| |
booker
Registered: Jul 2003
Posts: 334 |
[idea] a bit of fun for ratings - real Charts
Remember in the old days when there were votesheets and we were voting for _active_ groups/ppl and sometimes the first page to open in new mag was charts in a hope to see own nick in there in top 10? :):)
why not resurrect this?
Let's say the scoreboard to get divded into all-time rating and active-rating. With some ruling, for instance - no new prod in past 12 months you out of "active scoreboard".
This can still pump up the scene as it is now, same as it was pumping it 20 years ago.
What say you? :) |
|
... 15 posts hidden. Click here to view all posts....
|
| |
booker
Registered: Jul 2003
Posts: 334 |
..but :) if you would allow for an externall query of the ratings (ideally via API if you have it) I could try to get this done (or get a help) and hosted myself.
How about that? |
| |
Celtic
Administrator
Registered: Jan 2002
Posts: 807 |
ehmmm, not a coder here, so i leave this a bit, and hope perff will take a look and see if this is possible for him and what this entails. Be patient, this may take a while |
| |
booker
Registered: Jul 2003
Posts: 334 |
Cool okay no prob, thanks for checking up with me. |
| |
Axis/Oxyron
Account closed
Registered: Apr 2007
Posts: 91 |
I also dont think that CSDB is the place to do this, because its a database.
But to all the diskmag editors reading this: WE WANT DA CHARTS!!! |
| |
Mixer
Registered: Apr 2008
Posts: 449 |
I usually avoid rating discussions as they're pointless.(pun intended)
All statistical results are lies. In reality, there is no such thing as an average and even if there was some useful average measure, it is bad decision making to base judgement on that, especially if samples are people.
For instance, assume that there are two men, one that is 165 cm and one that is 185 cm. If you go and buy a shirt to both based on average height, you'll have a shirt for 175cm man and there is no 175 cm man. The shirt will be too big for one, and too small for the other and neither are happy.
Some think that votes of others are somehow wrong or purposedly manipulated up or down. That is just one consequence of the one dimensional average problem. Some believe their opinion don't matter and others think theirs should matter more, some ignore and some abuse the system.
In this light, Bookers suggestion of external charts is step into better direction. I hope it'll be taken far enough and produces more than one dimensional view of all items. Perhaps allow everyone their charts and own scales of awesomeness. |
| |
booker
Registered: Jul 2003
Posts: 334 |
Mixer, I'd be glad if you can elaborate your vision about multidimensional charts vision you mentioned. Assuming I'd get approval to use CSDb votes we will see the charts done outsourced. |
| |
Joe
Registered: Apr 2002
Posts: 229 |
I have to reply, since I still owe the top position in voting here, anonymous of course, (making it 19002 at the moment)
as it always should be, regardless real-life parliament votes or hobbyists making love to their machines.
How could such systems be outsourced? How are they helped in another form outside and most important,
why dont more take the opportunity of giving their honest say, without commenting a remark on what they see/hear/feel?
If it works better offline or in another forum and perhaps with a different viewpoint part from what has been tried in several magazines in the scene throughout the years.
Go ahead! |
| |
booker
Registered: Jul 2003
Posts: 334 |
Hi Joe. Glad you joined the discussion.
I'll just host "outsource" :) the code on a rented server. Again it would simply use existing votes from CSDb. Idea is you don't have to fill in any new votesheets, because this doesn't work and mags currenlty dropped the charts off. You can vote anonymously still as you do arleady AND no questions asked.
I'm trying very hard to emphasize this is not another thread to slash anyones decision on the voting. Vote as you please, that's the point, however the chart calculations code with the aim to reflect CROWD rating (majority) not INDIVIDUAL (or minority) of votes. And as I showed earlier on few occasions minority of ratings are disturbance to the viewpoint of crowd rating, be it 300 or 20 votes count total.
I've done it in this thread http://csdb.dk/forums/?roomid=5&topicid=106366#106513
Now, this all comes from the fact Perff or whoever else is doing the CSDb code rating and scoreboard calcs have not time/will/whatever_applicable to shake the ratings a bit, and also as Celtic said they don't see CSDb to be a place for charts. This is ok.
So I just simply though instead of harrasing admins/creators with a request to do something, I'd to it myself or if I'll feel I will fail in coding I have some friends who can do web coding, so I'll ask them for help.
The calculation would be then a bit diffrerent. The main idea is to cut out the minority of votes to show up the rating driven by the majority of votes. Currently the scoreboard is fucked up by 5% (or less) of votes I was able to move a 300+ votecount entry on a scoreboard (EOD demo) a position down(!) by entering just my single vote of a score of 1. I've removed it now, but seems someone followed the idea. If you look at it now there are two votes of 1 for EOD and they consits 0.06% of all votes, therefore my vote (at the time the only vote of value 1) mustvebeen somewhere 0.03% or something? Yet it moved the production on the scoreboard from position 1st ot 2nd :)
This is example for a big voting for EOD. For small voting on 19 or so votes on example musican I showed the minority of 5% of votes (a single vote) been able to control the scener scoreboad rank by 40 positions.
For the cuttof there hast to be some margin, which would need to be normalized. I'm thinking the margin to be proportional to the amount of votes. Ie 300 votes 0.05%, 20 votes 5% and so on - It's hard to say preciselly yet because I'd need to source the voting data from CSDb yet, put the code in and see the results. And while it's possible to just rip it out of the CSDb web I don't want to do that, and wanted to get the permissions to fetch the data first.
Secondly the real charts idea - this is simply to show ACTIVE scene movement as it is now. For instance you have Jeroen Tel & Hubbard top musicians. This is ok, but shows all time. These folks are hardly active, rating them on the scene charts is out of old good standards of old scene 20 years and more in past - you simply DON'T do charts on an inactive sceners :) If you want to be on the charts you need to 1) do something 2) deserve the high votes. Otherwise you get a kick out of the table, and you come back to it when you 1) do something 2) deserve high votes from seners. Simple old good rules.
So in short these things are moving in and out my head recently... totally open to suggestions (apart from those who'd say me to fuck off :) not interested :)) ) |
| |
Mixer
Registered: Apr 2008
Posts: 449 |
The multidimensionality simply means that an entry can be voted on more than one scale. For instance graphics, soundtrack, code etc.
To get more points of view, create a scale/chart for everyone. "this demo gets 7 on the deathmetal scale of Mixer".
Regarding time dimension, charts often take account the time between publications, but here per year precision is my guess for a working timescale. It means that votes should have timestamp. |
| |
Mixer
Registered: Apr 2008
Posts: 449 |
Addendum, I do not wish to exclude anyone that has voted from the voting, old or new active or inactive. That would be putting power into some hands instead of others. The timescale is my preferred method for making the continuum of popularity discrete enough for a chart. |
| Previous - 1 | 2 | 3 - Next |
