfect score was gone. After a few minutes it happened again, with 11401B. This is too much frustrating (and, above all, will affect how they will consider the candidates I identified); I am going to quit, I think.

Roswell_Crash
Posts: 1
Joined: Sat Aug 05, 2006 2:41 am

Calibration test errors

Post by Roswell_Crash »

I took the test twice. Several times in each test I was told my answer was incorrect when I click on a movie that had a track. I was puzzled about this until I read these forum entries. Now I feel better about flunking the tests. When will this be fixed so I can take the test successfully> ANd, how did anyone pass the test with so many errors in the tests??

DustBuster
DustMod
Posts: 694
Joined: Wed May 17, 2006 8:12 pm
Location: Horsetown, USA

Post by DustBuster »

This issue is being addressed. Here is the official thread on this subject http://stardustathome.ssl.berkeley.edu/ ... .php?t=473
No dessert for you- ONE MONTH!

splait
Posts: 2
Joined: Thu Aug 03, 2006 5:38 am

Me, three

Post by splait »

I believe I have been dinged at least seven times but I wasn't smart enough to register the movie number until this morning. (Actually, I was too competitive to go back and register the number.) 11217A. I said there was NO track, but was marked incorrect. I can find no track there.

Is there a way to go back and look at the ones I was "incorrect" on.? Is there a list somewhere of the ones I missed?

Several times I clicked on clear tracks in calibration movies and was marked as incorrect.

Arrgh!

Nanotrix
Posts: 1
Joined: Tue Aug 15, 2006 11:55 am
Location: Germany

importance of score/specifity/sensitivity

Post by Nanotrix »

So far i got 30 wrong calibration movie results - CalMovies with clearly visible tracks but marked as "wrong" when clicking. This decreases my specifity seriously.

But: is this bug of any importance? Will all the tracks we found in real movies be rejected at the end of the project because our specifity falls? I dont think so - the bug will bug all the members, so it will average out itself.

---
Be cool! Lets make science, not scores...

Stadt
Posts: 11
Joined: Fri Aug 04, 2006 6:26 am
Location: Netherlands

Post by Stadt »

But: is this bug of any importance? Will all the tracks we found in real movies be rejected at the end of the project because our specifity falls? I dont think so - the bug will bug all the members, so it will average out itself.
I do agree, however, it's does deflate my ego somewhat, and you don't know instantly that you were right and it was just an wrong calibration movie - it makes me doubt myself when it happens. I think the real bug is gone now, there might just be a few movies that were maybe classyfied wrong as having no tracks.

I think 2062521V1 is such a case. I thought I saw a possible inclusion on 220,20. I might be wrong though - I'd rather know in order to prevent making unnecessary false positives.
Image
Cheers!

albutterworth
Stardust@home Team
Stardust@home Team
Posts: 69
Joined: Wed May 17, 2006 12:24 pm
Location: UC Berkeley, CA
Contact:

Post by albutterworth »

In an effort to keep these FAQs organised, discussion of the calibration movies belongs in
I answered a calibration movie right but was scored wrong.

Thanks
Stardust@home
Useful Links: | Aerogel Scanning Progress | Forum Rules |

Locked