Suggestion Sorting Threads by Date, Replies, Views & External Resonance

  • Thread starter Thread starter A. Neumaier
  • Start date Start date
AI Thread Summary
The discussion centers on the desire to enhance thread sorting options in forums by introducing a metric for "external resonance," calculated as the number of views divided by a constant offset plus the number of replies. Participants argue that current sorting methods, which include date, replies, and views, do not adequately reflect the engagement or interest level of threads. The offset value significantly influences the ranking of threads, particularly when comparing older, popular threads to newer ones with few replies. There is skepticism about the impact of web crawlers on view counts, with some believing they do not skew the statistics significantly. Overall, the proposal aims to better highlight threads that generate genuine interest and engagement among users.
A. Neumaier
Science Advisor
Insights Author
Messages
8,679
Reaction score
4,739
It would be nice if, in addition to
- sorting by date
- sorting by number of replies and
- sorting by number of views
it were possible to
- sort by external resonance, defined as (number of views)/(20+number of replies).

Or another sensible number in place of 20. Note that 1+number of replies = number of posts, but using this in the denominator would give too much emphasis on posts with no or very few replies.
 
Physics news on Phys.org
I think the sort options are what are provided by VB, so it might take some significant hacking to provide more.

Changing the offset from 20 to 1 (or any other number) would not affect the rank by "external resonance" of a thread.
 
Gokul43201 said:
Changing the offset from 20 to 1 (or any other number) would not affect the rank by "external resonance" of a thread.

Unfortunately, the offset makes a difference:

Given the current PF statistics in the forum on quantum physics, a well-received thread W might have v=500 views and r=50 replies, while a new thread N might well have v=50, r=0 before the first reply arrives.

The thread W will have higher rank than N when k is 6 or higher, whereas for k=5 or lower, N will have the higher rank.
 
Relying on views doesn't make sense, many are by crawlers.
 
Borek said:
Relying on views doesn't make sense, many are by crawlers.

I don't believe crawlers are counted. They'd be much much higher.
 
Greg Bernhardt said:
I don't believe crawlers are counted. They'd be much much higher.

...and if they would dominate the numbers, the ratios of views to posts (which vary at least between 20 and over 500) should be much more uniform than they are.
 
A. Neumaier said:
Unfortunately, the offset makes a difference:

Given the current PF statistics in the forum on quantum physics, a well-received thread W might have v=500 views and r=50 replies, while a new thread N might well have v=50, r=0 before the first reply arrives.

The thread W will have higher rank than N when k is 6 or higher, whereas for k=5 or lower, N will have the higher rank.
Oops, I misunderstood what you were saying. Yes, agreed on the distinction. Not sure I've ever cared about an index of that nature in all my searching history though.
 
Greg Bernhardt said:
I don't believe crawlers are counted. They'd be much much higher.

Best known (google, yahoo and others, phpBB has a list of about 40, not sure about vBulletin) are filtered out, but I see from my server logs that there are some that show at random. I have never tried to delve deeper.
 
Gokul43201 said:
Oops, I misunderstood what you were saying. Yes, agreed on the distinction. Not sure I've ever cared about an index of that nature in all my searching history though.

Well, I noticed that some thread have a _much_ higher "external resonance" than others, suggesting that very many people look at it because they expect from the title something interesting. So good answers to such threads would have much more impact than answers to a thread with little "external resonance", which at best only satisfies the original poster and a few regulars...
 

Similar threads

Replies
3
Views
2K
Replies
6
Views
2K
Replies
7
Views
2K
Replies
6
Views
2K
Replies
15
Views
6K
Replies
7
Views
2K
Back
Top