Jump to content
Sign in to follow this  
Professor Hantzen

XRP Chat Top 100 Members By Reactions Per Post (RPP

Recommended Posts

When reading through the forum, I often unconsciously divide a members Reputation by their Posts in my head, to get a feeling for how their contributions are percieved by other members.  A higher ratio of Reactions Per Posts ("RPP" ;) ) shows that more people may be willing to demonstrate they value those members contributions.  I find it interesting, but also useful.  I might not read a long post by a member I don't recognise if they have a terribly low ratio, for instance.  Perhaps others do something like this too?

Recently, when @BobWay joined, I noticed that his RPP climbed very (x)rapidly up to being much higher than what I'm used to seeing on a regular basis.  It reminded me of something I've wanted to do for a while - scrape the member data from the site, and make a chart to see where all members sit by this metric.  With a few caveats (one briefly on the chart itself, more in detail below), here it is:

xrpchat_top100-byRPP.thumb.jpg.e5cc511b1c33225fff987311062c3e60.jpg


Caveats (this list may not be exhaustive...!):

1) One limiting factor with this chart is with how people digest and respond to information when all eyes are upon them.  For many people, truths can be uncomfortable, difficult or even completely unwelcome - and even when they do agree or value an insight, they don't always want to "own up to it".  I like to think people are pretty open-minded and flexible here, but I've still seen this pretty normal human behaviour from time to time, as anyone would reasonably expect.  So, while I believe this chart give some useful indication of something, it is not prone to including those who may regularly say things other people aren't willing to show they agree with, even though they may agree, or even though those things may be true, important and valuable.  There's probably nothing that can be done about that, except to note that if there's a member missing you expected to find on here - well, you may have a good point.

2) The chart does not use a complete set of members who have Reputation.  If it did, it wouldn't be very useful as it would likely be filled with outliers who made one post, received a handful of votes and never returned.  And their inclusion would be at the expense of significantly more established members who might be hard to see among them or even drop off the chart entirely.  Partly as a fix, and partly because I didn't have time nor inclination to click "Next" 200 times, I picked range of reputation to include and culled the rest below it. The solution is perhaps equal parts reasonable and unreasonable, given that the point at where some members are included versus excluded is essentially arbitrary.  Members on the boundary could theoretically be one reaction away from topping the chart in at #1, or disappearing entirely.  Nevertheless, a compromise simply had to be made.  If anyone has a suggestion about how to better go about it - I'm all ears.

3) As I understand it, the "Reputation" score on the forum counts any "reaction" of any kind (and hopefully I've got that right...!).  In theory, someone could game their way to the top of such a chart by posting only tremendously sad and confusing things.  In practice I doubt someone would be able to gain much traction with their posts - more likely they would be ignored - so I think the metric stands a pretty good chance of being at least vaguely indicative of how members posts are perceived.  (Also, in my anecdotal and unscientific experience, I see the positively-connotated Like, Thanks and Laugh buttons clicked much more often, in that order.  The negative Sad and Confused buttons seem to be used at the low ends of the bell curve.)

What gives me some hope that this metric is valuable in spite of these caveats, is that the resultant Top 100 is populated with almost all of the current and former Ripple Employees on the forum that I am aware of, and all of them within the top third of the chart.  That was pretty cool to see!

Anyway, see what you think, and if this is valuable/useful maybe I can do it again in a month or so.

Share this post


Link to post
Share on other sites
Guest
3 hours ago, vsyc said:

Comclusion: RPP is useless.

It’s a rare day that I fully agree with you sir or madam,  but today is that day.  :) 

Thats not to say I don’t respect the effort and thought that the good professor (@Professor Hantzen) has put into this but I do think at heart the metric is based on a fickle and unreliable statistic.  

Share this post


Link to post
Share on other sites

Given the amount of caveats and the lengths I needed to take to describe them, I somewhat agree with the negative assessments.  Though I still find it interesting that the Top 100 by this metric shows 90% of all of the ex and current Ripple employees that I am aware have ever used the board.  That they reliably cluster together at the very top when ordering by RPP makes me disagree that it's useless.

Moving the cut-off point to a higher reputation (~700, arrived at by limiting to the first ten pages of results), the list is still populated with many ex and current Ripple employees.  In that case, there are three within the top 5, two more within the top 15, and no others within the top 100.  The four excluded by the shift in cut-off point all only posted for a short period of time before disappearing, so it would appear to work better at a higher setting in that regard.  It's also populated more thoroughly with names even I recognise (I'm not a particularly frequent contributor), so maybe it does indeed show more useful results at a higher cut-off to remove a greater amount of noise/outliers.  I may redo it.

@vsyc RPP dilutes Reputation by weighting it against Posts, so at base its an improvement over Reputation alone. It penalises high-volume, low-medium-response posters in favour of low-medium-volume, high-response posters.  (Also, I must note you have a negative RPP... ;) )

Thanks for the kind words @Tinyaccount:)

Share this post


Link to post
Share on other sites
On 3/15/2019 at 8:08 PM, Professor Hantzen said:

might not read a long post by a member I don't recognise if they have a terribly low ratio, for instance.  Perhaps others do something like this too?

 

9 hours ago, Tinyaccount said:

this but I do think at heart the metric is based on a fickle and unreliable statistic.  

 

1 hour ago, Professor Hantzen said:

Also, I must note you have a negative RPP... ;) )

Oh lol. Good discussion here.  While RPP gives an indication of how members value certain contributors, it doesn’t indicate with 100% correlation what kind of posts they exactly value:  I think this varies in time: times of appreciation for the more technical posts alternate with more appreciation for jokes, TA and price speculation or marketing content. Would be interesting to analyse the changing preferences ( back and forth) in time. Depends on the mood of the day and the number and category of visitors online at any time. 

There are periods I only read media content , in other periods I like the mix of infotainment of the Zerpening, and spreading my likes accordingly.

I don’t really look at the RPP,  but ofcourse I pay more attention to members I know well or like more than some others. Yes, I’m human, that’s how social interaction works. Very deepdiving technical posts don’t interest me much, I just drive my car without looking under the hood.

But agreed: a negative RPP is a serious signal and maybe a red flag to put someone on the ignorelist:P.

 

Edited by Ripple-Stiltskin

Share this post


Link to post
Share on other sites

Most of the time I see the same forum avatars making quality posts so I spread some RPP to show them I like what they said and to keep posting good material. On any forum there's always those people that everyone likes bc of the way they form their thoughts and articulate it. Over time people recognize them more often since they are able to 'trust' their posts.. Some people take it more seriously, some people post infrequently and never really obtain much RPP, even though they might have good ideas. 

I think that people are more likely to give someone RPP if they recognize them as someone who provides value to the community on a consistent basis compared to an account that posts once a week. 

Pretty much if you have a recognizable avatar and are a regular that posts decent content, you'll obtain more RPP. 

None-the-less it's just like Karma and only here to inflate our e-peens

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  

×
×
  • Create New...