|
I like it (at least for high ranked articles).
What about low ranking articles with lots of votes? (guess I could pull out the calculator myself but call me lazy  )
(apologies for reformatting your post - I wanted others to see your idea more clearly)
cheers,
Chris Maunder
|
|
|
|
|
The last Top Ranked Articles normally have few votes (<10), so the "energy" is low. Anyway a "3/5" article needs 50 votes to overcome a "5/5" article with 1 vote.
|
|
|
|
|
I think the popularity ratings are biased. They tilts heavily toward MFC articles and toward "cool" subjects like grids. Meanwhile, other articles that are more specialized and receive less total views (and, consequently, less total votes) receive lower "popularity".
Frankly, I find it somewhat childish. It takes me back to grammar school days when we voted for the most popular boy and girl in class.
|
|
|
|
|
The whole point of the ratigs and popularity is two-fold
a) to allow those who have never been to CodeProject before a chance to look over the articles that most people found useful.
b) to have something that will give good articles a chance to appear on the homepage longer than would normally be the case.
No, it's not perfect, and I'm always up for other ways of accomplishing these two tasks.
cheers,
Chris Maunder
Rub your belly and pat your head simultaneously. Sometimes that helps me make sense of things - Jon Sagara
|
|
|
|
|
I agree that those two goals are important. You've got a tough job trying to please everybody.
|
|
|
|
|
Just realised that you've written Energy = rating * log10(max(10,#votes)) but your calculations are using Energy = rating * log10(#votes). I was originally wondering why the votes would be clipped at 10.
Your formula doesn't handle the case of an article rated 3 with 100 votes vs. an article rated 5 with 10 votes.
#votes rating popularity
Article A 10 5.00 5.00
Article B 100 3.00 6.00
What about something like: pop = (rating-3) * Log<sub>10</sub>(#votes). Then
#votes rating popularity
100 4.50 3.00
20 5.00 2.60
10 5.00 2.00
20 4.50 1.95
5 5.00 1.40
100 3.50 1.00
3 4.50 0.72
5 4.00 0.70
2 5.00 0.60
1 5.00 0.00
100 3.00 0.00
10 2.50 -0.50
100 1.90 -2.65
Unfortunately this breaks down if you are close to 3 but have a large number of votes.
cheers,
Chris Maunder
|
|
|
|
|
I made a quick search, and the 99.9% of the articles rated 3 or less have less than 40 votes. There are only FEW exceptions, like your article "Visual Studio.NET screen shots" with 2.93/5 and 67 votes (normal popularity : 5.35), all the other are more or less in the mud.
I choose "log10(max(10,#votes))" to avoid a handicap for the new entries and because log10 is <1 for values <10.
|
|
|
|
|
Seems like a good idea. 
"When a friend hurts us, we should write it down in the sand, where the winds of forgiveness get in charge of erasing it away, and when something great happens, we should engrave it in the stone of the memory of the heart, where no wind can erase it" Nish on life [methinks]
"It's The Soapbox; topics are optional" Shog 9
|
|
|
|
|
I think you have the right type of idea,
But it might need some tweaking.
Remember though these systems will always be controversial.
Regardz
Colin J Davies
Sonork ID 100.9197:Colin
More about me
|
|
|
|
|
the complete formula was:
rating * log10(max(10,#votes))*exp(-(#days since 1st post)/100))+100*(#clicked banners)+10*(#nish's threads);
but maybe it requires too much processing time for the server 
|
|
|
|
|
|
It looks that we are finally getting a new resonable formula.
Best regards,
Alexandru Savescu
|
|
|
|
|
Hello Davide,
I see that you are doing some R n D on article ratings. I have a humble request. Can you do a case study of my articles. I have always complained to Chris about low ratings and I honestly feel some of it is artificially fabricated!
I have 50+ articles and that should give you a wide enough range for comparisons. Most of my articles have good view counts but very poor ratings. Some of them are in the 2s and 3s and I have wondered why
warm regards
Nish
Author of the romantic comedy
Summer Love and Some more Cricket [New Win]
|
|
|
|
|
A case study on your articles based purely on page views, ratings and vote counts won't tell you anything. There are dozens of articles with high vote and view counts with deservedly low ratings, while just as many articles with viewership, high votes and undeserved low ratings.
The rating system has been changed to stop multiple votes and to weight votes by membership level. Things will sort themselves out eventually.
cheers,
Chris Maunder
he was a VB programmer, but he got better - Christian Graus
|
|
|
|
|
Chris Maunder wrote:
Things will sort themselves out eventually.
I hope so too. Anyway I've noticed that all my articles posted after you modified the rating system are a lot higher rated than my older ones. So I guess, that shows two things, (1) The rating system has got better and (2) My articles are improving, the improvement might be small, but it's still there 
Nish
Author of the romantic comedy
Summer Love and Some more Cricket [New Win]
|
|
|
|
|
Hi,
The results for the 55 articles are:
overall rating = 3.37 (low)
max energy = 7.60 (good energy - article: Some handy dialog box tricks...)
mean energy = 4.32 (low energy)
articles for beginners = 78%
articles with downloads = 33%
Not to be a teacher, or a guru, but I found some common defects in the
articles:
- Too much articles for beginners: the essay is easy to understand, so the
skilled CPians give a low rating.
- Too much articles without a playable demo: compile, run, and see what
happens is easier than read a good article.
- Too much preformatted monolithic blocks: interleaving some text between
the source code gives a better appearance.
Anyway the overall rating is really too low   , I have another
suggestion for Chris: when an author updates his articles, he could choose
to reset the rating (with a check-box in the submission wizard, or asking
directly to Chris): in this way he loses all the votes (bad and good) and
the article has a new chance to climb the charts.
|
|
|
|
|
Hello Davide,
I am thankful that you have done a case study for me ![Rose | [Rose]](https://www.codeproject.com/script/Forums/Images/rose.gif)
Davide Pizzolato wrote:
- Too much articles for beginners: the essay is easy to understand, so the
skilled CPians give a low rating.
Yeah, I have realized that by now that beginner level articles are not taken too well down here. Not just from the ratings, but also from the comments! Unless it is on a recent topic so that the majority of people reading it actually gain some information.
Davide Pizzolato wrote:
Too much articles without a playable demo: compile, run, and see what
happens is easier than read a good article.
Some of them cannot have a demo of any sort, but I guess you are right. People prefer a downloadable project file zipped rather than code they can copy/paste/modify to compile. Thanks for this suggestion. I had not thought of that before.
Davide Pizzolato wrote:
- Too much preformatted monolithic blocks: interleaving some text between
the source code gives a better appearance.
You mean that my sample source code is not spaced properly, eh? Thanks for that too! Anyhow if I actually give them a zipped project file I won't even have to put a lot of code in the actual article, unless they are short samples.
Thanks for your valuable tips and pointers Davide.
Davide Pizzolato wrote:
Anyway the overall rating is really too low
Yeah, puzzling as it may sound, that's true. Even allowing for an unpopular collection of articles the rating is till too low as I had voiced previously. But Chris M's new rating system has actually improved things quite a good bit.
Regards,
Nish
Author of the romantic comedy
Summer Love and Some more Cricket [New Win]
|
|
|
|
|
Nish,I wanna say happy birthday to you man.Hope we see you again as an active member in CP. 
Mazy
"If I go crazy then will you still
Call me Superman
If I’m alive and well, will you be
There holding my hand
I’ll keep you by my side with
My superhuman might
Kryptonite"Kryptonite-3 Doors Down
|
|
|
|
|
|
Nishant S wrote:
you are the first fellow to wish me this birthday
Nope.Take a look at soapbox.
|
|
|
|
|
Davide Pizzolato wrote:
suggestion for Chris: when an author updates his articles, he could choose
to reset the rating (with a check-box in the submission wizard, or asking
directly to Chris): in this way he loses all the votes (bad and good) and
the article has a new chance to climb the charts.
Good suggestion, a lot of articles start off poor but authors improve them over time and they become good articles. I second this idea.
Michael
Logic, my dear Zoe, merely enables one to be wrong with authority. - The Doctor
|
|
|
|
|
Hi all,
Can anybody point out any site for window XP tips?
thanks
regards
/rsasalm
|
|
|
|
|
*Please* don't cross-post. The OS forum was the correct place for this, just leave it at that.
... and sorry, i don't have an answer for you.
--------
Laugh at your problems; everybody else does. --Shog9 --
|
|
|
|
|
when no one answers in that forum,
then i think posting here is an understandable options!
|
|
|
|
|
I'm sure this question has been asked before, but I didn't look hard enough to find it.
How hard would it be to have each poster set his/her TimeZone so all the message dates and times would appear to be local times ?
Venet.
--------
Black holes are where God divided by zero.(Steven Wright)
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
