Points

[Melikki]

Can someone please explain how the points are calculated?

I can guess that it's based on:
Difficulty
Size
Map features
Region count
Solve time

Negatives:
Hints taken (if any)
Mistakes?
- Beyond the added time, are you charged if you clear the puzzle?
- Same question for "undo"

Positives:
Points for "in a row"? If so, does it increase based on how many in a row?

Thanks in advance for your help!
Jen

[Sting]

I have no idea, it seems very random.

Sean

[uigrad]

image.png
There's definitely something strange that's still going on. In this puzzle, I'm slower than the median, even though I have the fastest time.

My guess is that the fastest time and median are fake numbers until we finally have enough data to get real values. There's probably also a difficulty assessment (that we don't get to see). Then the difficulty assessment is used to set a maximum score, and then we get some fraction of the maximum based on our time, the fastest time, and the median.

This would fit what we've seen with other puzzles on the site. The "difficulty assessment" isn't seen in the older puzzle types, but it is shown directly in the new clueprint puzzles. I'm assuming that variable is here, but hidden.

If all my assumptions are accurate, I still have 2 questions:
1. Is the fake median related to the difficulty assessment? I've noticed a lot of puzzles with the exact same median, but the maximum points seem to be all over the board, so maybe they are not related.
2. How is the completion rate set? I played a puzzle with a 0% completion rate (although that may have been at the clueprint site, I can't remember for sure), and so it seems that it doesn't take my own play into account. But I've also seen a 75% completion amount when I'm the only person who has solved it. So, maybe the completion percentage is completely fake?

It would be nice if there was some tell for stats that are fake. Maybe just an asterisk or even a dagger (†) could be given with a footnote at the bottom saying "† - Not enough data"

[admin]

The vast majority of players are guests (not logged in) so their scores are reflected in "How you compare" but do not appear in "Top Solvers". As more and more registered players join the site though these two sets of numbers should start to converge.

[uigrad]

Ah, I guess that makes sense. I just assumed that all the regulars are logged in, so it would be very rare for the top solvers of a puzzle to all by anonymous.

Thanks!