Ratios

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

Moderators: Stardust@home Team, DustMods

Post Reply
PocketLint
Posts: 20
Joined: Mon Aug 07, 2006 1:02 pm

Ratios

Post by PocketLint »

I like that you have expaned the ratios to 2 decimal places to fix the rounding issue but..

It seems the same calculation is being used for both so each has the exact same ratio..
I missed a calibration movie that had a track and both ratios dropped by .5 %

I have answered 3354 calibration movies correct
I screwed up on 18
I have viewed 9365 real movies.

My ratios are:
Specificity: 99.05%
Sensitivity: 99.05%

PocketLint
Posts: 20
Joined: Mon Aug 07, 2006 1:02 pm

never mind

Post by PocketLint »

I see that is now fixed..

Thanks!!

jcbc
Posts: 65
Joined: Wed Aug 02, 2006 2:10 pm
Location: Pennsylvania, USA

Post by jcbc »

I don't see any decimal places... Maybe I haven't done enough to earn them. :?
My Events wrote:Statistics for jcbc

Your Overall Score: 255
Total Movies Viewed: 687
Your Rank: 1326 out of 10001
Specificity: 99%
Sensitivity: 99%
Virtual Microscope wrote:Calibration Movies Answered Correctly 257
Calibration Movies Answered Incorrectly 2
Your Overall Score: 255
Total Real Movies Viewed: 428
Your Rank: 1326 out of 10001
Specificity: 99%
Sensitivity: 99%

stevems
Posts: 14
Joined: Thu Aug 03, 2006 8:18 pm

Post by stevems »

jcbc wrote:I don't see any decimal places... Maybe I haven't done enough to earn them. :?
I'm in the same boat. No decimal places, and similar scores to yours.
VM wrote:Calibration Movies Answered Correctly 301
Calibration Movies Answered Incorrectly 2
Your Overall Score: 299
Total Real Movies Viewed: 739
Your Rank: 1171 out of 10018
Specificity: 99%
Sensitivity: 99%
Maybe we're not making enough errors. :wink:
Steve
---------------------------------------
Sometimes I wonder why I spend
The lonely nights
Dreaming of a song.

DaveJ
Posts: 13
Joined: Thu May 18, 2006 1:52 pm
Location: Joliet, Illinois
Contact:

Post by DaveJ »

The percentage I'm getting doesn't seem to be correct:

Calibration Movies Answered Correctly 511
Calibration Movies Answered Incorrectly 1
Your Overall Score: 510
Total Real Movies Viewed: 1386
Your Rank: 581 out of 10018
Specificity: 100.00%
Sensitivity: 99.64%


510/511 = 99.80%, not 99.64%. 99.64%=509 movies answered correctly, but I have 510 answered correctly.

Now, I may be nitpicking, but incorrect is incorrect.

ID is DJ257.

PocketLint
Posts: 20
Joined: Mon Aug 07, 2006 1:02 pm

strange...

Post by PocketLint »

I wonder if it does have to do with the number views.
It could be though that I have had to re-boot my PC a few times as I'm on a wireless lan and the drivers for this POS linksys card freeze up often.

Maybe I got the newest html code and you are still caching then old?

PocketLint
Posts: 20
Joined: Mon Aug 07, 2006 1:02 pm

ESP

Post by PocketLint »

They must have installed that latest ESP class script that just knew we would screw up at least one more movie and was posting our future results

jcbc
Posts: 65
Joined: Wed Aug 02, 2006 2:10 pm
Location: Pennsylvania, USA

Post by jcbc »

DaveJ wrote:510/511 = 99.80%, not 99.64%. 99.64%=509 movies answered correctly, but I have 510 answered correctly.
The calibration movies are split into "no track found" (specificity) and "track found" (sensitivity). Since you've only missed one, and it's showing up in sensitivity, then this means that for "track found" calibration movies, you have had something like 278 sensitivity calibration movies and missed one (277/278 = 99.64%) and you have gotten the 233 remaining specificity calibration movies right (233/233 = 100%).

JCBC

jcbc
Posts: 65
Joined: Wed Aug 02, 2006 2:10 pm
Location: Pennsylvania, USA

Re: strange...

Post by jcbc »

PocketLint wrote:Maybe I got the newest html code and you are still caching then old?
Hmm... that's very possible. I'll turn off my computer tonight.

JCBC

stevems
Posts: 14
Joined: Thu Aug 03, 2006 8:18 pm

Post by stevems »

DaveJ wrote:
Specificity: 100.00%
Sensitivity: 99.64%

510/511 = 99.80%, not 99.64%. 99.64%=509 movies answered correctly, but I have 510 answered correctly.

Now, I may be nitpicking, but incorrect is incorrect.
I think that they calculate specificity based the number of times we say we've found a track in "blank" calibration movies. Sensitivity would be based the number of times we've missed tracks in calibration movies that have tracks. I may have that backwards. :?

So, you've probably looked at about 275 calibration movies with tracks, and you missed 1 of them.

274/275 = 99.64%

The other 236 (or so) calibration moves were blanks (no tracks) and you got them all correct.

I think that's how it works.

Now, can anyone tell me why I don't have any decimal places in my percentages? :?: :?:
Steve
---------------------------------------
Sometimes I wonder why I spend
The lonely nights
Dreaming of a song.

stevems
Posts: 14
Joined: Thu Aug 03, 2006 8:18 pm

Post by stevems »

jcbc

You type faster than I do. :wink:

Or maybe I'm too long-winded.
Steve
---------------------------------------
Sometimes I wonder why I spend
The lonely nights
Dreaming of a song.

stevems
Posts: 14
Joined: Thu Aug 03, 2006 8:18 pm

Re: strange...

Post by stevems »

jcbc wrote:
PocketLint wrote:Maybe I got the newest html code and you are still caching then old?
Hmm... that's very possible. I'll turn off my computer tonight.
I was dusting at work today (during breaks only, of course) 8) and I had no decimal places.

I just booted up my computer at home, and still no decimal places.

I'm missing the point. :roll:
Steve
---------------------------------------
Sometimes I wonder why I spend
The lonely nights
Dreaming of a song.

DaveJ
Posts: 13
Joined: Thu May 18, 2006 1:52 pm
Location: Joliet, Illinois
Contact:

Post by DaveJ »

jcbc wrote:
DaveJ wrote:510/511 = 99.80%, not 99.64%. 99.64%=509 movies answered correctly, but I have 510 answered correctly.
The calibration movies are split into "no track found" (specificity) and "track found" (sensitivity). Since you've only missed one, and it's showing up in sensitivity, then this means that for "track found" calibration movies, you have had something like 278 sensitivity calibration movies and missed one (277/278 = 99.64%) and you have gotten the 233 remaining specificity calibration movies right (233/233 = 100%).

JCBC
Ahhh, that makes sense! Thanks to both you and stevems.

Howie
Posts: 331
Joined: Wed Jun 21, 2006 2:47 pm
Location: Florida

What about bad focus movies???

Post by Howie »

No one mentioned bad focus???

Where does bad focus come into play, other than a viewed movie in total movies watched...it is not added to real movies or calibration, so???

Ziggy
Posts: 10
Joined: Wed Aug 02, 2006 4:39 pm

Re: What about bad focus movies???

Post by Ziggy »

Howie wrote:No one mentioned bad focus???

Where does bad focus come into play, other than a viewed movie in total movies watched...it is not added to real movies or calibration, so???
I don't think the bad focus movies are considered anywhere. Your total movies watched figure should be:
Total Movies = Real Movies + Calibration (Correct) + Calibration (Incorrect)
I estimate that for about evey 4 real movies I see, then I will judge 1 to be bad focus (as the focus range is insufficient to conclusively deduce that a track exists) and it would be nice to have these viewings be represented within the statistics on the site.
Ziggy Stardust...

...searching for the Spiders from Mars.

Post Reply