A.D.A. Amiga Demoscene Archive

  Welcome guest! Please register a new account or log in

  

  

  

Demos Amiga Demoscene Archive Forum / A.D.A / ADA chart

 

Author Message
scicco
Member
#1 - Posted: 8 Apr 2008 14:22
Reply Quote
just because z5 asked about to post a calculation method for the top rankings.

if you want votes comparable, you need to compare the number of voters to the number of registered user who may vote and then apply this on the average vote result.
eg: 100 user at ADA, 15 voted, avg: 3.5/5
-> 100/15*3.5

this result in overall 100% for a prod which all voters gave 5/5.

so i don't think that the chart calc math is a problem, but the number of possible voters may. as written in a different thread, there might be some accounts which are just created to fake votes.

a side effect: prods having 10 5/5s are ranking lower than prods having 20 3/3s.
it may help to use
100/15*3.5^2
instead.

ah, well, just suggestions, not real checked. SQL-it a try and you'll see. :)

/scicco
scicco
Member
#2 - Posted: 8 Apr 2008 14:27
Reply Quote
[..]prods having 20 3/3s.[..]

should have been

prods having 20 3/5s of course. :)
dalton
Member
#3 - Posted: 8 Apr 2008 18:33
Reply Quote
check out the formula they use at imdb: http://en.wikipedia.org/wiki/Imdb#User_ratings_of_ films
scicco
Member
#4 - Posted: 8 Apr 2008 18:42
Reply Quote
this looks pretty good!
go, z5 :)
rload
Member
#5 - Posted: 8 Apr 2008 21:15
Reply Quote
"Only non-documentary theatrical releases running at least forty-five minutes with over 1300 ratings are considered;"

damn.. I shouldn't have made Psylteflesk a documentary...
z5_
Member
#6 - Posted: 8 Apr 2008 21:23
Reply Quote
damn.. I shouldn't have made Psylteflesk a documentary...

don't worry, rload. All Ephidrena releases are already disqualified because none are forty-five minutes. Those 1300 ratings on the other hand, that it something ada reaches with ease... (me is quickly inventing 1260 nicks to post fake votes...)
movew
Member
#7 - Posted: 8 Apr 2008 21:44
Reply Quote
haha cool...

...running at least forty-five minutes...

Maybe the proposed system could be applied to Odyssey by Alcatraz.
klipper
Member
#8 - Posted: 11 Apr 2008 22:52
Reply Quote
z5: wow, the charts suddenly look very different! :)
z5_
Member
#9 - Posted: 12 Apr 2008 00:15
Reply Quote
@klipper: i assume this is tongue in cheek for my laziness, right? :)
d0DgE
Member
#10 - Posted: 12 Apr 2008 09:50
Reply Quote
DARN ...movew was faster xD
z5_
Member
#11 - Posted: 27 Jun 2008 17:24 - Edited
Reply Quote
I did a little test with something i found on the web using bayesian averaging:
WR = (v/v+m)*R + (m/v+m)*C
WR = weighted average
v = number of votes
m = minimum number of votes required
R = average for this production
C = average of all productions (who have at least minimum number of votes)

You can see the result here. It's still the same chart but the weighted average is mentioned at the right side of the normal average (which still determines the chart position).

The problem with such a system is that C is a variable which depends on the votes. So whenever a vote is added or changed, each and every weighted average has to be recalculated.

It was a fast hack so i hope i didn't mess up. In any case, it was just of curiosity and not meant as a permanent thingie.
Frog
Member
#12 - Posted: 28 Jun 2008 20:48
Reply Quote
i think it could be interesting to add the same filter as in the Demo page so you can get for example TOP 10 demo AGA all year, TOP10 intro64k ECS 1922...

 

  Please register a new account or log in to comment

  

  

  

 

A.D.A. Amiga Demoscene Archive, Version 3.0