How many people did you speak to / spoke to you last week? (Engagement League scoring potential)
Sort of like a test run for this weekends Engagement League. Potentially, both these numbers will be added and scored.
I'm trying to improve, without breaking, how to gauge interaction and engagement with regards to the Engagement League I host each week, which is sponsored by @curie. The sponsorship and level of support shown for the post means that I take it seriously, and wish for it to be as fair and interesting as possible.
Counting the number of unique individuals engaged with each week seems like it might fit the bill for a new metric. And in turn, the total number of unique individuals that replied to you could also be of worth - maybe a combination/calculation of both is in order?
For the 'individuals spoken to', here's what I have so far:
select count(distinct parent_author) from Comments WHERE parent_author <> 'abh12345' and author = 'abh12345' AND created > DATEADD(day, -7, getutcdate()) AND depth > 0
So basically, the number of people that replied to me (who were not me!). This one is pretty straight-forward, and I think i'm OK with the result.
For the number 'who spoke to you', we have options.
- Do you think that the number of people that spoke to you, excluding top level replies to posts, is the way to go? Or do you think top level replies should also be counted?
- Also, do you think that one or both of the above counts should require a minimum character count for the comment/reply?
e.g. 'Nice post, resteem me sir' will not be counted if the reply is shorter than 40 characters
The SQL here could be something like:
-- PEOPLE WHO SPOKE TO YOU
select count(distinct author) from Comments WHERE parent_author = 'abh12345' and author <> 'abh12345' AND created > DATEADD(day, -7, getutcdate()) AND depth > 0
-- PEOPLE WHO SPOKE TO YOU excluding top level replies
select count(distinct author) from Comments WHERE parent_author = 'abh12345' and author <> 'abh12345' AND created > DATEADD(day, -7, getutcdate()) AND depth > 1
-- PEOPLE WHO SPOKE TO YOU excluding top level replies and replies less than 50 characters
select count(distinct author) from Comments WHERE parent_author = 'abh12345' and author <> 'abh12345' AND created > DATEADD(day, -7, getutcdate()) AND depth > 1 AND len(body)>50
Regarding scoring unique interactions
Engagement can be considered to me, an in-depth discussion requiring many replies between two people. It is also, a shorter two/three threaded (depth) conversation.
Which of the above holds more value?
Because If these metrics are added, the people who engaged with the greater number of people would benefit the most. Is that fair?
Gamification - just say 'thanks but no thanks, sir'
Since I released all the metrics used earlier in the year, I've always had slight concerned about the possibility of gaming the results.
Thus far though, I'm happy to report that there has been little to no match-fixing going on. And whilst I try to check each week on suspicious looking activity, I also place a little of this on the entrants and ask them to check on how folks are generating their numbers. Members have also come to me and ask to be removed. e.g. due to there copy/paste commenting for a community initiative, where they would go crazy if they were to type the same thing over and over.
I would like to hear peoples views on the potential changes above, before I go ahead and alter the scoring structure. Mainly because I don't want to balls things up, and well, the Engagement League is here because of you guys - you should have some sort of say, right?
Today i'm going to bribe you for your 'number of people spoken to', 'number of people who spoke to you' by offering these numbers, IF you give some sort of feedback on the above!
Let's hear what you got to say :)
Almost forgot, your reply might include this:
| Metric | Count |
|---|---|
| Unique accounts you spoke to | 178 |
| - | - |
| Unique accounts that spoke to you | 174 |
| Unique accounts that spoke to you, excluding top level replies | 116 |
| Unique accounts that spoke to you, excluding top level replies and replies < 50 chars | 88 |
In other news, my auto-voter is off for 12 hours starting from now, that's what you get for hitting the button 800/900 times each week :(
Thanks in advance!
Asher
Ok this is one I'm curious to see my numbers.
I'll have to read this several times to really absorb it (you know me! Can't you just write a personal touchy feely post? :), but the thing that jumped out most is the question about how many we spoke to as opposed to those who spoke to me. Not sure how or why "those who spoke to me" should affect my stats? because I can't control anyone here but me and my typing fingers :)
I may have more for you later, but for now, it's all I've got :)
This is enough for me, and is in line with how Glen and Mike have commented, and likely others who i've not got to yet.
I'm thinking just the 1 metric, and perhaps with a secret min character limit to keep y'all on y'all toes. yeehaa!
Dunno what happened there, I blame too much coffee :)
Whoops, almost forgot....
Thanks Asher! Some things are nearly always finished before they get started :)
Your list is at the top, so I haven't gone through to see how others fared, but you seem to have done pretty well over the last seven days. I was happy with mine but yours are quite a bit more, so good job. I know I have to widen my circle quite a bit and this just proves it. :)
I rarely compare myself to others, but after your comment, I scrolled down to have a look and wow, I "have done pretty well" if I say so myself :) Thanks Glen!
And for the most part, I think it's a good idea to avoid comparing ourselves to others. Too much to be lost and not enough to be gained. But every once in a while, I think it's worth a look, if nothing more to get a glimpse at what is and what could be. And if we measure up, then we can also take a moment to acknowledge and enjoy life's simple pleasures and small victories. :)
So true Glen; I can't add a thing more to that. Not even a question :)
I should have read your reply before I typed my own, hahaha.
I had the same problem, lol. That technical school we both attented didn't prepare us for Asher's posts :0)
I had the same thought. No matter how hard you try to keep a conversation going, the fact that you get a reply or not is not up to you...
Sorry!
I did realised that this post may not reach the engagement highs of others, but I've had some useful feedback so far :)
Lol, it's not you, Asher, it's us.
@lynncoyle1 and me are both very slow students when something technical is involved ;0)
hahaha @simplymike. We did both attend the same technical school after all, and I can only assume that your daily water bottle was also filled with vodka :)
Back them I was still a very good girl. I had to make it all up afterwards 🤣
haha @simplymike, that's awesome 😅
As always super interested in this type of data, would you provide me my counting?
Seeing as you asked nicely :)
As always! Thank you!
Yess! Great update!
Exclude top level replies, which have nothing to do with our engagement. When we reply and generate more engagement, it will count.
A character limit has pros and cons. Some of my best replies are haikus, which are usually pretty short, but could now provide bonus engagement by summoning the haikubot.
I'm not sure how worried you should be about gamification, this is gamification! I happen to take the goals seriously, ie I read every post and try to reply from the depths of my heart, but I am doing certain things like staying up late to engage more because you made a game out of it.
How can you compare sincerity among participants? If there are top users who's comments I believe are less sincere than mine, what metric would I use to prove it? Life is a game friend, and we are all playing it, in our own way.
I get upvotes on my hundreds of comments, and I think, not sure, that they are even more valuable than top prize here. I make more new friends, find more new people to follow, and get more new followers, all because you made a game offereing a few steem. And I like to win ;p
Oh, and I like the 'spoke to me' stat, because that does have some relevance to how engaging my comments are.
You are the first to side with this, and speak positively about the 'spoke to me' stat.
I've included top level comments this week, and not added a 'spoke to me' score for now. If i get more people agreeing that this is a useful metric then I will add it.
My thinking was that if a reply was engaging enough to warrant a response, this should be scored. But others suggested it was not fair to be scored on if someone replied to you or not...
Good point re:haikubot!
Thanks for the comments!
Wow, you are fast! I know I didn't specifically ask, but I suppose I will see my stats in a few hours on the main post right?
I have a tendency to have contrarian viewpoints, but I don't mind if I get outvoted, it happens a lot. Steem just happens to be a great place to share views and (probably because of rewards) most people are so nice!
Hi Asher. Just repayed the kind gesture SBI you sent the other day. Thank you again.
Very kind of you, thanks very much!
I did get sent a link to check how many shares are owned but forget where I put it :)
Call me dense if you want but I do not understand what a top level reply is, so not sure how to answer that one. Not a lot but sometimes when people visit my pages, the comment left really only requires a thank you. So it is good you will be filtering them out. (Polite does not mean engaged).
I'm fine with being a test case, to see the numbers. Also what is a top level reply?
I may not have worded it so well, but I class a top level reply a first reply to a post. There can be many of these of course, and you can reply to a post (in a new thread with a top level reply) as much as you like.
On my phone right now, if I forget your numbers (laptop required) then please remind me. Cheers!
Okay, I understand now. I was slightly confused about the term. basically the first reply to someone or to a post. Thanks.
Yes that's correct :)
Better late than never :D
thank you for that.
Hunh. Don't see where bashadow reminded you about this list, nor do I see the list, so I'm going to remind you for bashadow, since I'm pretty much curious about what everyone is doing. You can ignore this reminder if you like, though. :)
haha :)
I think i'm up to date with regards to recent list offerings for @bashadow, but thanks for the reminder all the same!
Okay, well, then I need to go back and look for it then. From what I can tell, i'm only trailing Lynn so far, but it's like comparing a cheetah to anything that might be significantly slower than a cheetah. :)
Just curious, how are image/meme replies calculated? Is it by the actual character count in the code that's embedded? Does that also hold true for any formatting, bolding etc?
Interesting one. I think it will count the code that's embedded. Don't tell people though or i'll have to add more code to fish out URLs :D
Yes I suspect formatting will add a few characters too - same as above will need to be applied!
Awesome work. Thanks for that. The 50 character thing is interesting, as shorter comments can still be engaging, but at the same time, spammy comments greater than 50 characters are pretty disengaging too. I don't envy you trying to quantify this on a qualitative basis at all.
Cheers :)
Yeah, it's nigh-on impossible to do so really - but this is what we have to work with so i'm giving it, with the help of you guys, my best shot.
That all comes down to what is the desired outcome. Is it better to engage with a wide audience or a select few. Can you have some really amazing engagement with 10 or some really vague with 1000. Which is better overall? Ideally none of these in my eyes its somewhere in the middle.
Part of engaging with others should be about going outside of your own bubble. I think at some point you are going need to have diminishing returns that are applied to the numbers. Where that is and how fast it is diminished is all up to you.
I personally know I’m not getting out there enough with different people. Which is not something you always think about unless it’s kind of pointed out in a way. It's easy to fall into a maintenance mode way of dealing with things and keeping in touch with people you have meet.
I do like your mention of a char minimum. 50 sound a little too high. I also think that eliminates anyone’s normal reply back to from something like “thanks, thank you, will do, will get on it.” While that in itself is engaging and could be meaning in letting someone else know. It is also rather generic.
I think the biggest indicate is also just going be too much work to try and track as it would be per person and case. Is someone engagement increasing their performance overall on the blockchain week after week with how much engagement and just votes towards their own comments and blogs?
That great if someone can put out 1k comments a week. But to what end for that person is that doing for them? Someone with only 20 comments could really be out performing that person.
Hey @enjar
Thanks for the reply. I think you part answer your own initial question with:
But i understand that 'somewhere in the middle' is best, it's just that:
Yes, and sadly I don't have the time or inclination to do that each week, so we have to work with numbers.
The char minimum of 50 or anything around that could end up being just a number that people know they have to pass to boost their score. Is that a bad thing though if it makes them produce a more complete sentence?
With regards to performance, I'm not sure, but hope that being a part of the EL helps - even if it's just via the prize money each week.
For sure. Likely someone with stake (applied to self) :)
I think a privately known changing from week to week min. char variable would be the way to go. If someone thinks they need to put out 100 char comments to meet the min when it was only 20 that week. I can't but help laugh that there plans to game the system caused them to be gamed by it instead.
Perhaps that is the solution to your issue. Just needing some slight changing variables every week. Try and shake up the top a little bit. Have some of that "magic sauce" in the numbers so people have to guess. While it does kill transparency. Real meaning human engage is the end result.
That is always the issue when running an experiment. You change the results and behaviors from studying the subjects.
Good plan, I had considered hiding a few variables, but thought transparency was nice - even though the 'weight' to each criteria I don't share.
Gamed by the system, muhahaha.
(added an extra 'ha' to hit the min!)
100% upvote for this comment! As here, I don't think a char limit will make my comments better, just longer.
To be honest this is the first rule that has seemed like work, dimishing just a little the fun of engagement league achievemet hunting.
This might be different for different types of posters. Doesn't seem like it would affect @glenalbrethsen at all ;p
I dont think it will affect many of the key engagers - it's there to stop 'nice post', but it would also stop 'thank you' as well. :/
We'll see how this weeks scoring is reflected upon, well done on reaching the top 5!
There is no perfect number. Recently I have been interacting with this guy, who loves puns.
Oddly enough, using the partiko app might be enough to get around any minimum character length; is the markup included or only the visible text?
Thank you! I am heating up, and having a blast on steem! This is unsustainable for me in the long run, but I am enjoying every minute of it while I can!
Otherwise you need some “guest judge” numbers every week. You have a pool of data you could consider but you only use X number sets. If someone wants to game the system well they have to do a lot more work since they don’t know per week to week what is going be used.
That is what I love about running contents with human judges. I’m constant but I switch up the other 2 for each one from a pool selection. Someone might be able to write something to what I consider “a winner” but you still have those other 2 that change per contest.
Yep that's another good suggestion to consider.
I'll see how things go with the additional column this week, and if I/others spot any fishy activity, then I may need to start a little gaming myself :)
I think anything you add will only make it better. Counting the number of unique accounts interacted with seems like a fair measure to me.
I am rather new to the leagues so it is hard to say but I suppose it would make sense to have a character length requirement.
I have been wondering about what sparkesy43 asked how are the images counted? Just curious as I am not really a meme type person myself.
Cool, I think that column will be there on Sunday.
I replied to Sparkesy, and think that the code to embed an image or use markup is will be counted, unless i write more code to fish it out :)
Thank you! I really love your little charts :)
Hey, finally, I get to come in on a list post within 40 minutes!
Thanks for the trial run on these possibilities, by the way. I'm interested to see what mine look like, as always, so if you please. Thank you.
You already know my thinking on the CL limit (yes, there should be one), and as far as engagement with unique accounts go, there are going to be many where it's one and done. I think as long as the the limiter is taking out the "Nice Post!" comments, the top level probably should be counted as long as it gets over the CL threshold. Problem is, are we just counting unique users, or are we counting engagement with unique users? There's a difference, obviously.
Also, the question about spoken to versus spoke to you: we only have control over who we speak to, so I'm kind of leaning to that, just because the other is depending on someone else. I mean there are things that we can do, but if we're talking for purposes of best portraying the engagement of individual members of the league, I'd think going with the metric that we have the most control over would be better, right? :)
Posted late on in the day for your early viewing pleasure :)
It would be a comment to a different user (over the min char limit) that would be counted...
I see what you mean with spoken to, spoken to you. I was wondering if a spoken to you would represent an engaging comment that was worthy of a reply - even though we can't force anyone to do that. Perhaps it could be given less 'weight' that the metric that can be controlled? Or just excluded for now :)
Well, it's your engagement league, so whatever you feel is best is fine with me. Sounds like there will be plenty of added metrics by the time you're through. :)
Thanks for the list. It's higher than I was expecting so, that's always good. Let's see how it holds up with some of the other Top 10 folks. :)
Cheers Glen.
I think I have enough to go with, just wanted a few more eyes on the situation. I'd say you are doing better than average for sure :)
You're right, you are far from Lyncoyles numbers. But of course you still beat me. While I was hoping a new metric would bring me to the top of the engagement league.
Next week I suggest that we also include the number of useless comments or something like that. Cause one day I will win the engagement league. One day in 2052 😋😜😆
Hi Asher. I think you have to have a minimum 50 character reply to count. A thank you is not a reply that should count. Also what if I reply and edit the reply does my word count go up as well? It was just a thought.
Yes I think i'll go with a minimum count, even though a thank you after a few replies back and forth is a reasonable way to end a conversation :)
The word count should go up on editing a reply, yes. Thanks for the reply.