Discussion Forum by Ziggy C.
August 16th, 2007 2:27 PMBefore you call me out on a retroactive completion, please understand my motive for submitting this: I could care less about the points. God knows I haven't submitted a task in months. Really, I felt that submitting this was appropriate because I sincerely want to hear what others think of my idea. Flag if you must, but give me your opinion before you do so!
Also, I apologize for the wall-o'-text. It's a long read, but please read through it, for I think it would benefit our game immensely.
---
The following was sent through SF0's contact feature:
"Just an idea that I thought of, since we're still in beta...
Maybe apart from raw point rankings, we can have something like "player satisfaction" rankings/scores?
It's fairly agreed upon that votes are symbols of satisfaction from other players. I propose you calculate a player satisfaction score simply by dividing a player's total score by the number of tasks they've completed.
For example, let's take two hypothetical players "A" and "B". Player A has completed 20 tasks with a total score of 400 points. It's obvious that this player is going for quantity over quality, and has not received many votes.
Player B, however, has only completed 5 tasks, but also has a total score of 400. It's obvious that this player's completions are well received by the rest of the community, seeing as it would take a lot of votes to reach such a score with only 5 tasks.
Player A's PS Score: 400/20 = 20
Player B's PS Score: 400/5 = 80
Even though they have the same total score, Player B's completions were enjoyed much more by the community, so they will have a higher player satisfaction ranking/score
Why would we do this? Of course, a player can just do as many tasks as possible with no creativity and get a high score, but if they do this, their player satisfaction score will be low. I believe this player satisfaction ranking and score would give new players an incentive to truly do something creative.
It also provides an interesting meta-game and balance, juggling between keeping up a high raw score and a high player satisfaction score. I think it could provide the type of middle-ground this game lacks, seeing the recent debate (http://sf0.org/taskDetail/?id=1282) about quantity vs. quality.
Even if you don't use my means of calculating the score, I personally think something like a player satisfaction level would be a great addition to SF0. Let me know your thoughts about this, even if you think it's a horrible idea.
Thanks!"
Also, I have since thought through a couple more mechanics of the situation. Of course, if someone right off the bat does an amazing completion that was only worth 15 points, and gets a ton of votes, their player satisfaction score would match their raw score, and would rank ridiculously high on the player satisfaction level.
To counteract this, maybe a player must complete a certain number of tasks before they are assigned a player satisfaction score. Not to mention the fact that their raw points score alone would still be incomparable to others, which would act as a motivator regardless to submit more tasks and still put a lot of effort on them to maintain their high player satisfaction score.
One could also argue that such a score could easily be measured simply by the total votes one has. I, however, feel this is inaccurate, and here's why: Let's say one player has 30 mediocre tasks, and another player has 5 near-perfect completions. While the former would technically have more votes, the latter's completions were generally much more well received by the community. If the player satisfaction score was simply the raw number of votes they've received, the former would still be ranked higher, despite the community being generally more satisfied with the latter's completions.
I have calculated many people's player satisfaction scores with my formula, and I recommend you give it a shot yourself. I feel it computes a number that is a fairly accurate measure of how satisfied the rest of the community has been with that player's completions. I will add a chart of the current PS scores to this completion if this idea is well received, or if you feel you need to see the current standings to see how well it works before you can judge.
I will add more to this as the idea further develops. Tell me what you think!
4 vote(s)
Terms
(none yet)17 comment(s)
It would be quite intriguing to see...
I tend not to think of points in sf0 in a competitive sense for the reasons you outlined. I see points more as a tool for leveling up, not as a statement of how good a player is. A PS score would give us a better idea of who is really doing well in the game.
I really like this. The idea of balancing a raw score in some way has a lot of appeal to me. This could be a nice additional metric to guage our own completions, and if coupled with some of the ideas given by Charlie Fish regarding Reframing the Reset it would be a great way to balance the quanity/quality situation. I generally support all movements towards quality, and therefore more emphasis on the vote.
only one random idea/concern
But, does this discourage the completing of certain tasks? In my estimation some tasks seem to be simply incapable of capturing more than 8/10 votes. I don't care if William Shakespeare wrote a Macrofiction, i would be surprised if 10 people read the whole thing and voted. Would it be possible to integrate into the 'satisfaction' formula some way to compare with the other completions?
If you get the max score with three votes on an unloved task, this should raise your overall 'satisfaction', while a completion with 8 votes when the community favorite has 25+ votes would not significantly improve it. I guess i am saying that there should be some stronger interelation between tasks and votes, rather than tasks & points (I know that votes & points are related, i'm just not sure if it is properly 'balanced' and if that is the goal then this should at least be considered and discussed.)
I like the idea, especially with something like YellowBear's modification. Finding a proxy for goodness other than straight score or vote count is intriguing, but it seems to me there's a real danger of discouraging people from doing little simple tasks.
My fall-back statement when such questions arise in life is generally to say that "more information can't possibly hurt." But, here that may not be true. We are, after all, playing a game, and the scoring rules of the game will influence the way people behave.
I know that I'm a little hesitant to embark on a decidedly non-awesome straightforward completion of an easy task even *without* a quantitative metric. I've been actively trying to convince myself that sometimes it's more fun to do a simple task than to *not* do one, even if it doesn't thrill the praxis readers.
In the case of passive media, it's more obvious that not everything one does has to be epic in order to be worth doing. There's a place for crossing national borders in order to see the entire Ring Cycle presented back to back on stage. And, there's a place for killing three minutes while you wait for coffee to brew by watching a StrongBad cartoon. That one is a more rewarding experience and more fun to tell your friends about doesn't detract from the other.
We don't want people to *not* spread SF0 to their myspace friends or sort the books on their bookshelf because they're worried about screwing up their satisfaction score.
I'd also suggest that if a scaling like YellowBear's is implemented, it uses something more sophisticated than just scaling by the highest score on a task. It's possible to take a simple little task and go over-the-top-apeshit-nuts with it and earn a lot of votes, but no everyone should be expected to do so for every simple task. Something like your votes divided by the median votes on the task might be more appropriate.
The metric I have used on occasion, which seems to work quite well, is points:votes. The closer that ratio is to 5:1, the more "other player satisfying" the players is. Another way of phrasing the same concept is the percent of your points that come from votes, rather than automatic upon completing the task. This was the metric that first brought Mouse to my attention, at the time at least substantially more than half of his points were from votes!
Now, both of these have their interesting side effects. Your metric favors people that avoid doing tasks that are below their current level. Mine favors people that do low-level tasks, because it's much harder to get 20 votes on every 100 point task you do than it is to get 2 votes on every 10 point task you do. Lord knows i've seen plenty of "10+45" scores and the like.
As for the idea of factoring in highest score in a task, I like it, but I agree with Loki in that it should not be the sole metric. Perhaps just making that number easily accessible elsewhere would be good. "This person has a player satisfaction rating of 40 and 5 highest score tasks."
So I started to crunch some numbers and I have decided that you may be a genious! I certainly like the way the results stack up, so far as my satisfaction number is concerned.
Don't flame me, you know you are all curious. I only did the first page of the Glastnost High Score ordering with the simple (points) / (tasks), what do you all think?
# Name Score Tasks Satisfaction
1 Lank: 2655 / 38 = 69.87
2 Spar: 2355 / 38 = 61.97
3 p00n: 2215 / 36 = 61.53
4 CyberKitty: 2090 / 36 = 58.06
5 Eddy: 1860 / 37 = 50.27
6 Burn Unit: 1860 / 27 = 68.89
7 Lincoln: 1640 / 32 = 51.25
8 YellowBear: 1554 / 20 = 77.75!
9 H. Kat: 1525 / 36 =42.36
10 Haddock: 1505 / 34 = 44.26
11 Anna 1: 1500 / 20 = 75
12 R. Bright: 1485 / 18 = 82.5!!! (Congrats My Friend)
13 El B: 1295 / 27 = 47.96
14 Lowteck: 1244 / 17 = 73.18
15 Yarmoth: 1195 / 26 = 45.96
16 Rongo: 1190 / 20 = 59.5
17 Darkaardv.: 1180 / 27 = 43.7
18 Kyle H: 1130 / 25 = 45.2
19 Ben Forbes: 990 / 18 = 55
20 Rabbit: 970 / 27 = 35.96
21 Bex: 970 / 18 = 53.89
22 El Cap: 885 / 17 = 52.08
23 Mouse: 870 / 11 = 79.09!!
24 Ziggy: 815 / 19 = 42.89
The players who are on the top vote page, but not on the high score page:
InkTea: 525 / 7 = 75
Jackie H: 660 / 8 = 82.5
Sean M: 800 / 10 = 80
Sam L: 705 / 8 = 88.16
Metatron: 760 / 15 = 50.67
Give Feedback!! do you agree/disagree with how you feel about your performance vs. the number you seem to have? should this be adjusted? do you agree that those with a higher number have consistently higher quality?
It seems like doing fewer tasks very well obviously improves your satisfaction, and it seems that keeping a very high satisfaction rating gets increasingly difficult as you go up in level (do more and more task). If a level 1 collaborated on a level 6 task and that was all they did their satifaction would be unrealistically high. Look at your own tasks, count fleurs or something and add ways you think this could be scaled to more acuratly depict your satisfaction level
I'd rather see calculations based on votes rather than score- regardless of the method, I'm probably not going to be in the top tier but as it stands, it feels like this method rewards people who do tasks "at their level" i.e. stick to the highest-scoring tasks they can- with an added bonus for those who do receive a lot of votes for their tasks as well.
I don't know. I guess to me, I like to judge from completion to completion- most of us have completions that got a lot of votes, and those that fell flat as well. The all-time votes board is another list that feels reasonable to me.
But I'm not as competition-oriented as a lot of people out there. To each his/her own and if people like this method, that's cool too.
The players who are on the top vote page, but not on the high score page:
InkTea: 525 / 7 = 75
Hot! It looks like I'm actually *accomplishing* something! I like this merely based on the fact that it makes me look less lazy and distracted than I am!
But in all earnesty, this gets at *exactly* something I was discussing with Burn Unit some time ago... but, even so- as an old school Impossible Exchanger, I'm hesitant to embrace change. My little high score fleur badges are plenty reward.
On the other hand? Anything to get players to push things to the limit one more time.
pretty acceptable metric--i think as another metric, not a pure replacement. more is more fun imho.
some players gain votes for non-quality reasons. trying to account for this in a formula would be a headache. it's not a reason to throw out the metric, but it's a grain of salt.
**Edit, after cocktail hour, same night: if you think that's hot Inky, check this one out
Ziggy & Yellowbear, thank you for some inspiring number work. For that matter, why not Votes/tasks? Hell, this could all be displayed almost baseball style, natch. (Although in baseball the number of hits per bat is limited at one, so obviously this metaphor isn't exact.) HOWEVER, there are some interesting things that happen when you start doing things like this.
Okay did a little number crunching of my own. Taking your list of players who had pretty high point counts per task, and instead of ranking them by points/task, went with #tasks and votes and then vpt. It would look like this (with some other players, not on your list, but in the top page + of results by # of votes:
Player-- #Tasks -- #Votes -- votes per task (vpt)
Inktea-- 7-- 141-- 20.14 vpt
Flameboy-- 7-- 118-- 16.86 vpt
JackieH-- 8-- 129-- 16.13 vpt
Sam L-- 8-- 126-- 15.75 vpt
Anna-- 20-- 275-- 13.75 vpt
Lowteck-- 17-- 233-- 13.71 vpt
Sean M-- 10-- 136-- 13.6 vpt
Rainbow Bright-- 18-- 232-- 12.89 vpt
YellowBear-- 20-- 254-- 12.7 vpt
Mouse-- 11-- 136-- 12.36 vpt
Lank-- 38-- 440-- 11.58 vpt
Cameron-- 9-- 94-- 10.44 vpt
Burn Unit-- 27-- 238-- 8.81 vpt
P00n-- 36-- 285-- 7.92 vpt
Spar-- 38-- 296-- 7.79 vpt
Mink-- 15-- 111-- 7.4 vpt
Katotoro-- 36-- 242-- 6.72 vpt
Bex-- 18-- 118-- 6.56 vpt
Rongo-- 20-- 130-- 6.5 vpt
El Cap-- 17-- 109-- 6.41 vpt
Heatherlynn-- 15-- 94-- 6.27 vpt
Villain-- 17-- 104-- 6.12 vpt
Eddy-- 37-- 201-- 5.43 vpt
Zemaluco-- 27-- 146-- 5.41 vpt
KristinawithaK-- 18-- 94-- 5.22 vpt
Lincoln-- 32-- 163-- 5.09 vpt
Ben Forbes-- 18-- 91-- 5.06 vpt
Ziggy C-- 19-- 89-- 4.68 vpt
Hemingway Kat-- 36-- 146-- 4.06 vpt
DarkAard-- 27-- 103-- 3.81 vpt
Yarmoth-- 26-- 87-- 3.35 vpt
Rabbit-- 27-- 82-- 3.04 vpt
Haddock-- 34-- 101-- 2.97 vpt
Kyle-- 25-- 67-- 2.68 vpt
This is not an unreasonable stat. It shows the trend of players who attract a lot of votes and is elastic for players with high raw nubmers of votes.
Now, let's reopen that baseball metaphor just for fun with some truly crazy statistics. In baseball they have that Slugging percentage, which is total number of bases divided by at bats, right? Well, consider this: you could count votes in a similarly fun/amusing way.
A task getting 1 or more vote would count as a hit, a single lets say.
If the points value of the task is matched by the points value of the vote, you have effectively doubled the score, yes?
If the point value of the votes is 2x or greater than the point value of the task, you have tripled.
And if the number of votes actually equals the point value of the task, it's safe to say that task is a home run, especially as the tasks get more difficult.
For example, a 15 point task receives 1 vote = single
a 15 point task receives 3 votes (15 points) = double
a 15 point task receives 7 votes (35 points, more than double) = triple
a 15 point task receives 15 votes = home run.
Abbreviate 1b, 2b, 3b, 4b... duh.
Thus a player can calculate their SF0 "slugging percentage" by
((1 *1B) +(2*2B)+(3*3B)+(4*4B))/tasks
Running these numbers gets involved, so I only ran the first four on the list above, plus myself
Inktea 20.14 vpt slugging metric 2.71
Flameboy 16.86 and 1.86
JackieH 16.13 and 2.13
SamL! 15.75 and 2
burn unit 8.81 and 2.19
Try it yourself! determine your Era-stats and all time stats! Play with the numbers!
I believe it would be possible to also do a quasi "on base plus slugging", but it's late, I've been drinking, and frankly, that's getting a little crazy.
about slugging, and ops
if we keep at it, we'll get a completely insane statistical picture of the best player(s) in the game. look at a player's rank on all time score, all time votes, number of task completed, points per task, votes per task, slugging percentage, number of tasks created by, number of other-player-comments per task, number of friends, number of collaborators per task, number of votes given to other players, delta of votes given & received. . .
average these things together and... I predict this statistical system will eventually reward Lowteck.
Thanks for the number work, Burn Unit!
The more I think about it, the more I believe my original metric was horribly flawed, especially since it's essentially just a measure of average points per task completion. If a higher level player completed a 125 point task and did a poor job on it, but it does not get X'ed, it would still ultimately boost their Player Satisfaction score/ranking.
However, toying with the numbers is great fun. I believe Burn's average VPT score is a much better measure than my original PPT score.
But hell, keeping track of both wouldn't hurt now, would it? Hell, if I had the time, I'd love to see all the stats that Burn just listed in his last comment. Maybe this is the realization of where my original idea was going: An in-depth player data page. Ranking pages for each catagory (or at least the important ones, such as VPT score, PPT score, etc.), as well as a summary that is displayed in each user's profile.
I don't feel like this will give the incentive to new players that I was hoping for, but it's something, anyway. Unless such data is displayed harshly, like as a certain number drops (or becomes less desirable) it'll get ever closer to a deathly red, while as the number improves, is becomes more and more an optimistic green. You can get a good idea of how well received they are by simply glancing at their data and judging by the colors.
I'm just spouting off ideas in a sleep-deprived state. Thanks for the comments, everyone.
Wow good numbers Mr. Unit! I love the idea of being able to toy with the stats in all different ways. Your newest stat makes me want to un post all my tasks with no votes thats for sure
Wow.
Um. My ego feels fluffed. And I feel guilty, because I thought I was kind of half-assing it.
on base plus slugging in baseball is interesting. it's
OPS = (AB (H + BB + HBP) + TB (AB+BB+SF+HBP))/AB(AB+BB+SF+HBP)
We don't use hit by pitches that I know of, and we don't use sacrifice flies so those would be zeroes anyway. So the SF0ops would be
OPS = (AB (H + BB) + TB (AB+BB))/AB(AB+BB)
where: H = Hits (raw votes)
BB = Base on balls- perhaps a proof with no votes?
AB = At bats - number of tasks TB = Total bases - per the slugging metric description above.
So I ran these numbers on myself:
TB= 61 votes = 241 tasks= 27
OPS = (AB (H + BB) + TB (AB+BB))/AB(AB+BB)
(29 (241 + 0) + 61 (29+0)) / 29 (29+0)
6989+1769/841
burn unit SF0ops, Glasnost=10.413
for comparison purposes, I used InkTea (because I'm lazy and it amuses me to boost her ego) 7 tasks, 142 votes, tb= 19
994+133/49
ink tea Sf0ops, Glasnost=23
Enjoy!
one thing to consider for statistics:
it may be best to keep impossible exchange out of it, seeing as there were many different voting systems.
-at the inception of SF0, there were no votes.
-then votes were limited and as such they weren't given out half as liberally as they are now. also during that period 6-8 votes would more than double the points of a task, as opposed to 5 points per vote.
- only towrads the end did we have our current voting system.
I don't follow baseball, but aren't statistics per season?
I didn't read a lot of this thread so forgive me if someone already brought this up
but also all time, and meta-stat comparisons are then done between seasons-- "the highest number of X in a season vs. the shortest number of seasons to reach Y benchmark" etc.
ImEpx will be like the pre-modern era--maybe the "dead vote" era (instead of the "dead ball" era of 19th century baseball)
I'm sorry, but I'm going to boast just this one time:
I've completed 19 tasks, with a Player Satisfaction score of 113.9 by Ziggy's formula. I rock.
Running Burn Unit's numbers gives a mildly less impressive result (13.68 vpt, and either 2.37 or 2.63 slugging metric depending on how the hell I was supposed to work that out), but still something I'm proud of.
I would love to see that chart...