Question

Derive the formula for SmartScore.

Answer 1

LOL. Ask @cshalvey

Answer 2

-_- not allowed.

Answer 3

Heh:-)

Answer 4

this might help http://openstudy.com/users/lgbasallote#/updates/4f8e5fb7e4b000310facc223

Answer 5

LOL. Else ask @farmdawgnation. If you say not allowed, I have few more names to mention :P

Answer 6

Trying to derive the smartscore from the 3 subscores seems fruitless to me.

Answer 7

^scroll down to the dumbcow's answer

Answer 8

The tendency I'd have would be to calculate a single metric from those three subscores, weighting PS and T more than E. Then I guess I'd assign smart scores so that they were normally distributed.

However, I'd want to exclude scores of 0 from that distribution to account for people who had joined and never posted.

Answer 9

I'm trying to decide, based on what I've seen around the site, whether the scores seem normally distributed. And if so, what kind of mean and SD there is. If it is normal, the mean seems to be higher than 50. But that could just be that people under that mean are inactive.

Answer 10

according to dumbcow..

score ps tm eng
score 1.00
ps 0.89 1.00
tm 0.86 0.88 1.00
eng 0.13 -0.02 0.44 1.00

i dont understand what it means though

Answer 11

I also perceive a clumping around 90-100, but that could also just be explained by the fact that higher scoring people are more active and therefore more visible.

lgb, that part makes sense to me, I think. It's a statistical analysis of how good a predictor each score is. The closer to 1, the better.
For example, the top left entry is 1 because score is a perfect predictor of score.

Answer 12

That would indicate that PS and TM matter much more than engagement for producing your score. This indicates that there's some weighting of which scores matter, as there should be.

Answer 13

As I read the entire thread pointed out by @lgbasallote , I came to know that the high score-->high actvivity and not no of medals. See this profile: http://openstudy.com/users/akshay_budhkar

See my profile. I have scored roughly 2/10 of the medals this person has but still I have my smartscore higher than him. So smartscore is directly proportional to smartscore.

Answer 14

Typo; So smartscore is directly proportional to *activity

Answer 15

Okay... interesting. I wonder if the subscores are too.

Answer 16

See , the profile to which I have pointed out has buggy total count of medals. See subject wise distribution of medals