Printable View
Don't know if this has been stated by others, but the following thought occurred to me after watching one idiot ruin a perfectly good thread by making it go suicidal.
Why not make the post that is getting all the negatives get suicided out rather than the entire thread ? Or at least not suicide a thread unless the original poster is the one getting all the negs for the post...
Just an idea...
Neb
nebulus: I was arguing much the same on IRC just now...
I think the problem (as appeared in that thread (http://www.antionline.com/showthread...hreadid=232386) ) might however be resolve if the suicide algos disregareded negative point outliers (in the purest statistical way).
As in stats, if you don't discard outliers or use resistent statistical functions, biased results come out... I think this thread shows that... (Not to say that JP's algo is bad, but I do think outlier detection might help...)
Besides, the suicide note says:
I really don't think this was the case here...Quote:
This thread has committed suicide.
Due to the overall poor quality of the discussion that was
taking place in this thread, the AntiPoints System has
automatically closed it from further posts.
My 0.02$...
Ammo
This has been brought up quite a few times before.
If the thread was good, then other posts should have been given postive points. Yeah, it sucks that 1 post can suicide a thread, but if nothing in the thread has gotten any postive points, then the thread isn't worth keeping anyway.
http://www.antionline.com/showthread...uicide+Threads
http://www.antionline.com/showthread...uicide+Threads
FYI, for those who haven't taken stats courses and don't know what outliers and etc. are:
From stats book:
IQR=Q3 - Q1Quote:
Call an observation a suspected outlier if it falls more than 1.5xIQR above thrid quartile or below the first quartile
AmmoQuote:
To calculate quartiles:
1. Arrange the observations in increasing order and locate the median M in the orderd list of observations.
2. The first quartile Q1 is the median of the observations whose position in the orderd list is to the left of the location of the overall median.
3. The third quartile Q3 is the median of the observations whose position in the orderd list is to the right of the location of the overall median.
I can understand what you are saying about no positive points being given out and it therefore wasn't a good thread, but I think in the case of the article (which btw was the one I was referencing), the BigD post was so outlandish, I never had a chance to give any positives before the whole post went suicidal...which leads me back to what I was originally saying, the whole thread died because of one person's post and the appropriate response to that post (and alittlebitnumb mentioned that in a post in that thread shortly before it went suicidal).
I remember seeing the discussions about suicidal threads quite often; however, I didn't see (which doesn't mean it didn't exist) my particular idea or at least it expressed in the same way.
I very much agree there should be a way for bad posts to die and I was only suggesting a tweak to the current system that might prevent the above from happening...
*shrug*
Neb
I believe that GG posted a similar idea not too long ago. I think it had quite a few responses but in the end nothing was ever done about it. It seems that JP just doesn't want to do that. Either that or he hasn't been checking out all the posts. One of those 2.
Sorry for intruding coz i'm new here but you're all right about the negative threads.
Just coz some person decides to ruin threads there might be some good posts in them.
Somethings got to be done. :D
nebulus200, there will always be people out there to spoil everybody's fun... like crap in the punchbowl if you will; it's too bad things have to be that way, but the best way to go about attention whores like this is to simply let them get banned; however, just like any other system, it will never be perfect because there is always those that will re-subscribe and go back to being a pain.
Maybe banning the IP? I have no clue...
*shrug*
Anyone care to comment on the outliers detection idea? I think it makes sence: evaluating the AP status of a thread is the same as evaluating a statistical sample. As so I think statistical principles should be respected. Statistics is a science for which mathematical tools were developped and tested...
The only exception I can see would be to include the first post APs even if they should normally be ingnored if it's an outlier.
Does anyone see any counter indications to outliers detection?
(Seriously, I don't think this has been considered on statistical grounds before...)
Ammo
I guess we could all look at it this way - these losers (actually, maybe the same person over and over) that are creating multiple accounts and coming on here to mess with everyone and start something should just be ignored. I was reading some of the posts that are being referenced here and I couldn't help but think to just ignore these guys. BigD obviously came on here to start trouble and ruin threads, so just ignore the idiot and give them no points at all. That way they couldn't ruin the thread and pretty soon, they'd get bored and go back to playing with themselves or whatever it is that they do. It's ridiculous to have to put up with these morons, but that's part of being online. I say just ignore them, they may take the hint and leave....
But I also agree that if a thread wasn't getting any positives then maybe it shouldn't be there anyway...
Thanks for reporting, Souleman.
Thread closed.