Should outlier votes fall off over time?

[LongDraw]

What I'm suggesting is relatively simple and is used to improve statistical accuracy by removing inaccurate results through collection of larger data sets. So basically it works like this:

Story A is published. Hour 1 there are three 5* votes, hour 2 there are four 1* votes, hour 3-5 there are ten 4* votes followed by another five 1* votes over next 3 hours.

So Story A would have a (64/22)/5 = 2.909/5 score.

Let's say over the course of the week, there are another 78 votes cast, majority 3* or 4*, with maybe another 5-10 votes being 1*. This should indicate 1* vote bombing occurring. So to correct the malicious voting, 30 days later if the story average score is 3.00-3.99, you remove all the 1* votes cast in prior month. If average score is 4.00-4.99, you remove all 1* & 2* votes from prior month.

Thoughts?

 

[KeithD]

My thoughts are that Literotica is not a "collaborating with the users" Web site, and "what ifs" are pretty much wasted effort. Tossing out outliers is probably done to some extent by the regular voting sweeps the Web site does.

 

[ElectricBlue]

The site carries out a regular sweep process (and more frequently during contests), whereby suspect votes are culled - we don't speculate as to how it works, but suffice to say, it does.

Your proposal targets what could be legitimate votes, which is unfair on those votes.

 

[jaF0]

Thoughts?

Main reason I bailed on the AH is all the pissing and whining about votes and ratings. When you can buy a cup of coffee with them, let me know. Otherwise, confine this silliness to the AH.