Conceptual metric for DU to track?

[OAITW r.2.0]

I was wondering if any of the statisticians/analytics types could offer up a OP stat that shows a numerical value that evaluates 3 things

Recs/Replies/Views. I'd love to read the threads that have lowest # of recs vs volumes of replies and views. Some value that ranks by 3 decimal points. 0.000 to 1.000.

[EarlG]

1. You can get a rough idea of how current threads are doing from the Greatest Discussions page

If you go there and click the "Most replies" button it will sort the page with the highest number of replies first, but still display the number of recs on the left. So as you scroll down you can get an idea of which threads have a lot of replies but not many recs.

https://www.democraticunderground.com/?com=greatest_threads&sort1=replies

[OAITW r.2.0]

4. Thanks EarlG.

I wonder if you could add the views and allow the user to sort by all 3 variables?

Also, TexasTowlie brings up an interesting hack on views....hope to see your comment on that.

[TexasTowelie]

2. Too easily manipulated.

The ratio can be altered significantly by holding down the F5 function key. That will significantly increase the number of views and lower the ratio.

[OAITW r.2.0]

3. At what point is it manipulated?

When you edit a post, making a post? Seems like this is a real bug if that is the case.

[TexasTowelie]

5. Keep pressing the F5 key, it refreshes the screen which also increases the view count.

The ratios for recs to replies might be meaningful, but any ratio that involves view count is subject to manipulation. It was a trick that I learned on a local newspaper site where the view count could jump up 500-1,000 views by holding down the F5 key for about a minute. Some sites require the F5 key to be pressed each time while others get caught in a loop and constantly refresh as long as the F5 key is pressed.

[TexasTowelie]

6. The reply count can also be manipulated

by replying to oneself multiple times or by including a large number of +1 replies or other replies that aren't substantial as far as adding value. While the number of recs cannot be as easily altered (without using multiple DU accounts ), I believe that a case is made that the other two variables in your conceptual metric are subject to manipulation. Therefore, a conceptual metric making comparisons of replies, recs, and views isn't going to provide any relevance after individuals learn online behaviors that will influence the site metadata.

[OAITW r.2.0]

7. Thanks for the explanations!