Challenges historical stats - DPChallenge Forums

Canon EF-S 15-85mm f/3.5-5.6 IS USM

02/07/2007 05:44:40 AM · #1

I was having a 5min break at work and was wondering how the challenge stats had evolved in 5 years of DPC.

Here a quick graph showing:
* That the average score is absolutely flat (at 5.2) over time: no increase or decrease in overall challenge quality (viloet, right scale).

* That the average number of comments you get in a given challenge depends very much on the challenge (light blue, right scale) but that all in all it's pretty stable over time (dark blue is the 1 year trailling average) at 10-15 comments on average per picture.

* That the average number of entries in a given challenge depends also very much on the challenge (light green, left scale), as one would expect, but that all in all it's now stabilizing (dark green) at 200-250 entries on avg per challenge.

ok I'll get back to work now...

Message edited by author 2007-02-07 05:51:02.

LN13

Canon EOS-60D

02/07/2007 01:58:52 PM · #2

Your graph looks like someone is lying really badly while attached to a polygraph or some really severe earthquakes.

asimchoudhri

Canon EOS-5D Mark III

Canon EF 17-40mm f/4.0L USM

02/07/2007 02:11:27 PM · #3

edit: never mind orig post...

great work! this is really cool to see... the scale for the score doesn't really allow us to see the variations, however...

Message edited by author 2007-02-07 14:13:30.

Canon EF 17-40mm f/4.0L USM

02/07/2007 02:37:12 PM · #4

Originally posted by asimchoudhri:

the scale for the score doesn't really allow us to see the variations, however...

agree but hard to have 3 different scales in 1 graph :D, and even with more zoom, variations remain limited

asimchoudhri

Canon EOS-5D Mark III

02/07/2007 02:54:45 PM · #5

Originally posted by mouten:

Originally posted by asimchoudhri:
the scale for the score doesn't really allow us to see the variations, however...

agree but hard to have 3 different scales in 1 graph :D, and even with more zoom, variations remain limited

any way to make the right scale 0-20? (or maybe 0-22, to include those early comment-heavy challenges)

edit: also, out of curiosity, what's the width of your moving average window?
(i'm not trying to be nit-picky... actually, i'm very excited and impressed, and that's why i have a bunch of questions... this is great!)

Message edited by author 2007-02-07 14:56:34.

02/07/2007 04:04:50 PM · #6

i'll post the answer to your questions when i'm back at work tomorrow ;)

Canon EF 17-40mm f/4.0L USM

02/08/2007 03:13:31 AM · #7

Originally posted by asimchoudhri:

also, out of curiosity, what's the width of your moving average window?
(i'm not trying to be nit-picky... actually, i'm very excited and impressed, and that's why i have a bunch of questions... this is great!)

here is the graph historical score alone. Not sure I got you question about the moving average window though?

Message edited by author 2007-02-08 03:14:15.

asimchoudhri

Canon EOS-5D Mark III

02/08/2007 10:41:33 AM · #8

cool... thanks for the post. you are right... not much variation at all.

with regards to the moving average, i was wondering how many values were being averaged for every given data point. for instance, if the "moving average" of data point n was calculated with a window width of 5, it would calculate the smoothed out value of data point n by averaging 5 values (either n-2, n-1, n, n+1, n+2 or n-4, n-3, n-2, n-1, n). just curious. but, looking at the new graph that shows more detail, doesn't seem like that would matter a whole lot.

my reason for originally asking was because the moving average on the first graphs didn't start until one year in... clearly the window wasn't one year, because the variation in the moving average tracked the raw data points fairly well without having the fluctuations (thus it served its purpose well). anyways, that's all irrelevant... thank you for posting this graph!