Reputation versus time

[michel123456]

I noticed that my own reputation increased considerably after passing 100 without any analogy to the quantity of my posts nor a significant increase in quality (IMHO). I am wondering maybe the reputation system is influenced by time: an ancient member getting reputation points only by the fact old posts are being voted positive by newer members who are influenced by previous positive voting. So as time passes by, I am getting more and more positive votes without doing nothing. Do administrators have statistical elements about this?

Edited September 17, 2011 by michel123456

[Hal.]

[michel123456]

Michel has 2053 posts divided by rep 149= 13,778523489932885906040268456376

 

Hal has 516 posts divided by reputation -34 = 15,176470588235294117647058823529

 

Hal is in about the same situation with me (reversed). As if members voted (negative) in function of numbers of posts, IOW time.

[Cap'n Refsmmat]

It's quite possible. Also, since the reputation points displayed on your profile are cumulative, you have a natural advantage from having more posts.

 

Some sort of reputation/post ratio, like you did above, would be less misleading. One could also hide the reputation display on individual posts, just leaving the +/- buttons, so you can thank the member without everyone knowing and following your lead.

[Hal.]

Michel ,

 

it takes you 13.8 posts to get a net effect of 1 positive reputation vote ,

it takes me 15.2 posts to get a net effect of 1 negative reputation vote .

Your number of posts is to 1 decimal place , 4 times the number of posts of mine .

 

Do you think that when I have 2000 posts I will have such a fine negative reputation as you have a positive reputation ? I don't . Most of the negativity happened in a few threads which I was on the opposing side of an argument to 5 - 10 people . This is beginners bad luck , or just natural argumentative nature , if I promised to pay a euro for each negative rep vote I'm sure I would be confirmed as very bad before too much more .

[swansont]

Do you think that when I have 2000 posts I will have such a fine negative reputation as you have a positive reputation ? I don't . Most of the negativity happened in a few threads which I was on the opposing side of an argument to 5 - 10 people . This is beginners bad luck , or just natural argumentative nature , if I promised to pay a euro for each negative rep vote I'm sure I would be confirmed as very bad before too much more .

 

There are a number of people (the majority, AFAICT) who give reputation based on the quality of an argument rather than agreement with its conclusions. I have disagreed with a post and gotten positive rep from the poster, and I have given positive rep without regard to agreement with the poster (i.e. I may disagree, but the poster made a good point and/or backed their position up with credible references) And I've seen others do this as well.

 

BTW, when you've made posts that earn you a suspension, that's not good evidence that it was being on the wrong side of the argument. >10% of your current rep deficit came from posting F**k off. That's not being on the "opposing side"

[Hal.]

Swansont ,

 

If trawling through previous threads to reacquaint myself with the circumstances of the said suspension and expletive has to be done , I can do that . Those in the higher echelons of power , you included , have already told me not to argue with decisions that are made , which , AFAICT in the here and now , looks like some want to have their cake and eat it . The time when I offered to explain what occured has gone , I'll put it down to experience and try to post with more wisdom .

[mississippichem]

There are a number of people (the majority, AFAICT) who give reputation based on the quality of an argument rather than agreement with its conclusions. I have disagreed with a post and gotten positive rep from the poster, and I have given positive rep without regard to agreement with the poster (i.e. I may disagree, but the poster made a good point and/or backed their position up with credible references) And I've seen others do this as well.

 

BTW, when you've made posts that earn you a suspension, that's not good evidence that it was being on the wrong side of the argument. >10% of your current rep deficit came from posting F**k off. That's not being on the "opposing side"

 

I know I give rep based on the quality of the argument, or I try to at least. Its not possible for me to totally distance myself from my own opinion or knowledge (sometimes lack of). In politics I often disagree with people but rep their sound logic. Science is a bit more cut and dry, for me at least (a post with agreeable logic usually yields an agreement).

[Phi for All]

I always give negative rep to posts that use insults. Likewise when someone repeatedly uses fallacious logic after it's been pointed out. I never give negative rep simply when someone is wrong and never just because they hold an opinion that differs from mine.

 

I don't use the positive rep as much as I should. Many people give such consistently great comments that I often forget to acknowledge them.

[michel123456]

the O.P. was not intended to enter an argument with Hal or anybody else, but once the question arises:

it is quite possible that HAL will never enter positive rep. just because his negative rep. will increase continuously due to new members reading old threads.

If I am not abused, even when a thread is closed the reputation point system is still available. I'll check this just now.

 

Personally, I never vote negative. I vote positive for a lot of reasons, entirely spontaneously when I like the comment, especially when it makes me smile.

 

If I am not abused, even when a thread is closed the reputation point system is still available. I'll check this just now.

 

 

Even in locked topics it works.

Edited September 18, 2011 by michel123456

[Hal.]

Michel ,

 

When new people with the ability to vote see a negative vote it's like a question , now that somebody has shown the way do you agree ? , whereas if neutrality is observed the person doesn't notice as much that they are being asked , do you want to make a positive or negative vote ?

 

Do you predict my past is on the way to haunt me ? - 75 this year , - 200 next , super bad - 500 before 2014 with a biased influence from a receding past ? Can I make a superpost to get 100 positive votes ? Will I challenge Newton's third law and risk immortal badness - 1 000 000 ?

[StringJunky]

While we are on the subject of reputation would it be possible to have a confirmation window of some sort after clicking plus or minus saying "Are you sure you want to vote this post up/down" (as applicable)" > Yes/No? It happens quite often that I click the wrong button unintentionally and it also may give a person a pause for reconsideration reducing knee-jerk reactions.

Edited September 18, 2011 by StringJunky

[michel123456]

Michel ,

 

When new people with the ability to vote see a negative vote it's like a question , now that somebody has shown the way do you agree ? , whereas if neutrality is observed the person doesn't notice as much that they are being asked , do you want to make a positive or negative vote ?

 

I agree.

 

Do you predict my past is on the way to haunt me ?

 

It is not a prediction, much more a question.

[michel123456]

Michel has 2053 posts divided by rep 149= 13,778523489932885906040268456376

 

Hal has 516 posts divided by reputation -34 = 15,176470588235294117647058823529

 

Hal is in about the same situation with me (reversed). As if members voted (negative) in function of numbers of posts, IOW time.

Keeping track today:

Michel has 3,103 posts divided by rep 260= 11,934615384615384615384615384615

 

Hal has 585 posts divided by reputation -32 = -18,28125

 

Conclusions yours.

But Hal has improved. (Hi Hal, are you here?)

Edited October 21, 2012 by michel123456

[Phi for All]

Keeping track today:

Michel has 3,103 posts divided by rep 260= 11,934615384615384615384615384615

 

Hal has 585 posts divided by reputation -32 = -18,28125

 

Conclusions yours.

But Hal has improved. (Hi Hal, are you here?)

His post quality was much improved after he was banned. He re-joined as ACUV and earned -2 rep in 85 posts.

 

It could be argued as well that the longer he was here, the worse his post quality became as more and more people objected to his posts. It tends to work that way with people who troll discussion forums. If they start out uber abrasive, they get banned uber quickly, and they know that, so they usually start out with more civility.

[John Cuthber]

While I agree that some sort of average points per post would be good, I thought about marking michel down for spurious accuracy.

[Klaynos]

I also think that rep/post would be good but the rep system has not been in place as long as some people have been posting which would complicate matters.

[ydoaPs]

I also think that rep/post would be good but the rep system has not been in place as long as some people have been posting which would complicate matters.

 

Yeah, our old system was completely different. It got to the point where swansont was so powerful, he was nearly a god when it comes to giving reputation. One click from him could make or break your rep.

[ecoli]

While I agree that some sort of average points per post would be good, I thought about marking michel down for spurious accuracy.

 

My posts per rate, as suggested by Michel, is somewhat abyssal: 26.48 reps/post, so I'm going to protest this

 

though it really should be rep/posts so the number is fractional and increasing with reputation.

 

Actually I think would have to be a bit more nuanced than this simple rep rate. Maybe a 'bayesian' model where each rep vote is taken as evidence for your true reputation, with current rep as a prior. Could hopefully counteract longevity a bit.

[Klaynos]

A good quantifier might be the fraction of the total rep handed out during the last week that went to your posts...

[ecoli]

A good quantifier might be the fraction of the total rep handed out during the last week that went to your posts...

how would you handle zeros? Either from zero posts or zero rep

[Klaynos]

how would you handle zeros? Either from zero posts or zero rep

 

Post count wouldn't matter. Which I now see is a fatal flaw.

 

Your rep in the last week / rep given in the last week

 

You'd need some logic to take account if rep given in the last week was 0...

[ecoli]

Post count wouldn't matter. Which I now see is a fatal flaw.

 

Your rep in the last week / rep given in the last week

 

You'd need some logic to take account if rep given in the last week was 0...

So if you have a slow week your rep disappears? This seems like a good idea to encourage active posting, but probably a good way to kill the whole rep system.

[Klaynos]

So if you have a slow week your rep disappears? This seems like a good idea to encourage active posting, but probably a good way to kill the whole rep system.

 

Yep, that's the flaw, needs to be normalised against post count as well. It could encourage people to stop posting for a week after getting a well received post though.

[ecoli]

how about something like this:

 

[math] \frac{r_w}{R_w + P_w} + \frac{r_t}{R_t + P_t} [/math]

 

[math] r_w [/math] is a user's rep from the past week

[math] R_w [/math] is the SFN total for rep given in the past week

[math] P_w [/math] is the SFN total post count from the past week

[math] r_t [/math] is a user's Total rep

[math] R_t [/math] is SFN's total rep

[math] P_t [/math] is SFN's total post count

 

haven't fully vetted this or simulated a range of variables but think it accounts for a reasonable mix of close range, total factors and reputation and posts, though perhaps not in the exactly correct balance. Also could add a term to balance the other user's average score.

 

Weekly could easily be changed to any arbitrary time period.

Edited October 21, 2012 by ecoli
redundant variables