Learning from my score

Questions
#1

I want to learn as much as I can from the cubes I submit.


Does a score of 20 points or less always mean I made significant mistakes?

After submitting a cube, is there any way to go back & look at what I did so I can try to figure out where I went wrong?

Are there any other ways to learn from what I’ve done?
#2
The points you'll get consist of:

Time: The more time you spend, the more points you'll get. The function is somehow sigmoidal (rather than linear). After approximately 2 minutes you are at the maximum of 20 points.

Volume: Adding more missing pieces will give you more points. Interestingly, the volume of the missing pieces is very important (more volume, more points)... maximum is 980.

Accuracy: Each cube is done by multiple players. Since the team doesn't know how the cube should look like (except for the tutorials), all your results are compared to those of other players and vice versa. High similarity means: no points deduction

Trailblazing: Of course, someone has to begin with a new cube - and there is nothing to compare the result with. In this case you'll become a "trailblazer" and honored with an immediate 50 points bonus. After 2 or 3 (not sure) other players did the same cube like you, you'll get your normal points retroactively.

What does that mean for you:
  • There is no need to hurry with a cube. Take all the time you need. (Even though trailblazing after coloring a cube for half an hour is unlikely.)
  • Going to bed with a cube open won't give you additional (free) time points.
  • The maximum amount of points you can get is 1000*
  • If there is nothing to add in your cube, you'll only get 20 points at max: the time bonus.
  • If you end up with 20 points and added quite a lot of stuff, you did something different from what most other players did. The underlined "different" is important since it doesn't necessarily mean you did something wrong - maybe you found another large branch the majority missed and therefore got a huge penalty.

To verify the last point, you can click "Back to overview" on the upper left corner of your screen and in most cases you can then see the current look of this cube (without your additions). This also works after you hit "I'm finished". With another click on "Start playing" you'll get back to the cube with all stuff you already added.
Note: If you try this trick after you pressed "I'm finished" and came back to the cube, you'll have to press "Skip cube" to move on to a new one. Clicking on "I'm finished" a second time won't work in that case.


* The team introduced a "Starburst Amacrine Cell" challenge. Those cells are a bit tougher to trace and all points you get are doubled. So the maximum time bonus becomes 40; maximum overall points 2000!
#3

Nice! Posted this on the blog!

http://blog.eyewire.org/how-are-points-calculated-in-eyewire/
#4

When I finish a cube, I’d like to know what the cube was worth in addition to how many points I got.  For example: 20 out of 20… 380 out of 500… 135 out of 135… 756 out of 1000.  That way I’d know if I screwed up or not.  Is this possible?


Thanks.
#5

I agree with jero. The score doesn’t tell the user anything about their actual accuracy relative to the “model” (deemed to be the compilation of all the other finished cubes). And even if it did, like 170/195 (195 being awarded when the cube is submitted in a form identical to the most often submitted form), it would not tell the user where they differed (rightly or wrongly, the most often submitted form being not necessarily the exact one).


Maybe I miss something but it seems that this lack of feedback is a major flaw as this prevents users from being able to correct and readjust their tracing strategy. Right now, I feel I am probably repeating the same mistakes again and again and having no feedback, it’s counterproductive.

Why not display the most submitted form of a cube after a user submit his/her finished cube, giving him/her a chance to understand his score, compare his work to the “model”, and realize where he/she may have gone wrong?

Otherwise, is there any other way to improve, i.e. know how to deal with difficult slices, etc…?
#6

Ah… got from the chat that the “overview” is a way to check one’s tracing against the consensus. I never used it. I thought it was a quick way to get an overview of the environment before tracing. I did not realize it actually gives away the updated cube…


Seems to me that giving away the consensus shape before hand and letting it accessible during tracing is counterproductive. I would do the exact opposite, i.e. revealing it only once the cube has been submitted, with color enhancement to show discrepancies. That would :
1) eliminate any bias during tracing
 2) warn the user about possible misinterpretation of pattern, hence learning from others, not copying others.

But that would also probably open another can of worms… :slight_smile:
#7

Indeed it would open another can of worms.  We tell people not to use the overview when they can help it, to help prevent against bias, but it’s also really useful, even necessary sometimes. Newer players can learn what a correct path looks like when they’re stumped or, if there’s a merger in a seed piece you can ascertain which seed piece is the correct one to trace.


Basically please don’t abuse it!