Main Area and Open Discussion > Living Room
Congratz to the US Military Forces!
Tuxman:
I started late, so my count is reset, eh?
Shades:
Great comment, worstje. +1
The only relief you get from the 'eye for an eye' sentiment...is your money filling the pockets from weapon sellers.
Politics are not the main ingredient from DC. And I for one am happy about that.
tsaint:
I have come up with a new heuristic to help people tell whether they should post another reply in an opinion thread, here it is:
The probability of making a useful contribution to an opinion thread ~= 1 / (NumberOfRepliesYouPost^3) (graph).
* So your FIRST post has a 100% probability of having a useful impact.
* By the time you make your 3rd post, you are down to a 4% probability.
* By your 5th post you are below 1% probability.-mouser (May 04, 2011, 04:44 PM)
--- End quote ---
Aren't you overlooking the fact that there should be a summation of probabilities to determine the p of a member posting usefully?
Your suggestion runs the risk of encouraging posters to post an infinite number times in their desire to contribute usefully.
(To anyone contemplating this course of action, it should be pointed out that as number of posts -> infinity, p(useful contribution) -> .20205)
Let the robust discussion begin!
mouser:
It's a very interesting hypothesis, but I dont think the probabilities sum in this case -- I think the probability calculation covers ALL replies as a single entity -- i.e. the usefulness probability score covers your entire contribution to the thread.
So by posting twice you decrease the likelyhood that ANY of your contributions on the thread are going to be helpful.
Good you clarified that though -- otherwise yes i think we might see some infinite posters :)
tsaint:
If that were true, we'd have to consider the following:
1 post made: p(useful)=1. Therefore a useful contribution HAS been made.
2 posts made: p(a useful contribution has been made|1st post was useful) <1.
Contradiction?
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version